The Automated Daily - Hacker News Edition

Please support this podcast by checking out our sponsors: - Prezi: Create AI presentations fast - https://try.prezi.com/automated_daily - Effortless AI design for presentations, websites, and more with Gamma - https://try.gamma.app/tad - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Static x86 to ARM translation - An arXiv paper unveils Elevator, a static binary translation system that converts x86-64 executables to AArch64 without source or symbols, enabling pre-deployment testing, certification, and signing. Europe-first digital sovereignty stack - A detailed migration story shows how swapping US cloud and SaaS tools for European and Swiss providers improves data jurisdiction control, reduces vendor dependence, and makes “values-based” infrastructure practical. Open-source Bambu printer connectivity - A new fork of OrcaSlicer restores full BambuNetwork remote printing for Bambu Lab printers, highlighting the ongoing tug-of-war between vendor lock-downs and community-driven device control. Seawater electrolysis stainless breakthrough - Researchers report SS-H2, a stainless steel alloy with dual-passivation that survives harsh seawater electrolysis voltages—potentially lowering green hydrogen costs by replacing titanium components. Pixter handheld preservation and emulation - A reverse-engineering effort documents Fisher-Price/Mattel Pixter hardware, dumps ROMs and cartridges, and delivers working emulators—preserving early-2000s kids’ software that was close to vanishing. Tiny on-device function-calling AI - Needle is an open 26M-parameter model distilled from Gemini aimed at reliable single-shot function calling on small devices, pushing private, low-latency tool use closer to phones and edge hardware. Bell Labs unsung operations work - A first-person Bell Labs interview spotlights applied operations research—inventory control, PBX simulation, and practical tooling—showing how disciplined optimization kept telecom systems efficient. Why sci-fi fonts look futuristic - A typography piece explains the repeatable visual cues—slants, cuts, tight kerning, metallic glow—that instantly signal “the future,” revealing how sci-fi design has become a codified shorthand. - Author Migrates Digital Infrastructure to European Providers to Boost Digital Sovereignty - Elevator proposes deterministic static x86-64 to AArch64 whole-program translation without heuristics - OrcaSlicer Fork Releases With Restored BambuNetwork Remote Printing for Bambu Lab Printers - HKU Develops Dual-Passivation Stainless Steel for Seawater Hydrogen Electrolyzers - Reverse Engineering Brings Full Emulation and Preservation to Fisher-Price Pixter Devices - Google Teases AI-Focused Googlebook Laptops Powered by Gemini, Due Fall 2026 - Substrate seeks Technical Success Manager to scale AI-driven healthcare billing operations - Cactus Compute Open-Sources Needle, a 26M-Parameter On-Device Function-Calling Model - Inside Bell Labs’ Applied Division: The Unglamorous Work Behind Telecom Innovation - Six Common Typography Tricks Films Use to Make Text Look ‘Futuristic’ Episode Transcript Static x86 to ARM translation Let’s begin with research that could reshape how we move software across CPU architectures. A new arXiv paper introduces “Elevator,” a static binary translation system that converts complete x86-64 executables into AArch64 binaries—without needing source code, symbols, or convenient assumptions about how the original binary is laid out. What makes it stand out is the philosophy: instead of guessing what ambiguous bytes mean and patching things up at runtime, Elevator explores the plausible interpretations ahead of time and keeps multiple paths when necessary, only discarding paths that would clearly crash. The payoff is big for security and compliance-minded environments: the resulting ARM binary is fully determined before deployment, so you can test and validate the exact code that will execute, then sign it. The cost is larger output binaries, but the paper claims performance that can compete with, or even beat, established dynamic approaches like user-mode QEMU—while shrinking the runtime “translator” surface area you’d otherwise have to trust. Europe-first digital sovereignty stack Staying with the theme of control and trust, another popular story is a first-person account of migrating a personal and business “digital stack” away from mostly US-based services and toward European—often Swiss—providers to improve digital sovereignty. This isn’t framed as anti-American tech; it’s about jurisdiction, policy risk, and reducing the chance that a vendor decision or legal shift suddenly changes the rules for your data. The author swapped Google Analytics for a self-hosted Matomo setup, moved email and password management into Proton’s ecosystem, and shifted compute and storage off US clouds and onto providers like Scaleway and OVH. They also replaced several developer-facing building blocks—transactional email, error tracking, even some OpenAI API usage—arguing that Europe’s ecosystem is more mature than many people assume. What’s most useful here is the realism: they kept exceptions where network effects and feature gaps still dominate, like Cloudflare for edge security and Stripe for payments, plus some US-based AI tooling. The broader takeaway is that “values-based infrastructure” is no longer a purely ideological slogan—it can be a manageable, mostly planning-heavy project that results in a professional, reliable setup. Open-source Bambu printer connectivity On the maker and device-control front, there’s a new fork of OrcaSlicer from the FULU Foundation aimed at restoring full BambuNetwork support for Bambu Lab 3D printers—specifically, remote printing over the internet instead of being restricted to LAN-only use. This matters because it sits right at the intersection of convenience, ownership, and security. Remote printing is a major workflow feature for many people, but it also raises questions about who controls the connectivity layer—users, the vendor, or the community. The project is early, but the intent is clear: bring back a prior “normal” workflow that some users feel was taken away. Expect plenty of debate around tradeoffs—because the same features that add convenience can also expand the attack surface if they’re not designed carefully. Seawater electrolysis stainless breakthrough Now to energy and materials science: researchers at the University of Hong Kong report a new stainless steel alloy, SS-H2, designed for the punishing conditions of seawater electrolysis for green hydrogen. The headline is durability. Seawater is corrosive, and the voltages involved in splitting water can destroy conventional stainless steel protection layers. This team claims their alloy forms a second protective layer at higher electrical potentials—driven by manganese—which is surprising because manganese is usually associated with worse corrosion resistance in stainless steel. If this holds up in real industrial designs, it could reduce cost by letting electrolyzers use cheaper, easier-to-manufacture stainless components instead of relying on expensive titanium in key places. The group says they’ve moved toward commercialization with patents and pilot-scale wire production, though turning that into full electrolyzer parts is still an engineering journey. Still, it’s a notable example of “boring” materials breakthroughs unlocking practical climate-tech gains. Pixter handheld preservation and emulation One of the most delightful preservation stories today comes from a reverse-engineering project focused on Fisher-Price and Mattel’s Pixter handhelds. The author describes what may be the first comprehensive effort to document the Pixter line—hardware, ROM and cartridge dumping, and emulators spanning multiple generations. Why it matters: Pixter wasn’t just a toy; it’s a time capsule of early-2000s kids’ software, custom cartridge ecosystems, and quirky hardware design. And it had a reputation for being hard to emulate and poorly documented. The project uncovered that many games run on custom virtual machines rather than native code, and it tackled unusual hurdles like preserving cartridge audio that relied on separate “melody chip” blobs. The end result—open tools and emulators—means this ecosystem doesn’t have to disappear as aging devices and cartridges fail. It’s a reminder that digital history isn’t only about famous consoles; it’s also about the everyday tech a generation grew up with. Tiny on-device function-calling AI In AI, there’s a smaller-is-the-new-useful story: Cactus Compute released “Needle,” an open 26-million-parameter model designed mainly for reliable, single-shot function calling on very small devices. The significance isn’t that it’s a general-purpose chatbot. It’s that it aims to do one job—tool use—predictably, in a footprint that starts to make on-device assistants and local automation feel more realistic on constrained hardware. And because the weights and dataset-generation tooling are open, developers can inspect it, tune it, and test how it behaves in their own environments. If this trend continues, we’ll likely see more AI components that are narrow, auditable, and fast—rather than one giant model trying to be everything. Bell Labs unsung operations work Two culture-and-craft notes to close. First, a first-person interview with a Bell Labs veteran highlights the applied, less-glamorous side of the legendary institution—work like PBX simulations, inventory control for expensive circuit packs, and practical tools that helped teams make decisions before modern software was everywhere. The story is a good counterweight to the usual Bell Labs mythology: breakthroughs mattered, but so did rigorous operations research and the day-to-day optimization that kept enormous systems reliable and affordable. And finally, a typography piece breaks down why sci-fi logos so often look the same. It argues there are familiar visual cues—slanted forms, sharp cuts, fused letters, missing strokes, metallic textures—that instantly communicate “the future,” even when the underlying design is pretty conventional. It’s interesting because it frames futuristic type not as prophecy, but as a shared design language that’s become a shortcut for genre signaling. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad - Prezi: Create AI presentations fast - https://try.prezi.com/automated_daily - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: SQLite gains archival format status - The U.S. Library of Congress now lists SQLite as a Recommended Storage Format for datasets, boosting long-term digital preservation with a transparent, widely adopted database file. AI makes work look done - A critique of generative AI at work warns of 'output-competence decoupling'—polished docs and updates that hide weak understanding, raising verification costs and failure risk. Efficient open model on AMD - Zyphra’s ZAYA1-8B mixture-of-experts model claims strong math results with few active parameters and highlights credible AMD MI300X training as an alternative to CUDA-centric stacks. Valve releases Steam Controller CAD - Valve published full Steam Controller and 'Puck' CAD files under CC BY-NC-SA, enabling community-made mounts, grips, skins, and repair-friendly accessories while protecting RF-critical areas. Permacomputing principles for sustainability - The Permacomputing working group proposes principles like 'Not Doing' and 'Expose the Seams' to cut e-waste, extend device lifetimes, and resist rebound effects from efficiency-driven compute growth. Booting Debian diskless over network - A hands-on guide shows how to run Debian 13 without touching local disks by PXE/iPXE booting into an iSCSI-backed ZFS volume—useful for dual-boot avoidance and lab-style setups. TI-83 BASIC programming revival - Boris Cherny’s updated TI-83 Plus BASIC tutorial offers a structured path from simple programs to graphics and input handling, keeping calculator programming accessible for students and hobbyists. Photoshop 2026 UI workflow regressions - A Photoshop power-user argues Adobe’s 2026 'modern UI' introduces focus and keyboard-navigation regressions, turning tiny dialog interactions into daily productivity drains for professionals. - Valve Publishes Steam Controller CAD Files Under Creative Commons License - AI Productivity Theater Is Flooding Workplaces With Unverifiable Output - Guide Explains TI-83 Plus BASIC Programming from Basics to Graphics and Game Logic - Library of Congress Lists SQLite as a Recommended Storage Format for Datasets - Permacomputing group publishes 10 principles for sustainable, resilient digital practice - GovernGPT seeks backend engineer to build database and infrastructure for AI fundraising agents - Agent Harness Kit auto-scaffolds multi-agent orchestration for existing code repos - Zyphra’s ZAYA1-8B claims DeepSeek-R1-level math performance with sub-1B active parameters - How to Boot Debian Disklessly via PXE/iPXE with a ZFS-Backed iSCSI Root Disk - Photoshop 2026 UI Refresh Criticized for Breaking Keyboard Focus and Pro Workflows Episode Transcript SQLite gains archival format status First up, a big win for boring, reliable tech: SQLite is now listed by the U.S. Library of Congress as a “Recommended Storage Format” for datasets. That’s the same preservation-minded category that previously included formats like CSV, JSON, and XML. Why it matters is simple: this is an institutional stamp that SQLite is transparent, well-documented, widely used, and realistically readable in the future. If you care about archiving, civic data, research reproducibility, or even just not losing access to old datasets, this pushes SQLite from “convenient” toward “trusted for the long haul.” AI makes work look done Staying with data—but moving from preservation to workplace reality—one essay argues generative AI is incentivizing people to appear productive instead of being reliable. The key idea is that LLMs can produce outputs that look polished—requirements docs, status updates, diagrams—without the human operator truly understanding or validating what they’re shipping. The author calls this a kind of competence split: the output seems credible, but the ability to explain or verify it is missing. The warning here isn’t abstract. If managers start rewarding volume and velocity of AI-authored artifacts, organizations can end up paying more for reading and verification, while quietly eroding the learning loops that used to create real expertise. Efficient open model on AMD On the AI model front, Zyphra released an open-weights mixture-of-experts model called ZAYA1-8B, and the headline claim is intriguing: strong math performance with a relatively small number of “active” parameters at inference time. The company also emphasizes that training happened on an AMD Instinct MI300X cluster built with IBM, which is notable in a world that still defaults to NVIDIA and CUDA for most serious training runs. Even if you take benchmark claims with healthy skepticism, the trend is clear: smaller, more efficient architectures and smarter test-time compute are being used to chase frontier-like reasoning—while the hardware ecosystem slowly becomes less single-vendor by default. Valve releases Steam Controller CAD Now to hardware openness in a more literal sense. Valve published the full CAD package for its new Steam Controller, including files for the controller and its “Puck” accessory, plus engineering notes about areas you can’t cover without hurting wireless signal or operation. The license is Creative Commons BY-NC-SA, so it’s squarely aimed at the modding and maker community, not mass-market accessory companies—though Valve says commercial makers can reach out for separate terms. Why this matters: it dramatically lowers the friction for custom stands, mounts, grip extenders, skins, and repairs. It’s also consistent with Valve’s pattern of treating hardware as an ecosystem—open enough that the community can extend it, not just consume it. Permacomputing principles for sustainability If you’ve been feeling uneasy about compute growth, the Permacomputing working group published a set of principles that tries to make that unease actionable. The framing is that computing isn’t just digital—it’s material, political, and resource-intensive, especially when you factor in chips, supply chains, and e-waste. A few principles stand out: “Observe First” and, more provocatively, “Not Doing,” arguing that refusing unnecessary tech can beat efficiency gains that sometimes backfire through rebound effects. It’s a timely counterweight to the idea that every problem deserves more software, more data, and more GPUs—especially when the environmental bill comes due later. Booting Debian diskless over network For the hands-on tinkerers, there’s a walkthrough of installing Debian 13 as a diskless system that boots over the network. The goal wasn’t just novelty—it was practical: run Linux on a gaming PC without repartitioning local drives or risking a Windows update messing with a dual-boot setup. The approach uses PXE and iPXE for boot, and then points the machine at an iSCSI-backed ZFS volume hosted on a server. The tradeoff is some performance and extra moving parts, but the payoff is clean separation: your Linux install lives on the network, your local disks stay untouched, and you learn a lot about how booting actually works along the way. TI-83 BASIC programming revival A different kind of nostalgia-meets-practice story: Boris Cherny’s TI-83 Plus BASIC Programming Tutorial got an updated release. If you grew up with these calculators, you already know the appeal—tiny programs that feel immediate, tactile, and surprisingly empowering. What makes this guide useful is the structure: it starts with the basics of output and flow control, then builds toward interactivity, handling keypresses, and even drawing graphics. The bigger point is that constrained platforms can still be a fantastic way to learn programming fundamentals, because every decision is visible—and every bit of polish is earned. Photoshop 2026 UI workflow regressions And finally, a very modern frustration: a critique of Photoshop 2026’s updated “modern” UI, focused on something that sounds small until it ruins your day—keyboard focus and text-field behavior in dialogs. The author argues Adobe introduced regressions that break fast, muscle-memory workflows: fields that don’t auto-focus, values that don’t auto-select, clicks that don’t reliably put you where you think you are, and modal interruptions that steal control. For casual users it might be minor, but for professionals who live in these dialogs for hours, these micro-frictions add up to real productivity loss. It’s a reminder that UI work isn’t just cosmetics—interaction details are part of performance. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad - Consensus: AI for Research. Get a free month - https://get.consensus.app/automated_daily Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: GitHub outages visualized as heatmap - A satirical “Red Squares” page reframes GitHub’s contribution grid as an outage heatmap, spotlighting reliability risk and cumulative downtime for critical developer infrastructure. AI coding agents and org bottlenecks - A .txt engineer argues AI coding agents make implementation cheaper, but shift the real constraint to specs, coordination, and organizational context—raising the stakes on prioritization and culture. Agent-ready cloud onboarding protocols - Cloudflare and Stripe Projects push “agent-ready” provisioning with authorization and tokenized payments, hinting at standardized, programmable deployment workflows—and new questions about control and safety. Battery-free ultrasonic smart tags - Georgia Tech’s battery-free ultrasonic tags identify object movements via unique acoustic fingerprints, enabling low-maintenance smart sensing with a more privacy-preserving footprint than always-on audio. Quadruped robotics on a budget - CARA 2.0 shows how low-cost motors, custom firmware, and pragmatic mechanical redesigns can yield dynamic legged robots—while exposing the tradeoffs of bargain components. Ultima Online server reverse-engineered - A developer translated a 1998 Ultima Online demo server from x86 binary to portable C, preserving MMO history and enabling research into dormant systems like ecology and AI behaviors. Plant electrical signaling history - Jagadish Chandra Bose’s contested 1926 plant-signal demonstrations highlight how instruments, metaphors, and academic gatekeeping shape what counts as legitimate science—now echoed in modern plant signaling work. - Satirical ‘Red Squares’ Chart Tracks GitHub Outages as a Contribution Heatmap - Coding Agents Shift the Bottleneck from Code to Organizational Context - Guide Details How to Run Sun Ray Server Software on OpenIndiana Hipster 2025.10 - Cloudflare and Stripe Projects Enable AI Agents to Provision Accounts, Payments, and Domains for Deployments - Star Labs details StarFighter 16-inch Linux laptop with coreboot, removable webcam, and 4K 120Hz display - Aaed Musa Debuts CARA 2.0, a Lower-Cost Capstan-Drive Quadruped Built for Hobbyist-Level Dynamics - Jagadish Chandra Bose and the Long Fight Over Whether Plants Have Nerve-Like Signals - Ten-Year Project Rebuilds the 1998 Ultima Online Demo Server in Portable C - Kate Davies Warns AI-Generated Knitting Media Is Polished, Profitable, and Unmoored from Truth - Georgia Tech Creates Battery-Free Ultrasonic Metal Tags for Smart Home Activity Sensing Episode Transcript GitHub outages visualized as heatmap Let’s start with that GitHub reliability gut-check. A satirical site called “Red Squares” reimagines GitHub’s contribution chart as an outage heatmap—each square representing a day with incidents, and darker shades meaning longer disruptions. The punchline is the cumulative story: lots of small incidents can add up to a surprisingly large amount of lost time over a year. It pulls data reconstructed from GitHub’s public status history, skipping scheduled maintenance, and then presents it in a format every developer instantly recognizes. It matters because GitHub isn’t just a website—it’s plumbing for modern software work. When it wobbles, deployments stall, collaboration slows, and teams lose momentum in the most expensive way possible: by interruption. AI coding agents and org bottlenecks Staying in the “software work is more than code” theme, one engineer wrote about evaluating structured-generation algorithms by checking whether a model’s token distribution is correct—not just whether the final string “passes.” Along the way, there’s a telling detail: after a short explanation, a coding agent produced a working prototype quickly. The broader point, though, is less hype-y and more useful: coding agents can make individuals faster, but that doesn’t automatically make organizations faster. When implementation gets cheap, the bottleneck moves to specifying what you actually want, aligning stakeholders, and keeping priorities sharp. The warning is classic Jevons paradox: cheaper “making stuff” can lead to feature bloat unless teams get even better at saying no. The proposed antidote is to externalize context—having agents continuously read repos, issues, and discussions to extract decisions into durable artifacts that both humans and other agents can rely on. Agent-ready cloud onboarding protocols That dovetails with a broader industry push toward making production deployment “agent-friendly.” Cloudflare announced an integration with Stripe Projects aimed at letting AI agents provision what they need—accounts, services, and billing—without the human doing the usual dashboard scavenger hunt for tokens and settings. The human is still present for permissions and terms, while payments are handled in a tokenized way so the agent doesn’t see raw card details. Zooming out, the interesting part isn’t one vendor combo; it’s the direction of travel. If onboarding and billing become programmable building blocks, deploying software could look more like calling an API than following a checklist. The flip side is governance: as it gets easier for agents to spin up real infrastructure, teams will need clearer guardrails for cost, access, and accountability. Battery-free ultrasonic smart tags On the sensing and smart-home front, Georgia Tech researchers showed tiny, battery-free metal tags that act like sensors by emitting brief ultrasonic pulses when they’re physically moved—like when you open a drawer or door. Each tag has a unique acoustic fingerprint based on its shape, so a nearby wearable or microphone-equipped device can tell which object moved and log the event. What’s notable here is the practicality: no batteries to replace, no charging habits to enforce, and no expectation that you’ll maintain a home full of fragile gadgets. The team also leans on lightweight rules instead of heavy ML, and because the signals are ultrasonic and short-range, it’s designed to be quieter and less intrusive than always-on listening. The use cases—elder care routines, activity tracking, general “did this get opened?” monitoring—are exactly where low-maintenance reliability beats fancy dashboards. Quadruped robotics on a budget In robotics, a creator shared CARA 2.0, an upgraded quadruped built under senior-project constraints, with a very grounded mission: keep it dynamic, but lower cost and weight. The most interesting takeaway is how much real engineering sits between “cheap parts” and “working robot.” They leaned on quasi-direct-drive actuators using inexpensive motors and controllers, then had to wrestle reliability issues—especially around feedback stability—before the whole machine could walk consistently. Mechanically, they iterated toward simpler structures, better traction, and more balanced packaging, and they even ran into a classic symmetry gotcha: legs that look identical aren’t necessarily mirror-correct, and that can show up as a persistent turning bias. This is the kind of write-up that matters because it documents the messy path from prototype bravado to repeatable behavior. Ultima Online server reverse-engineered For the preservation crowd, there’s a big reverse-engineering release: a developer reconstructed the 1998 Ultima Online demo server, translating thousands of disassembled functions into portable C with careful, instruction-level verification. Beyond nostalgia, this is a serious archival move—taking a historically important MMO server codebase, preserving its behavior, and making it inspectable. Along the way, the work uncovers dormant systems, including pieces of Ultima Online’s old “ecology” logic, and it adds compatibility improvements while marking exactly where it diverges from the original. The author is also calling for surviving server data files to improve historical accuracy. Projects like this sit at the intersection of software history, digital preservation, and the practical reality that online worlds disappear when their servers do. Plant electrical signaling history Finally, a historical piece with modern resonance: in the 1920s, Indian scientist Jagadish Chandra Bose demonstrated electrical signaling and fluid responses in plants, arguing they show heartbeat-like rhythms and nerve-like activity. He was admired by some giants of physics, but fiercely criticized by many biologists who felt the claims outran the evidence and leaned too heavily on metaphor. The story is a reminder that what gets accepted as “real science” depends not just on ideas, but on instruments, reproducibility, and the boundaries a field chooses to enforce. Today, plant electrical signaling is an active research area—like plant-wide warning signals after damage—so Bose’s legacy is being reexamined. It’s a useful case study in how a topic can be scientifically interesting and still get professionally sidelined for decades. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Consensus: AI for Research. Get a free month - https://get.consensus.app/automated_daily - Prezi: Create AI presentations fast - https://try.prezi.com/automated_daily - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Chrome silently downloads 4GB AI - A researcher says Google Chrome is quietly pulling down a ~4 GB on-device Gemini Nano model file, raising consent, transparency, bandwidth, and GDPR/ePrivacy concerns. Companies stuck in AI messy middle - Organizations are adopting AI tools unevenly, and the real challenge is turning individual Copilot-style wins into reusable workflows, better decisions, and faster learning loops—without surveillance. Agentic coding changes software practice - As AI agents make coding cheap, the article argues teams should invest more in intent, architecture, and end-to-end tests, because maintenance, security, and operations stay expensive. OpenAI scales low-latency voice via WebRTC - OpenAI describes an updated WebRTC architecture for ChatGPT voice and a Realtime API, prioritizing low latency, fast session setup, and global routing while avoiding massive UDP exposure. Async Rust code size bloat fixes - A Rust developer shows how compiler-generated futures can bloat binaries—especially embedded and wasm—and proposes optimizations like removing post-completion panics and collapsing nested futures. Bun experiments with Zig-to-Rust port - Bun published a Zig-to-Rust porting guide and tooling to systematically translate parts of its codebase while preserving its event loop model, sparking debate about a possible rewrite. Docker Compose still viable in production - A 2026 ops take says Docker Compose can still run real production—especially single-node and customer-hosted—if you handle log growth, orphaned containers, pinned images, and safer updates. Hand-drawn QR code actually works - A developer successfully drew a scannable QR code by hand on sticky-note grid paper, illustrating QR constraints, error correction, and how physical presentation affects scan reliability. - Developer Proposes Compiler Changes to Cut Async Rust Future Bloat - Docker Compose Can Work in Production in 2026—If You Fill Its Operational Gaps - Why Widespread AI Use Often Fails to Produce Organizational Learning - Bun Adds Zig-to-Rust Phase A Porting Guide and Batch Porting Script - Website Tracks AMC Showtimes With Zero Tickets Sold - Ten Lessons for Building Software with AI Coding Agents - Seth Larson Hand-Draws a Working Version 1 QR Code on Grid Paper - Report Claims Chrome Quietly Downloads 4GB Gemini Nano Model Without User Consent - OpenAI Rebuilds WebRTC Stack with Relay-and-Transceiver Design to Cut Voice Latency Episode Transcript Chrome silently downloads 4GB AI First up: a privacy researcher claims recent versions of Google Chrome are silently downloading a roughly 4 GB on-device model file—described as Gemini Nano weights—into user profile directories. The report says it can happen even on fresh profiles with essentially no interaction, and that deleting the file may trigger a re-download unless you disable certain AI-related features via flags or policy controls. Why it matters: this isn’t just a privacy and consent story under frameworks like GDPR and ePrivacy—it’s also about control over your device, unexpected storage and bandwidth costs, and the environmental footprint of pushing huge files at browser scale. If the allegation holds, it sets a precedent for “vendor-managed” payloads landing on personal machines with limited transparency. Companies stuck in AI messy middle Zooming out to the workplace, there’s a strong theme in two different essays today: many companies are drifting into an AI “messy middle,” and teams are still figuring out what it means to build software when code generation is fast and cheap. One piece argues the gap isn’t access—lots of people already have copilots and chatbots—it’s conversion. Individual employees learn clever workflows, but those insights don’t reliably become shared capability. The author’s takeaway is to measure outcomes like faster decision cycles and better verification, not raw usage metrics, and to build lightweight operational scaffolding for permissions and reuse. The warning is important too: if you try to capture learning by watching employees too closely, you’ll get compliance theater and hidden experimentation. A second essay pushes a practical mindset shift for “agentic coding.” The idea is that when implementation is cheap, you should expect to iterate more, treat the build as a way to discover the real spec, and put your human effort into the parts agents don’t magically solve—architecture, security, resilience, and long-term maintenance. The common thread between both: the advantage isn’t having AI. It’s increasing learning velocity without creating chaos. Agentic coding changes software practice On real-time AI: OpenAI published details on how it scaled voice interactions for ChatGPT voice and a Realtime API, and the headline constraint is latency. Natural conversation demands quick session setup and consistently low lag, and that gets brutal at global scale. The interesting bit here isn’t the brand name—it’s the architectural bet. They describe separating “routing” from “termination” so they can keep the public network footprint smaller while still steering UDP media traffic to the right place fast. Why this matters to the broader ecosystem: it’s a reminder that as voice and live multimodal experiences grow, the hardest problems often look less like model quality and more like distributed systems engineering—network paths, state, reliability, and operational safety. OpenAI scales low-latency voice via WebRTC Now to programming languages and compilers—starting with async Rust, and a complaint you’ll recognize if you’ve shipped anything size-sensitive. A Rust developer argues async still feels like a “minimum viable product” in one key way: compiler-generated futures can bloat code size, which is particularly painful on embedded targets and also relevant to wasm. By looking at compiler internals and the shape of generated state machines, the author points to overhead that shows up even in trivial async blocks—extra states and panic paths that can block downstream optimization. They propose several compiler-level improvements, like not forcing a panic in release builds if a completed future is polled again, and skipping state machine generation entirely when there are no await points. The bigger opportunity they highlight is collapsing nested futures so thin wrappers don’t become separate state machines. The reason to care: these changes push async closer to Rust’s “zero-cost” promise in places where every kilobyte and every instruction counts, and the author says they’ve already seen measurable binary-size and performance wins with prototype hacks. Async Rust code size bloat fixes Staying with Rust, but from the ecosystem angle: Bun added a fairly strict Zig-to-Rust porting guide as part of an experiment to translate portions of its Zig codebase into Rust. The guide isn’t just style advice—it sets guardrails about what Rust features and libraries are acceptable so the runtime behavior stays aligned with Bun’s existing event loop and system-call model. What makes this notable is the process: it’s an attempt to make a large-scale port repeatable and auditable, without prematurely committing to a full rewrite. That’s an increasingly common pattern in big codebases—running two worlds in parallel long enough to compare maintainability, testing confidence, and performance before making a point-of-no-return decision. Bun experiments with Zig-to-Rust port On the operations side, there’s a pragmatic argument that plain Docker Compose can still be production-worthy in 2026—especially for single-node deployments and software that runs in customer environments—if you’re honest about what Compose doesn’t do for you. The post calls out real failure modes operators keep rediscovering: orphaned containers sticking around after configuration changes, disks filling up from images and unbounded JSON logs, and health checks that mark a service unhealthy without automatically triggering recovery. The bigger lesson is that “simple” tooling isn’t automatically safe—it just shifts responsibility to operational discipline. The author also highlights two habits that matter a lot in the field: pin images by immutable digests to avoid silent drift, and treat access to the Docker socket as effectively root access to the host. Compose can work, but it needs guardrails and a plan for updates when you have many customer installs. Docker Compose still viable in production Finally, a lighter engineering story with a real lesson inside it: a developer managed to draw a working QR code by hand on gridded sticky-note paper. They ran into the classic constraint—small QR versions don’t hold much data—and ended up exploring encoding choices and error correction in a very tangible way. Why this is more than a novelty: it’s a great demonstration of how robust well-designed formats can be in imperfect real-world conditions. Even with a small drawing mistake, scanning still worked thanks to error correction—and the physical setup mattered too, like whether the paper was curled. It’s a reminder that “digital” systems often live or die on messy physical details. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: GameStop bids to buy eBay - GameStop launched a surprise cash-and-stock bid for eBay, raising questions about activist strategy, leverage, and whether cost-cutting can revive a shrinking marketplace. ASML EUV machine as Lego - ASML’s EUV lithography tools power advanced chips, but a scarce Lego replica became the unlikely hot item—showing how tangible models help explain complex engineering. Notepad++ macOS impersonation site - Notepad++ creator Don Ho warned about a fake website claiming an “official” macOS release, highlighting trademark abuse, user confusion, and malware-risk lookalike distribution. Reliable text in AI images - A practical workflow pairs deterministic layouts (SVG/HTML) with generative rendering so AI image models preserve accurate text and numbering—useful for design, UI mockups, and diagrams. Why huge AI contexts disappoint - A CACM report argues long context windows are constrained by memory bandwidth and KV-cache pressure, leading to latency, paging, and ‘context rot’ despite big token counts. Cheaper backends for coding agents - The open-source deepclaude project reroutes Claude Code-style agent workflows to cheaper Anthropic-compatible APIs, cutting costs and accelerating experimentation with coding agents. Reverse-engineering hard drives via timing - Researchers used read-only microbenchmarks to infer HDD physical geometry and remapping behavior, showing timing alone can reveal modern firmware-controlled storage quirks. Unplugged programming education for kids - NHK’s “Texico” teaches computational thinking without computers, reflecting a broader push for accessible digital literacy through ‘unplugged’ problem-solving skills. Giant RC airliner engineering goes mainstream - A creator-built, flyable RC Airbus A380 drew attention to how hobby projects now rival aerospace-style systems engineering, amplified by online audiences. - GameStop launches $55.5bn bid to acquire eBay - ASML’s Most Coveted Item Becomes a Scarce Employee-Only Lego EUV Set - Notepad++ Warns of Fake ‘Mac Version’ Site Using Its Trademark - ‘Underdrawing’ technique improves text and number accuracy in AI-generated images - NHK WORLD-JAPAN Introduces ‘Texico,’ an Unplugged Programming Education Series for Kids - Ramy RC Reveals Giant 1/8-Scale Lufthansa A380 for Airline’s 100th Anniversary - Nullagent Unveils BYOMesh Dual-Band LoRa Companion Dev Kit - deepclaude Lets Claude Code Run on DeepSeek and Other Anthropic-Compatible Backends - Microbenchmarks Reveal Modern Hard Drive Geometry, Layouts, and Defect Handling - Nvidia’s Rubin CPX Signals a Push Toward Billion-Token AI Contexts Episode Transcript GameStop bids to buy eBay Let’s start with the headline that made everyone double-take: GameStop has made an unsolicited cash-and-stock offer to buy eBay. The proposal puts a big premium on eBay’s share price, and Ryan Cohen says if the board says no, he’ll take the pitch directly to shareholders. GameStop is also promising aggressive cost cuts—particularly in sales and marketing—and says it has a major debt-financing commitment lined up. Why it matters: it’s a high-stakes bet that financial engineering and consolidation can fix two businesses facing real competitive pressure, and markets immediately began pricing in both the upside and the risk of heavy leverage. ASML EUV machine as Lego Sticking with markets—but moving to chips—there’s a delightful contrast coming out of ASML. The company’s EUV lithography tools are among the most important, hardest-to-build machines in modern industry, enabling the leading edge of smartphone and AI chips. And yet, the most coveted ASML “product” right now is a detailed Lego replica of the EUV machine, originally built as a side project by an employee. It’s scarce, restricted internally, and that scarcity has created a resale frenzy. Beyond the novelty, it’s a reminder that the most complex technologies are often invisible to the public—and sometimes a physical model is the best way to make abstract engineering feel real. Notepad++ macOS impersonation site Next, a security-and-trust story that’s depressingly familiar: Notepad++ creator Don Ho is warning about a website that claims to offer an official macOS version of Notepad++. According to Ho, it’s not affiliated with the project, misuses the trademark, and even copies personal bio details to look legitimate—enough to fool casual readers and, apparently, some tech coverage. The key point is simple: there is no official Notepad++ for macOS. Why it matters: impersonation sites don’t just confuse people; they can damage open-source reputations and become distribution points for unwanted or risky software. Reliable text in AI images On the AI front, there’s a practical lesson about image generation: models are getting better at text, but they still regularly mess up numbers, labels, and sequences—exactly the stuff you need to be correct in a board game, a UI mockup, or an infographic. One approach gaining traction is to split the work: generate a precise ‘underdrawing’ with deterministic tools—think SVG or HTML—so the text is correct, then have a multimodal model re-render it in a chosen style while preserving the layout. The significance is less about art tricks and more about reliability: it’s a workflow that treats AI as a stylist, not a typesetter, which is often the right division of labor. Why huge AI contexts disappoint Related to AI expectations versus reality, Communications of the ACM has a sobering take on huge context windows. Vendors now advertise massive token limits, but the report argues the bottleneck is frequently memory bandwidth and the sheer weight of the KV cache during inference. As the context grows, systems end up compressing, evicting, or paging data to slower memory tiers, raising latency and sometimes degrading output. On top of that, there’s the human-facing issue: attention gets spread thin, and long prompts can turn into noise—what some call ‘context rot.’ Why it matters: the next leap in AI usefulness may come less from bigger context numbers and more from smarter memory, retrieval, and hardware designed for inference-first workloads. Cheaper backends for coding agents If you’ve been watching coding agents evolve, there’s a notable open-source project called deepclaude that aims to keep the Claude Code-style workflow—where an agent iterates through edits and tool calls—while routing the model requests to cheaper, Anthropic-compatible backends. The pitch is straightforward: agent loops can get expensive fast, so swapping the underlying model provider can dramatically lower the cost of experimentation. The bigger story is ecosystem gravity: once an API style becomes a de facto standard, a whole layer of compatibility tools emerges, and developers start treating model choice like an interchangeable component rather than a locked-in platform decision. Reverse-engineering hard drives via timing For the storage nerds—and anyone who likes clever measurement—there’s a write-up on inferring the hidden physical geometry of hard drives using read-only timing tests. The author disables caches, times carefully chosen sector reads, and uses those measurements to tease out things like rotation behavior, track boundaries, and even how the drive works around defects. The key takeaway is that modern drives are no longer the neat, cylinder-aligned devices older textbooks imply; firmware-controlled layouts and remapping strategies make them far more complex. Why it matters: even when hardware is a black box, performance fingerprints can reveal design choices—and that can inform everything from benchmarking methodology to forensic analysis. Unplugged programming education for kids In education news, NHK WORLD-JAPAN has launched a kids-focused series called “Texico” that teaches core programming concepts without requiring a computer. It’s part of a broader movement toward ‘unplugged’ computational thinking—using stories and everyday reasoning to teach analysis, abstraction, and simulation as mental tools, not just screen skills. Why it matters: access to devices is uneven, and even when kids have screens, foundational thinking can get lost in the apps. Teaching the concepts away from the keyboard can make coding feel less like a club you need gear to join—and more like a way of thinking anyone can practice. Giant RC airliner engineering goes mainstream And finally, a lighter story with a serious engineering undertone: a YouTube creator has unveiled an enormous, flyable radio-controlled Airbus A380 replica. This isn’t a desk model—it’s the kind of build that forces you to think about structure, control systems, and ground handling like a real aircraft program, just scaled down and driven by hobbyist obsession. Why it matters: the line between hobby engineering and professional-grade systems keeps blurring, and online platforms are turning niche feats into mass-audience showcases—pulling more people into the mindset of building complex things carefully. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Crypto proof verifier flaw - Researchers found a soundness bug in Dusk Network’s PLONK verifier that could let attackers forge zero-knowledge proofs and mint coins from nothing—highlighting verifier-binding and KZG opening checks. Universal DO_NOT_TRACK telemetry opt-out - A proposed DO_NOT_TRACK=1 environment variable would standardize opt-out for telemetry, crash reporting, and other non-essential network calls across CLIs, SDKs, and developer tools—privacy by default. Haskell reliability at fintech scale - Mercury’s engineering write-up shows how a large Haskell codebase stays reliable by designing for graceful degradation, encoding safe defaults in types, and using Temporal for durable workflows and retries. Ladybird browser performance sprint - Ladybird’s April update delivered big usability gains like an inline PDF viewer plus faster parsing, smoother rendering, and better responsiveness—evidence the independent browser is maturing quickly. Windows 11 Insider program reset - Microsoft is reshaping Windows 11 Insider into Experimental and Beta channels, reducing surprise rollouts, tuning Windows Update restarts, and “right-sizing” AI surfaces like Copilot integrations. 3D Gaussian splats in-browser - The ml-sharp-web project demonstrates end-to-end 3D reconstruction in the browser—ONNX Runtime Web, workers, and exports—while exposing real constraints like COOP/COEP and model licensing. Why brain averages mislead - A Stanford Medicine study shows group-averaged fMRI results can hide or reverse within-person brain–behavior links, supporting individualized models for cognition and conditions like ADHD. Neanderthal bone-grease “fat factory” - Archaeologists report Neanderthals processed bone grease at scale 125,000 years ago, implying coordinated labor and planning—pushing back timelines for complex resource management. Building usable maps on Watch - A six-year Apple Watch mapping effort highlights the tradeoffs of on-device tile rendering, tiny-screen UI design, and why some developers bypass MapKit to achieve better outdoor navigation. - How Mercury Runs a Two-Million-Line Haskell Codebase in Production - ml-sharp-web Brings Apple’s SHARP Gaussian Splat Generation to the Browser - Ladybird April 2026 Update Brings Inline PDFs, Faster Parsing, and Major Performance Gains - Pedometer++ Creator Details Six-Year Push for Better watchOS Maps - Stanford study shows group-averaged brain scans can misrepresent individual cognitive control - Microsoft details Windows 11 quality improvements rolling out to Insiders - Critical Bug in Dusk’s PLONK Verifier Allowed Forged Proofs and Token Minting - Proposal Introduces DO_NOT_TRACK Environment Variable to Disable Telemetry Across Tools - Study Finds Neanderthals Rendered Bone Grease in a 125,000-Year-Old ‘Fat Factory’ Episode Transcript Crypto proof verifier flaw First up, a serious security disclosure in the zero-knowledge world. Researchers say they found a soundness flaw in Dusk Network’s dusk-plonk verifier—the component that decides whether a cryptographic proof is valid. The issue wasn’t about a clever new attack technique so much as a missing check: some values the verifier relied on weren’t properly bound to the trusted commitments. In plain terms, that can turn “math-proof security” into “attacker gets to pick the answer.” Why it matters: the authors demonstrated a proof-of-concept where they created new coins from zero and sent funds to an honest wallet on a local testnet, with nodes accepting the transactions. Dusk has shipped a patch, but the bigger takeaway is ecosystem-wide: verifier implementations need more mechanical, standardized safeguards so “unbound evaluation” bugs don’t quietly slip into production code. Universal DO_NOT_TRACK telemetry opt-out Staying on trust and control, there’s a new proposal aimed at cleaning up the mess of telemetry opt-outs in developer tools. The idea is a single environment variable—DO_NOT_TRACK=1—that would tell CLIs, SDKs, and frameworks to disable non-essential network calls like usage analytics, crash reporting, and similar reporting. Why it matters: today, every tool seems to invent its own switch, which means privacy becomes a scavenger hunt. A universal signal wouldn’t fix the broader debate about what software should collect—but it could make user intent unambiguous, and it nudges the industry toward opt-in instead of opt-out defaults. Haskell reliability at fintech scale On production engineering, Mercury published a detailed look at running a roughly two-million-line Haskell system while moving real money at high volume—much of it maintained by people who learned Haskell after joining. The core argument is refreshing: reliability isn’t the absence of failure; it’s “adaptive capacity”—systems that fail in predictable ways, are understandable under pressure, and make the safe path the easiest path. A key theme is using the type system as an operational tool, not an ideology. Mercury’s engineers encode institutional knowledge into APIs so that critical actions—like emitting the right transactional events—are hard to forget by accident. But they also warn against over-encoding invariants so tightly that teams can’t evolve. And one very practical reliability win: adopting Temporal for durable workflows, so retries, timeouts, and crash recovery are handled consistently in a way that fits nicely with Haskell’s “pure core, impure shell” mindset. The post also calls out a perennial Haskell pain point—observability—and recommends patterns like explicit instrumentation hooks and avoiding library-level logging so production teams can trace behavior when it counts. Ladybird browser performance sprint Now to browsers—specifically, the Ladybird project’s April progress report. They merged hundreds of pull requests, brought in new contributors, and landed meaningful user-facing upgrades like an inline PDF viewer powered by Mozilla’s pdf.js. That’s paired with profiling-driven work to keep large PDFs from bogging down the experience. Under the hood, Ladybird is also pushing on the stuff that makes a browser feel “fast” in daily use: getting earlier starts on parsing, moving more work off the main thread, improving rendering responsiveness with more parallelism, and tightening up the JavaScript engine’s speed and memory use. Why it matters: independent browser engines are rare, and Ladybird’s steady climb in standards testing and performance work suggests it’s moving from “interesting experiment” toward “credible alternative,” which is healthy for the web ecosystem. Windows 11 Insider program reset Microsoft also shared a progress update—this time on Windows 11 quality improvements and how it’s reworking the Windows Insider Program. The headline is simplification: two main channels, Experimental and Beta, with fewer confusing rollout mechanics. Beta is positioned to be more predictable, while Experimental becomes the place to try new features more explicitly—without trapping people in a maze of hidden switches. They’re also trying to make Windows Update less disruptive, including efforts to consolidate restarts and keep shutdown options visible even when updates are pending. And in a notable bit of restraint, Microsoft says it’s “right-sizing” AI integration—removing “Ask Copilot” from some app surfaces and making AI entry points feel more intentional. Why it matters: Windows reliability isn’t just about fewer crashes; it’s also about fewer surprises. Clearer channels, calmer defaults, and less forced AI could rebuild some trust with power users who’ve felt like involuntary beta testers. 3D Gaussian splats in-browser On the AI-in-the-browser front, a new open-source project called ml-sharp-web shows a complete, browser-based playground for generating 3D “Gaussian splats” from a single image using Apple’s SHARP model. It runs end-to-end in the browser: inference happens with ONNX Runtime Web in a worker, then the result is previewed and exported as a downloadable file. Why it matters: this is a concrete demo of heavyweight AI workloads moving closer to the client side, which can improve privacy and reduce server costs. But it also documents the less glamorous reality: browser constraints like cross-origin isolation requirements for high-performance WebAssembly, real deployment friction around large model files, and licensing gotchas where model weights have very different terms than the surrounding code. Why brain averages mislead Switching to science, Stanford Medicine published results that challenge a common habit in brain imaging: averaging people together and assuming the result reflects individuals. Using fMRI data from thousands of children doing an inhibitory-control task, the researchers compared group-level results with within-person, trial-by-trial dynamics. In some cases, the relationship flipped—what looked true in the average wasn’t true inside an individual. Why it matters: it’s a reminder that “one chart for everyone” can be misleading in neuroscience, especially when the goal is understanding behavior and mental health. The work also suggests conditions linked to inhibitory control—like ADHD—may involve multiple underlying pathways, which argues for more tailored interventions rather than a single story built from averages. Neanderthal bone-grease “fat factory” And now for a deep history datapoint: a study reports that Neanderthals in central Germany were rendering bone grease at scale around 125,000 years ago. The evidence points to a deliberate, repeated processing operation—bones smashed into massive quantities of fragments and heated to extract fat—at what appears to be a centralized lakeside work site. Why it matters: this pushes sophisticated food processing and coordinated labor further back in time than many people assume. It also adds weight to the picture of Neanderthals as strategic planners who managed resources across landscapes, not just opportunistic hunters. Building usable maps on Watch Finally, a developer story from the small-screen frontier: one engineer described a multi-year effort to build a genuinely usable map experience on Apple Watch. Early approaches leaned on server-generated images, which meant latency and no offline capability. Over time, the project shifted toward on-device rendering for speed and reliability, and a lot of the work was simply interface design—making panning, zooming, and metrics readable on a tiny, one-handed display. Why it matters: wearables keep getting marketed as “mini phones,” but they succeed or fail on interaction design. This story is a good reminder that performance, offline behavior, and UI constraints are inseparable—and sometimes the default platform tools aren’t flexible enough for demanding outdoor navigation use cases. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: macOS VMs near-native speed - Tests on Apple silicon with macOS 26.4.1 “Tahoe” show virtualization can hit near-native CPU and GPU performance, but Neural Engine results lag, shaping AI workload choices. Why Windows has TMP and TEMP - Raymond Chen traces TMP vs TEMP to MS-DOS history and Windows API quirks, explaining why compatibility keeps both environment variables alive across apps. Dashboards-as-code with semantic layer - Bruin Data’s open-source DAC brings “dashboard-as-code” using YAML and TSX plus a reusable semantic layer for metrics, improving reviewability, consistency, and SQL generation. Common Lisp targeting .NET runtime - Dotcl compiles ANSI Common Lisp to .NET CIL, enabling cross-platform Lisp images and tight interoperability with modern .NET frameworks like ASP.NET Core and MAUI. Self-hosted private daily journaling - Piruetas is a minimalist, privacy-first journal that’s self-hostable, telemetry-free by design, and focused on data portability with exports and shareable links. - Tests show macOS VMs on Apple silicon are near-native speed, but need enough disk to update - Why Windows Has Both TMP and TEMP Variables—and Why Neither Is Definitively Correct - Bruin Data releases DAC, an open-source dashboard-as-code tool with semantic layer and AI-assisted authoring - MLJAR Studio Promotes Fully Local, Private AI for Data Analysis and ML Workflows - Texas Instruments launches TI-84 Evo with faster performance and redesigned interface - SimplePDF launches Copilot demo for chatting with and filling PDF forms - dotcl Updates Bring Common Lisp to .NET with Cross-Platform CIL Compilation - Piruetas launches as a minimalist, privacy-first journal with export and self-hosting Episode Transcript macOS VMs near-native speed First up, a reality check—in a good way—for anyone who relies on macOS virtualization on Apple silicon. The Eclectic Light Company revisited VM performance using macOS 26.4.1, “Tahoe,” on an M4 Pro Mac mini, and the results are surprisingly close to native in the places people actually feel day to day. CPU single-core performance in a VM landed in the high nineties relative to the host, and GPU Metal performance was similarly close. In plain terms, that suggests a lot of everyday macOS work in a VM—apps, UI responsiveness, many dev workflows—can feel a lot like running on bare metal. The catch is AI acceleration. The virtualized Neural Engine performed dramatically worse than the host in tests that target that hardware. The implication is pretty practical: if you’re doing ML tasks inside a macOS VM, you may end up leaning more on CPU and GPU paths than expecting the Neural Engine to bail you out. There was also a useful note for anyone trying to squeeze this onto smaller storage: the limiting factor isn’t so much CPU or RAM as disk. The testing suggested macOS VMs need on the order of tens of gigabytes—around 60 GB—to update reliably, even if sparse files can soften the real-world footprint. That makes “small Mac as a VM host” less ridiculous than it used to sound, but storage planning still matters. Why Windows has TMP and TEMP Sticking with the theme of “why is it like this,” Microsoft veteran Raymond Chen dug into a small Windows oddity that has refused to die: the fact that Windows still has both TMP and TEMP for temporary-file locations. The short version is that it’s not a design choice so much as accumulated history. Early microcomputer software didn’t have a strong convention, different developers picked different names, and then the platform itself made choices that didn’t fully align. Even Windows APIs contributed to the split by preferring one in certain calls and the other elsewhere. Why it matters today is that “set one variable and you’re done” isn’t always true. Some programs follow TMP, others follow TEMP, and some follow their own logic entirely. Keeping both is basically a compatibility tax—and a reminder that once an ecosystem ships widely, tiny inconsistencies can become permanent infrastructure. Dashboards-as-code with semantic layer On the data tooling front, Bruin Data released an open-source project called DAC, positioned as “dashboard-as-code.” The central idea is to treat dashboards more like software: defined in code, validated, reviewed, and deployed in a repeatable way. The interesting twist is that it’s not just about putting layouts in YAML. DAC also ships with a semantic layer, so teams can define metrics and dimensions once and reuse them across widgets, instead of rewriting slightly different logic in every chart. That’s the kind of discipline that tends to reduce “why don’t these two dashboards agree” arguments. They also lean into AI-assisted iteration, where an agent can help edit and evolve dashboards while keeping the output standardized in code. If that works well in practice, it could make dashboards more maintainable without turning them into a free-for-all of one-off edits. One note people will care about: the project documents telemetry as anonymous command usage metadata, with an opt-out. Whether that’s acceptable depends on your environment, but at least it’s stated plainly. Common Lisp targeting .NET runtime For language and runtime enthusiasts, the dotcl project posted updates to its open-source repository, continuing its push to compile Common Lisp into .NET’s Common Intermediate Language. This matters because it’s not just a novelty compiler; it’s aimed at practical interoperability. If you can write Lisp that runs across Windows, macOS, and Linux under the .NET runtime—and also call into modern .NET libraries—you can mix expressive Lisp code with mainstream tooling where it helps. Projects like this are often about expanding your “reuse surface area.” Instead of porting everything or rewriting in the trendy language of the day, you potentially bring a mature language into today’s frameworks and deployment patterns, including web apps and cross-platform UI stacks. Self-hosted private daily journaling And finally, a small but telling entry in the ongoing pushback against bloated apps: Piruetas, a minimalist journaling web app built around one entry per day, with privacy and self-hosting at the center. The pitch is intentionally simple: keep “today” front and center, avoid telemetry, support basic formatting and attachments, and make it easy to export your writing so you’re not trapped. It also supports sharing individual days via links, which is useful if you want selective sharing without publishing your whole journal. Why it’s interesting is that it reflects a broader trend: people increasingly want tools that do less, store locally or under their control, and make data portability a default instead of an upgrade. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: PyPI lightning supply-chain malware - A supply-chain compromise hit the PyPI package "lightning" (PyTorch Lightning), with credential-stealing malware that can leak secrets from dev machines and CI. Keywords: PyPI, supply chain, malware, tokens, CI security. Linux CopyFail backport dilemma - The Linux kernel "CopyFail" local privilege escalation fix is tricky to backport to older long-term branches, leaving many systems waiting or relying on mitigations. Keywords: Linux kernel, LPE, CVE, backport, mitigation. Room 641A and NSA spying - EFF recounts how AT&T whistleblower evidence pointed to backbone-level internet traffic copying in a secret room, shaping the modern debate on mass surveillance and legality. Keywords: NSA, AT&T, EFF, mass surveillance, Patriot Act. Rethinking GitHub-style code forges - A critique argues modern forges overfit the GitHub model, and proposes workflows with earlier feedback, richer review states, and better offline-first collaboration. Keywords: GitHub, GitLab, forge, PRs, CI workflow. OpenWarp brings your own AI - OpenWarp, a community fork of Warp, aims to make terminal AI provider-agnostic so users can choose their own models and endpoints with a privacy-first posture. Keywords: terminal, AI, BYOP, privacy, open source. USB-C cable truth on macOS - WhatCable is a macOS menu bar tool that translates USB-C capabilities into plain language, helping diagnose slow charging and mismatched cables. Keywords: USB-C, Thunderbolt, charging, macOS, diagnostics. Fixing Bluetooth MIDI on Windows - A new Windows utility bridges Bluetooth LE MIDI devices into Windows MIDI Services so keyboards reliably appear in traditional DAWs and Web MIDI apps. Keywords: Windows 11, BLE MIDI, DAW, interoperability, MIDI ports. Websites derailed by stakeholder taste - A web design essay explains how leadership “taste edits” can slowly override research, turning a site into an internal mood board instead of a tool that converts users. Keywords: UX, research, stakeholders, conversions, usability. Lost Caedmon’s Hymn manuscript found - Researchers uncovered an early ninth-century manuscript containing Caedmon’s Hymn embedded in the main text, strengthening evidence that Old English was actively valued and copied. Keywords: Caedmon’s Hymn, Old English, manuscript, Bede, discovery. - Websmith Studio: Why Your Website Should Serve Users, Not Leadership Tastes - Open-source utility bridges Bluetooth LE MIDI into Windows MIDI Services for DAWs - WhatCable for macOS reveals the real capabilities of USB-C cables and charging setup - AT&T Whistleblower Exposed NSA Backbone Surveillance via Secret Room 641A - xAI Releases Grok-4.3 API Model Documentation with 1M-Token Context and Tooling Features - Ninth-Century Rome Manuscript Reveals Rare Early Copy of Caedmon’s Hymn - Kernel CopyFail (CVE-2026-31431) Fix Doesn’t Cleanly Backport to Older LTS, Workaround Shared - Author Proposes a Modular, Offline-Friendly Replacement for Modern GitHub-Style Forges - OpenWarp Fork Lets Warp Users Plug In Custom AI Providers and Keep Keys Local - PyTorch Lightning PyPI Package Compromised, Malware Steals Secrets and Spreads via npm Episode Transcript PyPI lightning supply-chain malware First up in security: researchers are warning about a supply-chain compromise of the PyPI package “lightning,” better known to many as PyTorch Lightning. Two recent versions were published with malicious code that can run simply through normal install-and-import behavior, aiming to siphon off secrets from developer machines and CI—think repo tokens, environment variables, and cloud credentials. What makes this one especially concerning is the attempted cross-ecosystem spread: the campaign doesn’t just want to steal—it wants to propagate, using whatever publishing credentials it can find to hop into other package registries and workflows. If your org builds AI models or training pipelines, this is a sharp reminder that “just a library update” can become an incident across your entire automation stack. Linux CopyFail backport dilemma Staying with security, there’s a tense discussion around the Linux kernel vulnerability dubbed “CopyFail.” The fix landed upstream, but backporting it cleanly to older long-term kernels is proving difficult, which is where the real-world pain shows up. Distributions and operators often rely on long-lived branches specifically for stability, yet those same branches can be the hardest places to land a complicated security change safely. In the meantime, maintainers are sharing mitigations—pragmatic stopgaps that reduce risk, but aren’t the same as a proper patch. The takeaway: if you run older kernel lines at scale, plan for a window where mitigations and careful configuration matter as much as updates do. Room 641A and NSA spying And now, a reminder that the security debates of today have long roots. Cindy Cohn revisits the moment in 2006 when retired AT&T technician Mark Klein walked into EFF with documents describing what appeared to be mass, untargeted internet surveillance—centered on a secret room at an AT&T facility where backbone traffic could be copied. The significance isn’t just the history lesson. It’s how concrete evidence changed the conversation: from rumors about surveillance to something that could be argued in court, contested in public, and scrutinized by experts. It also highlights a recurring tension: when national security secrecy collides with constitutional challenges, simply getting claims heard can become its own battle. Rethinking GitHub-style code forges Switching gears to software development workflows: one essay argues that code forges—GitHub, GitLab, and the rest—have converged on a familiar shape that doesn’t reflect how many teams actually work anymore. The point is less “Git is broken” and more “the forge has become the product”: reviews, CI, identity, issues, releases, automation—those are the center of gravity. The proposal is basically a next-generation forge that tightens feedback loops so quality checks happen earlier, supports more nuanced review states than a binary approve-or-reject, and treats stacked changes as first-class. Why it matters: as teams push for faster shipping with more automation, the ergonomics of collaboration tools start to influence both code quality and developer sanity. OpenWarp brings your own AI On the tooling side, an open-source fork called OpenWarp is trying to make AI inside the terminal more flexible by letting users bring their own AI provider. Instead of locking you into one model or one endpoint, the idea is a provider-agnostic layer where you control where prompts go and what they cost. This matters for two reasons. One is privacy and compliance—teams increasingly need to decide where data is sent. The other is resilience: model availability, pricing changes, and product decisions can shift quickly, and developers are looking for setups that don’t collapse when a single vendor changes the rules. USB-C cable truth on macOS If you’ve ever stared at a pile of identical USB‑C cables and guessed wrong, there’s a new macOS menu bar app called WhatCable that tries to end the guessing. It translates what macOS can see about a connected cable, charger, or device into plain English—so you can tell whether you’re actually getting fast data, video output, or the charging speed you expected. The broader point here is that USB‑C’s “one connector for everything” promise has a usability tax: when performance varies wildly but looks the same, troubleshooting becomes a time sink. Tools that turn invisible negotiation into readable explanations save real time, especially for people hopping between docks, monitors, and travel chargers. Fixing Bluetooth MIDI on Windows Windows musicians also got a practical win: a developer released a free, open-source utility aimed at making Bluetooth LE MIDI keyboards behave reliably with DAWs and Web MIDI apps. The core problem on Windows 11 is that devices can pair successfully, yet not show up in the places musicians expect—because different apps rely on different MIDI stacks. By bridging BLE MIDI into the newer Windows MIDI Services layer and presenting it like a normal MIDI port, this tool tries to turn a fragile workaround into a predictable setup. The interesting footnote: the author also uncovered a silent “nothing works but nothing errors” failure mode tied to unexpected MIDI channel behavior—exactly the kind of issue that wastes hours when the system gives you zero feedback. Websites derailed by stakeholder taste Now for the human side of tech: a Websmith Studio piece argues that company websites often fail for a surprisingly simple reason—leaders treat the site like personal expression instead of a user-focused tool. Designers come with research, testing, and evidence, but decisions get overridden by preferences about colors, layout, or what “feels right.” The author calls it an expert paradox: in high-stakes domains, stakeholders defer to specialists, but in web design, familiarity creates overconfidence. The slow damage comes from tiny compromises—each one seemingly harmless—until the site becomes polished, expensive, and quietly worse at helping customers complete tasks. If you own a website roadmap, the practical filter is strong: is feedback improving the user’s outcome, or just reflecting internal taste? Lost Caedmon’s Hymn manuscript found Finally, a discovery that stretches well beyond tech—but speaks to preservation, digitization, and scholarship. Trinity College Dublin researchers found an early ninth-century manuscript in Rome containing Caedmon’s Hymn, often cited as the earliest known poem in English. What’s notable is how the Old English appears: not as a later scribble in the margins, but embedded into the main Latin text tradition. That strengthens the case that early readers actively valued and transmitted Old English material, even when most surviving records are sparse. It’s also a reminder that “lost” doesn’t always mean destroyed—sometimes it means miscataloged, moved, or waiting for the right combination of careful bibliography and modern digitization. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: OpenAI’s “goblin” metaphor bug - OpenAI traced a surprising GPT style quirk—“goblins” and “gremlins” metaphors—to reward-model incentives and personality prompts, highlighting AI auditing and RLHF risks. Meta smart-glasses privacy controversy - Meta’s split with Sama in Kenya after claims of reviewing intimate Ray-Ban smart-glasses videos is driving privacy, consent, and outsourced AI labor scrutiny, with UK and Kenya regulators involved. IBM Granite open-source enterprise models - IBM released Granite 4.1 under Apache 2.0, arguing data curation and training discipline can make smaller enterprise LLMs competitive while reducing deployment cost and complexity. Web Prompt API standards fight - Mozilla pushed back on a proposed browser Prompt API that would let websites prompt built-in models, warning about interoperability, vendor lock-in, and policy-driven fragmentation of the web. Zed editor hits version 1.0 - Zed reached 1.0 after years of development, betting on a Rust-built, GPU-accelerated editor and AI-assisted workflows as developer tools shift toward collaborative AI coding. Zig’s strict anti-LLM policy - Zig’s ban on LLM-generated contributions is reshaping downstream collaboration—Bun reports big performance gains but says it won’t upstream—raising questions about trust, maintenance, and community health. Belgium rethinks nuclear phase-out - Belgium plans to halt nuclear decommissioning and negotiate potential nationalization with ENGIE, making nuclear a central lever for energy security, prices, and reduced gas dependence. - Belgium halts nuclear decommissioning and opens talks to nationalize reactors - Meta Ends Kenya AI Contract After Workers Report Viewing Graphic Smart-Glasses Footage - BidProwl Aggregates Government Surplus Auctions Into One Search Platform - IBM Releases Granite 4.1 Open-Source Models, With 8B Variant Rivaling Larger Systems - Mozilla Opposes Proposed Web Prompt API Over Interoperability and Model Neutrality Concerns - OpenAI traced GPT’s ‘goblin’ metaphors to a rewarded Nerdy personality training signal - Zed Code Editor Reaches 1.0, Expands AI and Collaboration Focus - Zig Explains Its Strict Ban on LLM-Assisted Contributions Episode Transcript OpenAI’s “goblin” metaphor bug First up, a strange but revealing AI story from OpenAI. The company says newer GPT versions developed a noticeable habit of using “goblins,” “gremlins,” and similar creatures as metaphors—especially after GPT‑5.1, with another surge later. It wasn’t random: the spike was tightly linked to a “Nerdy” personality setting, and internal audits showed the reward model was effectively paying the system extra points for that style. What matters here isn’t the vocabulary—it’s the lesson. Tiny incentives in training can create sticky behavioral quirks that spread into broader data pipelines, and you don’t always notice until it’s already in production. OpenAI says it removed the personality, adjusted reward signals, and added filtering and mitigations, framing it as a case for stronger model-behavior auditing. Meta smart-glasses privacy controversy Staying in AI, Meta is facing tough questions after ending a major AI-training contract with outsourcing firm Sama in Kenya. Workers allege they were exposed to highly sensitive videos captured by Meta’s Ray-Ban smart glasses—content that reportedly included people in private situations. Sama says the termination puts over a thousand roles at risk, while Meta argues standards weren’t met, which Sama disputes. Labour advocates in Kenya claim it looks like retaliation for speaking out, and privacy regulators are circling: the UK’s data watchdog has contacted Meta, and Kenya’s data protection authority is investigating. The bigger issue is the collision of three realities: smart devices that can capture intimate moments, AI systems that still rely on human review at scale, and outsourced workforces that can end up absorbing the emotional and legal risk. This story adds pressure for clearer consent, stronger safeguards, and more transparency about when “AI training” really means people watching the raw footage. IBM Granite open-source enterprise models On the model front, IBM released Granite 4.1—an Apache 2.0–licensed family of enterprise language models. The headline claim is that an 8B dense model can match or beat what used to require a significantly larger mixture-of-experts setup, across a broad set of evaluations. IBM’s angle is that the win isn’t a magic new architecture so much as a disciplined training pipeline: careful data mixing, aggressive filtering, and post-training that tries to improve chat behavior without torpedoing other skills. Why it matters: this is part of a growing push to make smaller models more dependable for real-world use—cheaper to run, easier to govern, and simpler to deploy inside companies that care about predictability and licensing. If these results hold up in practice, it strengthens the case that “bigger” isn’t the only path to better. Web Prompt API standards fight Now to the browser world, where the fight is brewing over a proposed web “Prompt API.” Google’s Blink team has signaled plans to prototype an API that lets web pages send prompts to a built-in browser model. Mozilla is openly skeptical and has staked out a negative position in its standards review process. The concern is familiar to anyone who remembers the bad old days of web compatibility: if developers start writing sites around the quirks and policies of a particular browser’s model, we could slide into a new kind of vendor-specific web—this time driven by AI behavior instead of HTML differences. There’s also a thorny policy angle: if model-provider usage rules seep into the API surface, sites might start detecting or blocking “unknown” models to manage compliance risk. The meta-point is that AI inside browsers isn’t just a feature—it’s a power shift. Standards bodies are being asked to bless an ecosystem before we even agree on what interoperability should mean for model-backed features. Zed editor hits version 1.0 Developer tools got a notable milestone too: Zed has hit version 1.0 after years of rapid releases. The team’s pitch is straightforward—build a modern editor with performance as a first principle, using Rust and a GPU-driven UI rather than a web-based shell. They also lean into an “AI-native” direction, with features aimed at working alongside multiple coding agents. The reason this is interesting now is timing: editors are turning into coordination hubs, not just text boxes. Whether you love AI coding assistants or tolerate them, the next wave of tooling competition is about integrating them without turning the editor into a slow, fragile stack of plugins. Zig’s strict anti-LLM policy And that transitions nicely into a culture-and-maintenance debate in open source. The Zig project enforces an unusually strict rule: no LLM-generated content in issues, pull requests, or even bug-tracker comments. That policy is having ripple effects. Bun—a JavaScript runtime built on Zig and now owned by Anthropic—says it achieved a big compile-performance improvement in its Zig fork, but doesn’t plan to upstream the work because of Zig’s prohibition. Supporters of Zig’s stance argue that reviews aren’t just about code—they’re about building trusted contributors, and AI-generated patches can waste scarce maintainer time without growing human expertise. Critics counter that it may slow collaboration and strand improvements in forks. Either way, it’s a clear signal that open-source communities are still negotiating what “contribution” means in an AI-assisted era: is the goal maximum throughput, or long-term maintainability and trust? Belgium rethinks nuclear phase-out Finally, a major energy-policy shift in Belgium. The government says it will halt the decommissioning of nuclear plants and is moving toward negotiations with ENGIE about potentially nationalizing the country’s nuclear assets—including reactors, staff, and long-term obligations like dismantling. Belgium’s nuclear phase-out was set in motion decades ago, but energy security concerns and political change have kept pushing the timeline around, and parliament has already voted to end the phase-out. The stakes are practical: Belgium remains heavily exposed to gas imports, renewables have not scaled fast enough to fully compensate, and electricity price stability is politically sensitive. If Belgium does take more direct control over nuclear generation—and pursues new nuclear build as signaled—it could reshape its supply risk and bargaining power in a volatile European energy market. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: ChatGPT ads and conversion tracking - A researcher observed how ChatGPT ads are injected into responses and how OpenAI-style attribution can follow users onto merchant sites via tokens, SDK calls, and cookies—raising privacy and measurement questions. Governments self-hosting open source forges - The Netherlands soft-launched code.overheid.nl on Forgejo as a self-hosted, government-wide Git platform, signaling digital sovereignty, reduced vendor dependence, and a push for shared public-sector collaboration. Open-source projects leaving GitHub - Ghostty’s maintainer Mitchell Hashimoto says GitHub outages and flaky Actions have become frequent enough to block releases, prompting an incremental migration away while keeping a read-only mirror. Decentralized code hosting with Radicle - HardenedBSD began moving repositories to Radicle, betting on more resilient, decentralized distribution and less reliance on a single forge—while acknowledging early rough edges for large repos. Rust security audit of uutils - An Ubuntu-driven audit found dozens of CVEs in uutils, showing that Rust can prevent memory bugs but still leaves teams exposed to filesystem semantics, TOCTOU races, and compatibility pitfalls. Rip.so memorializes dead internet services - Rip.so is a hand-built “digital graveyard” archiving once-major services like ICQ and Google Reader, highlighting how quickly online culture disappears and why preservation needs community effort. Tindie downtime shakes maker marketplace - New owners of Tindie apologized for prolonged outages and weak communication during a migration, a reminder that marketplace reliability and timely payouts are existential for independent hardware sellers. - Dutch government soft-launches code.overheid.nl open-source development platform - Ghostty Project Plans Exit From GitHub Over Frequent Outages - Rip.so launches as a ‘digital graveyard’ memorializing defunct internet services - Canonical audit of Rust coreutils highlights security bugs beyond Rust’s compiler checks - Mitchell Hashimoto to Move Ghostty Off GitHub Over Frequent Outages - Stardex seeks first Founding Customer Success Lead to scale support for recruiting firms - HardenedBSD Begins Migrating Code Repositories to Radicle - Researcher Maps ChatGPT Ad Delivery and Merchant Attribution Tracking Loop - Armin Ronacher Warns GitHub’s Decline Threatens Open Source Memory - New Tindie Owners Apologize for Weeks-Long Downtime, Promise Stabilization and Payout Fixes Episode Transcript ChatGPT ads and conversion tracking First up: a security researcher reports observing how OpenAI’s ads appear inside ChatGPT, and how conversions may be attributed after you click. The key detail is that the ad payload seems to be delivered inline while the model is streaming its response, and clicks can open in an in-app webview. On the merchant side, the researcher saw a tracking flow that can set first-party cookies and send event data back to OpenAI endpoints for measurement. Why it matters: even without proving whether chat history is used, this outlines a full measurement loop—chat context to ad impression to click to downstream activity. For users, it’s a reminder that “inside the chat” can still connect to “across the web.” For sites and regulators, it raises the familiar questions of consent, transparency, and what a reasonable boundary looks like when the interface is a conversation. Governments self-hosting open source forges Now to code hosting and digital sovereignty, starting with the Netherlands. The Dutch government has soft-launched code.overheid.nl as a central place to publish and develop open-source software across government. It’s fully self-hosted and runs on Forgejo, positioning it as a practical move away from relying on commercial code-hosting providers. Access is limited during the pilot, and the team is explicitly asking developers to help shape the platform. Why it matters: government software isn’t just code, it’s infrastructure and accountability. A self-hosted forge can improve control over policy, security posture, and long-term availability—especially when public bodies need to collaborate without being at the mercy of external platform changes. Open-source projects leaving GitHub That theme—dependence on a single platform—keeps coming up, because a prominent maintainer is walking away from GitHub. Mitchell Hashimoto, maintainer of the Ghostty terminal emulator, says the project will leave GitHub after nearly two decades of him using it daily. He describes the decision as personally painful, but says reliability has degraded to the point that outages routinely block the work Ghostty depends on: pull request review, issues, and CI through GitHub Actions. He emphasized this isn’t about Git itself—it’s about the collaboration layer around it. The plan is to migrate incrementally, keep a read-only GitHub mirror, and announce a new primary home after evaluating both commercial and open-source options. Why it matters: GitHub has become a default dependency for open source—not just hosting, but the entire workflow. When that workflow becomes flaky, maintainers pay the price in lost time and delayed releases, and projects start reconsidering what “single point of failure” really means. Decentralized code hosting with Radicle A related reflection comes from Armin Ronacher, who looks back at the pre-GitHub era and asks what we lose if we fragment again. His argument is that GitHub didn’t just centralize code—it centralized context: issues, reviews, design debates, and the public record that helps outsiders understand why software is the way it is. He warns that if the ecosystem disperses back into many forges and self-hosted instances, we may regain autonomy but also recreate an older problem: broken links, missing artifacts, and vanished project history. His proposed direction is refreshingly unglamorous: a stable, well-funded public archive for open source that preserves not just repos, but releases and surrounding metadata. Why it matters: software supply chains increasingly rely on provenance and discussion trails. If those trails disappear, trust gets harder—especially for security reviews, compliance, and long-term maintenance. Rust security audit of uutils If you want a glimpse of what “post-centralized forge” could look like, HardenedBSD just moved its code hosting to Radicle. The project says Radicle is now usable for them, rough edges included, and they’ve published initial repositories with plans to migrate the rest over time. The headline isn’t the tooling details—it’s the intent: making the project less dependent on one company’s uptime and policies, and more resilient through decentralized distribution. Why it matters: the migration pressure we’re seeing—whether from outages, governance worries, or supply-chain concerns—is pushing projects to experiment. Even imperfect alternatives become attractive when the status quo feels increasingly fragile. Rip.so memorializes dead internet services Switching to security: there’s a sharp lesson in a new write-up about Rust, uutils, and what “memory safe” does—and doesn’t—buy you. Rust consultant Matthias Endler reviewed Canonical’s disclosure of 44 security CVEs found in uutils, the Rust reimplementation of GNU coreutils, following an external audit ahead of Ubuntu 26.04 LTS. The takeaway is that many issues weren’t classic memory corruption problems. They were about filesystem reality: race conditions between checking a path and acting on it, permission timing mistakes, mismatches between string comparisons and actual filesystem identity, and subtle behavior differences compared to GNU coreutils that can become dangerous in widely scripted tools. Why it matters: Rust reduces an entire class of vulnerabilities, but security still depends on semantics, OS boundaries, and compatibility expectations. If a tool is “drop-in” for scripts, behavior differences can turn into security problems—even if the code is perfectly well-behaved from a memory standpoint. Tindie downtime shakes maker marketplace For something lighter, but still surprisingly meaningful: Rip.so has launched as a “digital graveyard” for dead or hollowed-out internet services. It’s a hand-coded memorial site with epitaph-style entries for platforms and technologies that once dominated—think ICQ, Internet Explorer, Vine, GeoCities, Google Reader, and more. It even leans into early-web community rituals with a guestbook and a webring. Why it matters: digital culture disappears fast, often without durable public records. Projects like this may look nostalgic, but they also function as grassroots archiving—documenting what vanished, when it vanished, and how quickly the internet reinvents itself by forgetting. Story 8 Finally, a reliability story from the hardware world: Tindie’s new owners apologized for prolonged downtime and poor communication during an ownership transition. They say the site is now owned by EETree LLC and that the outage came from migrating an older stack with lots of interconnected services. Their stated priority is stabilizing the marketplace and resolving payment, refund, and order issues through support, while promising more updates going forward. The community response has included understandable skepticism, especially around transparency and seller payouts. Why it matters: for independent hardware makers, marketplaces aren’t just storefronts—they’re cash flow. Extended downtime and unclear payouts can push sellers to leave permanently, and trust is hard to rebuild once it’s shaken. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Vintage AI trained pre-1931 - Researchers released “talkie,” a 13B vintage language model trained only on pre-1931 English, raising new questions about data provenance, anachronisms, and evaluation contamination in AI. OpenAI cloud exclusivity ends - Microsoft and OpenAI revised their partnership to end Microsoft’s exclusive resale rights, reshaping cloud competition as AWS and others may now distribute OpenAI models. AI data centers’ gas emissions - A WIRED review of air permits suggests behind-the-meter natural-gas plants for AI data centers could enable massive CO2 emissions, challenging Big Tech climate pledges and grid strategy. Social Edge Paradox in AI - An essay argues heavy AI automation can weaken human apprenticeship and debate, risking creativity loss and “model collapse” as AI-generated text feeds future training data. ASML EUV lithography chokepoint - ASML’s early bet on EUV lithography turned a Dutch spinoff into the key bottleneck for leading-edge chips, with major geopolitical implications for semiconductor supply chains. GTFOBins and misconfig risk - GTFOBins catalogs how everyday Unix tools can be abused when systems are misconfigured, helping defenders audit sudo/SUID permissions and “living off the land” attack paths. WebAssembly: stack vs register - A developer argues WebAssembly isn’t truly a classic stack machine in practice, which matters for how compilers express optimizations and how humans reason about hand-written Wasm. Satellites streaking comet photos - NASA’s APOD shows Comet C/2025 R3 (PanSTARRS) obscured by satellite trails, highlighting how dense satellite constellations are increasingly affecting astronomy and astrophotography. More RAM made Quake slower - A retro PC benchmark mystery showed adding RAM reduced Quake performance due to chipset cacheability limits—an old-school reminder that specs don’t always match real behavior. - How ASML’s EUV Bet Made It the Chokepoint for Advanced Chips - Talkie Debuts as a 13B ‘Vintage’ Language Model Trained Only on Pre-1931 Text - Microsoft and OpenAI End Exclusive Model-Sales Deal, Drop Revenue Sharing - APOD Photo Shows Comet C/2025 R3 PanSTARRS Hidden Among Satellite Trails - GTFOBins catalog highlights how common Unix tools can be abused on misconfigured systems - WIRED: Gas-Fueled AI Data Centers Could Add Nation-Scale Emissions - Why WebAssembly Isn’t Really a Stack Machine - Essay Warns AI Automation Could Undermine the Social Foundations of Intelligence - Why Adding More RAM Slowed a 1997 Quake PC Episode Transcript Vintage AI trained pre-1931 First up, a big shift in the AI business landscape: Microsoft and OpenAI are dialing back one of the most consequential terms of the AI boom—Microsoft’s exclusive right to sell OpenAI’s models. That exclusivity helped make Azure the default home for “the hot models,” because if you wanted OpenAI at scale, you typically came to Microsoft. Under the updated arrangement, OpenAI can pursue distribution with other cloud providers, potentially including AWS. Microsoft, meanwhile, stops paying a revenue share on OpenAI products it resells through its cloud platform. The significance here isn’t just who sells what—it’s that AI model access is becoming more like a competitive market than a single-lane highway. For buyers, that could mean more options and better pricing pressure. For cloud vendors, it’s a reminder that model partnerships are powerful—until they aren’t exclusive anymore. OpenAI cloud exclusivity ends Sticking with AI, one of the more unusual research releases today is a “vintage” language model called talkie. It’s a 13B-parameter model trained only on English text published before 1931, designed to simulate a pre-modern-knowledge assistant. Why is that interesting? Because modern models are soaked in the web—memes, post-2000 facts, and a lot of repeated benchmark material—so it’s getting harder to tell whether a model is reasoning or just recalling. A historical model gives researchers a cleaner lens: you can ask, “What does the model do when it can’t possibly have seen the modern answer?” The team says talkie lags behind a comparable web-trained “modern twin” on many knowledge questions, as you’d expect, but the gap narrows when you remove anachronistic prompts. It also surfaces very practical hurdles, like OCR noise and “temporal leakage,” where post-1930 facts sneak into datasets. The bigger takeaway: data curation isn’t just paperwork—it shapes what a model can be, and what we can honestly claim it learned. AI data centers’ gas emissions Now, zooming out from models to the real-world footprint powering them: a WIRED review of state air-permit documents points to a surge in natural-gas projects tied to a small set of US data center campuses serving major AI players. The headline number is huge—permits that could allow emissions on the scale of entire countries. A key detail is the strategy: “behind-the-meter” generation. Instead of waiting years for grid interconnections, developers put gas turbines next to the data centers and largely bypass the grid. The argument is speed and predictability—AI workloads don’t like power uncertainty. The counterpoint is that this can lock in a new wave of fossil infrastructure right when many expected gas and coal to shrink. Even if permitted emissions are upper bounds, the concern from analysts is simple: data centers are steady, always-on demand, which can keep those turbines running hard. For Big Tech, it’s a credibility stress test for climate commitments. For everyone else, it’s a preview of how AI can reshape energy policy by sheer urgency. Social Edge Paradox in AI That pairs uncomfortably well with an essay making the rounds about what it calls the “Social Edge Paradox.” The idea: today’s AI looks smart partly because it reflects a vast record of human collaboration—debate, apprenticeship, institutional review, and all the messy social friction that produces robust knowledge. But if companies deploy AI in a way that reduces those human processes—cutting entry-level roles, shrinking mentorship, automating away the back-and-forth—then you weaken the very pipeline that creates future expertise and future high-quality training data. The essay also points to research suggesting AI can lift individual productivity while making group output more uniform. In other words, you may get faster drafts but fewer distinct viewpoints. Tie that to warnings about “model collapse,” where AI-generated text floods the ecosystem and future models learn from a blur of averaged content, and you get a strategic question: are organizations optimizing for short-term efficiency at the expense of long-term originality? The practical takeaway is surprisingly managerial, not technical: if you want durable advantage, use AI to amplify high-interaction work, not replace it. ASML EUV lithography chokepoint On the hardware side, there’s a detailed look at how ASML—once a scrappy Philips spinoff—became the chokepoint for manufacturing the most advanced semiconductors. The core bet was extreme ultraviolet, or EUV, lithography: turning a brutally hard physics problem into the only viable path for leading-edge chipmaking. What matters isn’t the gadgetry of EUV, impressive as it is. It’s the compounding effect of being the company that solved it while rivals exited. ASML benefited from long-running research ecosystems, deep collaboration with places like IMEC, and a critical moment in the early 2010s when EUV commercialization wobbled—only for ASML to raise money by selling stakes to the very customers who needed EUV to exist, and to pull key suppliers closer. The result is an effective monopoly that now influences the pace of compute progress and the geopolitics around who can build top-tier chips. If you’re watching export controls, supply-chain resilience, or the next leap in AI hardware, this is one of the central leverage points. GTFOBins and misconfig risk In security, a reminder that some of the most dangerous “exploits” aren’t vulnerabilities at all—they’re features used in the wrong context. GTFOBins is a curated reference of common Unix-like executables that can be abused to bypass local restrictions on misconfigured systems. The value here is defensive clarity. It’s easy to think, “This binary is harmless.” GTFOBins shows how ordinary tools can become stepping stones when sudo rules are too broad, SUID bits are set casually, or capabilities are handed out without a threat model. For blue teams, it’s a practical checklist for auditing what’s allowed and under what permissions. For attackers, it’s a playbook for “living off the land.” The important point is the same either way: configuration is security-critical, and convenience defaults can quietly become escalation paths. WebAssembly: stack vs register For developers, there’s an interesting argument about WebAssembly: people often call Wasm a “stack machine,” but if you actually try to write Wasm by hand, that label can mislead you. The claim is that classic stack-based VMs tend to offer richer stack manipulation, so you can reuse values and restructure computations without naming temporary storage. Wasm, in practice, pushes you toward introducing locals for non-trivial reuse, which makes it feel closer to a register machine with a stack-like encoding style. Why should you care? Because the mental model affects everything from how you reason about performance to how you expect compiler optimizations to show up in the output. If your intuition about “it’s just a stack VM” is off, you’ll be surprised by what’s awkward—or what becomes more verbose—when you leave the comfort of higher-level languages. Satellites streaking comet photos On the science and imagery side, NASA’s Astronomy Picture of the Day featured a long-exposure shot of Comet C/2025 R3, with a dense web of satellite trails cutting across the scene. To the naked eye, satellites can look like slow-moving points near twilight. But with a multi-minute exposure, they become bright streaks that can overwhelm faint celestial targets. It’s a vivid illustration of an ongoing tension: satellite constellations are transforming connectivity on Earth while complicating observing conditions in the sky. And in this particular case, the comet itself is also hard to spot right now because it’s close to the Sun from our viewpoint—though it’s expected to be better placed for viewing soon from the Southern Hemisphere. The broader point is that “light pollution” increasingly includes objects in orbit, not just city glow. More RAM made Quake slower Finally, a fun retrocomputing lesson with a real performance punchline: a Quake enthusiast upgraded a 1997-era PC to a hefty 384 MiB of RAM—because the modules were cheap—and then watched frame rates drop dramatically. After swapping parts, reinstalling software, and chasing ghosts, the fix was simply removing RAM. The culprit was an old-school limitation: the motherboard chipset could only cache a certain amount of main memory. Anything above that behaved like “uncached” RAM, slowing things down—especially because the operating system tends to allocate from higher memory addresses. It’s a great reminder that on older systems, “more” can be “less,” and that real-world constraints don’t always match what the spec sheet promised. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: PostgreSQL backups face uncertainty - pgBackRest, a widely used PostgreSQL backup/restore tool, is now obsolete and unmaintained—raising risk around security fixes, compatibility, and long-term support for production backups. Cassandra storage engine modernization - Apache Cassandra 5 brings storage-engine upgrades like trie-based structures and newer compaction work, aiming for better efficiency and more predictable operations at scale. On-device AI in Chrome - Chrome’s Prompt API expands access to on-device Gemini Nano for web pages and extensions, highlighting the shift toward local AI features with new governance, permissions, and hardware constraints. Quantizing vectors for cheaper AI - TurboQuant proposes compressing embeddings and KV caches to a few bits with a shared rotation and a reusable codebook, but debate continues around bias, benchmarking, and lineage versus prior work like EDEN. AI and engineering judgment gap - A growing concern in software teams: AI can remove drudgery, but it can also encourage “outsourced thinking,” weakening debugging instincts, risk awareness, and real technical judgment—especially for juniors. Docs screenshots that self-update - A documentation workflow auto-regenerates screenshots via headless browser runs driven by directives in Markdown, reducing stale help-center images and keeping docs aligned with UI changes. Cheap DIY USB audio DSP - DSPi turns inexpensive Raspberry Pi Pico-class boards into a plug-and-play USB audio interface with built-in DSP, making room correction and active audio setups more accessible to hobbyists. Authenticity fights: brands and social - Friendster’s domain revival experiments with real-world friend connections, while Moleskine’s AI-disclosure backlash shows how trust can erode when generative content labeling is inconsistent. - pgBackRest PostgreSQL Backup Tool Declared Unmaintained and Obsolete - DSPi firmware turns Raspberry Pi Pico boards into a multi-output USB audio DSP - flipdisc.io - Friendster Domain Bought and Relaunched as an In-Person-First iOS Social App - Moleskine Faces Backlash Over AI-Generated Imagery in Lord of the Rings Notebook Launch - Interactive Walkthrough Details TurboQuant’s Random-Rotation Quantization for 2–4 Bit AI Vectors - Blog Warns AI Can Create ‘Outsourced Thinking’ in Software Engineering - Jelly Automates Self-Updating Screenshots for Its Help Center - Chrome’s Prompt API expands on-device Gemini Nano prompting with sessions, multimodal input, and origin trials - Cassandra PMC Member Branimir Lambov on Tries, Compaction, and What’s Next for Cassandra Episode Transcript PostgreSQL backups face uncertainty First up, databases and reliability. The pgBackRest project—one of the most common tools used for PostgreSQL backup and restore—has been declared obsolete and is no longer maintained by its longtime author. The reasoning is blunt: years of personal time, corporate sponsorship drying up after a business change, and no sustainable path to keep up with reviews, bug fixes, and support. Instead of letting quality slowly degrade, the maintainer chose a clean stop. Why it matters is simple: backups are only “boring” until the day you need them. If your production backup chain depends on pgBackRest, you’re now betting on forks, frozen compatibility, or your own ability to take maintenance in-house. The repo explicitly says forks may happen, but they’ll need a new name and, more importantly, new trust. Cassandra storage engine modernization Staying with data infrastructure, an IBM engineer and long-time Cassandra committer shared progress on modernizing Cassandra’s storage engine. The headline isn’t a shiny feature so much as steady, unglamorous work: reducing memory overhead, improving how data structures behave under load, and making compaction more resilient. Cassandra 5’s direction here is about fewer operational surprises—especially in the messy reality of dense datasets and long-lived clusters. One detail worth noticing: even when the “main path” is solid, secondary code paths can still threaten correctness, and the post calls out a bug where certain access patterns made data appear missing. It’s a reminder that reliability work isn’t just performance tuning—it’s continuously hunting down edge cases that only show up at scale. On-device AI in Chrome Now to AI on the client side. Google’s Chrome team has updated documentation for the Prompt API, which lets web pages and Chrome Extensions send prompts to an on-device Gemini Nano model. This is being tested through Origin Trials, and it’s clearly part of a broader push: making local, in-browser AI feel like a standard platform capability rather than a series of cloud calls. The big why: local AI can mean lower latency, fewer privacy tradeoffs, and less dependence on external APIs. But Chrome is also drawing boundaries—who can embed it, how permissions flow, and what devices can realistically run it. In other words, it’s a glimpse of the next web platform debate: AI features as “native browser primitives,” with all the security and governance questions that implies. Quantizing vectors for cheaper AI On the research-and-systems side of AI, there’s an interactive walkthrough of TurboQuant, a method aimed at heavily compressing high-dimensional vectors—think embeddings and LLM attention caches—down to just a few bits per coordinate. The interesting twist is the claim that you can get near-optimal compression without needing per-block metadata or calibration training, by transforming vectors so they fit a predictable distribution and then using a reusable codebook. That’s attractive because it promises simpler pipelines and cheaper memory use. But the discussion doesn’t stop at the pitch: it also highlights a practical problem—certain “best by the math” quantization choices can bias results by shrinking vectors and distorting inner products, which matters for attention and search. And there’s some academic friction too, with comparisons to prior work like EDEN and questions about how reproducible some benchmarks really are. Net-net: promising ideas, but the community is still arguing about what’s truly new and what’s truly better. AI and engineering judgment gap One of today’s more human stories is a blog arguing that AI is splitting software engineers into two camps. In one camp, engineers use AI to clear out repetitive work so they can focus on system design, tradeoffs, and risk. In the other, people use AI to bypass thinking entirely—submitting fluent output without building understanding. The author’s core point is that judgment can’t be outsourced: AI can speed you up, but it can’t magically give you the instincts you only get from wrestling with problems. This matters for teams because it changes what “good” looks like. Hiring and performance reviews can’t just reward polished deliverables—they have to detect real comprehension, especially in debugging, incident response, and architecture decisions where confidence and correctness are not the same thing. Docs screenshots that self-update From philosophy back to practical craft: a developer described a neat way to keep documentation screenshots from going stale. Instead of manually capturing images every time the UI changes, their docs include small embedded directives that tell an automated tool exactly what to load and what to capture. A headless browser logs in, navigates to the right states, and regenerates screenshots in a single run. Why it matters: outdated screenshots quietly degrade product trust. This approach turns screenshot maintenance into something that can ship alongside code changes, so the help center evolves with the UI rather than lagging behind it. Cheap DIY USB audio DSP Let’s switch to hardware and maker energy. WeebLabs released DSPi, an open-source firmware project that turns inexpensive Raspberry Pi Pico-class boards into a USB audio interface with an onboard DSP pipeline. The point isn’t audiophile bragging rights—it’s making practical speaker and headphone tuning accessible: EQ, mixing, time alignment, loudness tweaks, and dynamics control, all with low latency. What’s compelling here is the value curve. The barrier to experimenting with room correction or active crossover setups has often been cost or complexity. Projects like this compress that gap: cheap board, plug into USB, and you have a playground for real audio processing ideas. Authenticity fights: brands and social And in a similar spirit—hardware as art—someone shared a build of a large interactive wall display using flipdiscs. Flipdiscs are those satisfying mechanical pixels that physically flip between two colors with a little click. The author leaned into what they’re good at: readability, durability, and that unmistakable sound. They paired a grid of panels with custom control electronics and modern software for visuals and interaction, then treated the wall like a living canvas—news feeds, music-inspired scenes, and more. Why it matters isn’t that everyone needs a flipdisc wall. It’s that we’re seeing a renewed appetite for “tangible computing”—displays that feel physical, legible at a distance, and built to last, especially as so much of our daily UI becomes abstract and disposable. Story 9 Finally, a quick two-part note on trust and authenticity online. First, the domain friendster.com has reportedly been acquired and revived in a new form. The experiment is intentionally anti-algorithm: an iOS app where adding friends is tied to physically tapping phones, nudging relationships toward real-world contact rather than endless online accumulation. Whether it succeeds is an open question—but it’s an interesting attempt to redesign social networking around fewer, higher-signal connections. Second, Moleskine caught backlash after promotional images for a Lord of the Rings notebook collection appeared to include an “generated by AI” disclaimer—while other marketing posts didn’t clearly disclose AI use, and some visuals looked suspiciously low-quality. The larger issue is transparency: when brands mix human design with AI-generated or AI-assisted imagery without consistent labeling, it becomes hard for customers to know what they’re buying into—and that uncertainty can damage trust faster than the tech can improve workflows. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: AI nudges a math breakthrough - A 23-year-old used GPT-5.4 Pro to spark a novel approach to a classic Erdős conjecture on primitive sets, with experts verifying and rewriting the messy AI proof. Keywords: Erdős conjecture, primitive sets, GPT, proof verification, Terence Tao. Statecharts tame complex software behavior - Statecharts.dev argues statecharts extend finite state machines to avoid state explosion, improve testability, and model exceptions more clearly, while noting adoption hurdles and the role of SCXML semantics. Keywords: statecharts, finite state machine, SCXML, executable diagrams, software modeling. AI efficiency vs engineering know-how - A defense-manufacturing analogy warns that heavy AI reliance and reduced junior hiring could erode tacit software expertise, creating a future “Fogbank for code” where critical debugging skill disappears. Keywords: tacit knowledge, junior hiring, AI coding tools, resilience, skills pipeline. Scan-to-playable browser FPS pipeline - PlayCanvas shows how Gaussian Splat scans can become a playable web FPS by generating collision, streaming-friendly assets, lighting coherence, navmesh pathing, and lightweight NPC AI—making scan-to-game more practical. Keywords: Gaussian splats, PlayCanvas, browser FPS, collision mesh, navmesh. iOS apps reappearing after deletion - Hacker News users report the Headspace app reappearing on iPhones after deletion, raising questions about iOS restore/offload behavior, App Store logic, and transparency around installs. Keywords: iOS bug, unwanted reinstall, App Store, privacy, diagnostics. USB naming chaos finally clarified - Fabien Sanglard’s USB cheat sheet cuts through confusing branding by mapping what cable/port labels typically mean in practice, helping developers avoid time-wasting configuration mistakes. Keywords: USB-C, cable compatibility, bandwidth reality, naming confusion, troubleshooting. GnuPG update and 2.4 EOL - GnuPG 2.5.19 ships with fixes and small features, highlights post-quantum work in the 2.5 line, and warns that GnuPG 2.4 is nearing end-of-life—prompting upgrades. Keywords: GnuPG 2.5.19, security, OpenPGP, post-quantum, end-of-life. Floating-point explained without myths - Bartosz Ciechanowski explains IEEE 754 floating-point as constrained base-2 scientific notation, showing why decimals like 0.2 don’t round-trip cleanly and why output formats matter for correctness. Keywords: IEEE 754, rounding, NaN, subnormal, numeric bugs. - Statecharts.dev Introduces Statecharts and Their Case for Modeling Complex Software - Amateur’s ChatGPT Prompt Leads to New Proof of 60-Year-Old Erdős Conjecture - Defense Production Failures Offer a Warning About AI and the Future of Software Talent - Kanban Task Management Service OOKO Updates on npm to v0.121.0 - PlayCanvas Demo Shows How to Turn a Gaussian Splat Scan into a Playable Browser FPS - iPhone Users Report Headspace App Reinstalling Itself After Deletion - Engineer Publishes USB Naming and Speed Cheat Sheet to Cut Through Branding Confusion - GnuPG 2.5.19 Released With New Options, Bug Fixes, and 2.4 End-of-Life Reminder - A Visual Guide to IEEE 754 Floating-Point: Encoding, Special Values, and Precision Limits Episode Transcript AI nudges a math breakthrough A surprising AI-assisted moment in pure math: Liam Price, a 23-year-old amateur, reported a solution to a long-standing Erdős conjecture after prompting GPT-5.4 Pro and then posting the result to the Erdős Problems site. The conjecture is about “primitive sets” of integers—collections where no number divides another—and how a classic scoring function, the Erdős sum, behaves as the numbers get huge. What’s grabbing mathematicians isn’t that an AI typed up a proof; it’s that the model appears to have made an unconventional connection, pulling in a known formula from a neighboring area that researchers simply hadn’t tried in this context. Experts, including Terence Tao, describe the raw AI output as messy and not publication-ready, but with human reconstruction the core insight looks sound and has been streamlined. Why it matters: it’s a real example of AI acting like a cross-domain suggestion engine—useful for novel direction-finding—while human experts still do the heavy lifting on verification and clear exposition. Statecharts tame complex software behavior Staying with the human side of expertise, one essay made a sharp comparison between defense manufacturing and software engineering. The author points to cases like the struggle to restart Stinger missile production—where retirees, obsolete parts, and lost shop-floor knowledge turned “just spend money” into “wait years.” They also cite other examples where optimized peacetime supply chains couldn’t surge when the world changed. The warning for software is pointed: if companies lean hard on AI to generate code while also cutting junior hiring, they may be thinning the pipeline that produces future senior engineers. The claim isn’t that AI is useless—it’s that organizations can accidentally trade away tacit skills like debugging judgment, systems thinking, and technical leadership, because those are built through years of real incidents and mentorship. The takeaway: resilience is the theme. If the industry optimizes only for short-term throughput, it risks a future where key knowledge quietly disappears until there’s a crisis and nobody left who knows how to “make it work.” AI efficiency vs engineering know-how On the topic of building reliable systems, Statecharts.dev published an accessible explainer arguing that statecharts are essentially “finite state machines, but grown up” for real software. The article frames statecharts as a visual formalism introduced by David Harel in 1987 to address the classic pain of big state machines: state explosion, where behavior becomes an unmanageable web of cases. The practical pitch is team-oriented—clearer behavior modeling, better separation between behavior and components, more straightforward testing, and a cleaner way to represent exceptional paths without turning the whole design into spaghetti. It also doesn’t ignore the downsides: there’s a learning curve, it’s a mindset shift, and for small systems it can feel like extra overhead. The interesting connective tissue is SCXML, a W3C standard that aims to pin down semantics so tools and libraries can handle edge cases consistently. And the article nods to “executable statecharts,” where the same definition drives runtime behavior and generates diagrams—reducing translation mistakes, while raising familiar concerns about tooling maturity and type-safety. Why it matters: if your product’s logic is getting harder to reason about, this is one of the few modeling approaches that tries to make complexity visible and testable rather than hidden in conditional code. Scan-to-playable browser FPS pipeline Now for something more visual—almost sci-fi, but very practical. PlayCanvas shared a walkthrough that turns a photorealistic Gaussian Splat scan into a playable browser FPS scene. Splats look great, but they’re typically just renderable geometry without solid surfaces for gameplay. The demo stitches together an end-to-end pipeline that makes the scan act like a real level: it streams progressively so devices don’t stall, generates collision so physics and shooting behave correctly, and adds navigation data so NPCs can move through the space. One clever bit is handling lighting: splats often have baked appearance that doesn’t naturally light regular game assets, so the demo approximates local brightness to keep characters and weapons visually consistent with the scanned environment. The big deal here is feasibility—this is a credible “scan-to-playable” workflow using open tooling and a public project others can fork. If this approach keeps improving, it lowers the barrier to building believable game spaces, simulations, or training environments directly from the real world. iOS apps reappearing after deletion A smaller story, but one that hits a nerve for platform trust: a Hacker News thread reports the Headspace app repeatedly reappearing on some iPhones after users delete it—even with automatic downloads disabled. Several commenters describe seeing the icon return daily, sometimes greyed out as if it’s pending download, often after iOS updates. The discussion leans toward this being an Apple-side regression—something in restore behavior, offloading logic, App Store server decisions, or device sync—rather than anything Headspace would intentionally attempt, since that would be incredibly conspicuous and risky. Why it matters: deleting an app is a basic user expectation. If the system can override that without a clear audit trail, it raises questions about control, privacy, and accountability. The recurring request from users is simple: better diagnostics and clearer logs that explain what triggered an install and why. USB naming chaos finally clarified If you’ve ever lost an afternoon to cables and ports, you’ll appreciate this one. Fabien Sanglard published a USB cheat sheet after chasing a bug that turned out to be USB terminology confusion. Rather than drowning you in spec history, the page tries to map marketing names to what you can expect in real-world performance, and it highlights the gap between advertised signaling rates and actual throughput. It also clarifies why two cables that both look “USB-C” can behave completely differently, depending on what’s actually wired and supported. Why it matters: USB is a reliability trap—especially in labs, dev kits, and hardware-adjacent software work. A simple, trustworthy reference can prevent misdiagnosis, bad purchasing decisions, and those maddening “it works on my machine” hardware inconsistencies. GnuPG update and 2.4 EOL In security and privacy news, Werner Koch announced GnuPG 2.5.19. The release is a mix of fixes and small features across core components, plus ongoing work in the 2.5 line that’s especially focused on better support for 64-bit Windows and modern cryptography directions, including post-quantum algorithms. The headline for many users is lifecycle: GnuPG 2.4 is expected to hit end-of-life in about two months, and the maintainer is urging people to move to 2.5.19 while keeping compatibility with older setups. Why it matters: crypto tools tend to live at the bottom of dependency stacks, and upgrades get postponed until something breaks. An explicit EOL clock is a good reason to plan migrations before they become emergency work. Floating-point explained without myths Finally, a classic topic done exceptionally well: Bartosz Ciechanowski published a deep-but-readable explanation of IEEE 754 floating-point. The framing is simple: floats are base-2 scientific notation with strict limits on precision and exponent range, and those constraints create the “surprises” developers trip over—like why common decimals can’t be represented exactly, why rounding errors accumulate, and why some ranges of numbers have big gaps between representable values. The article also covers the special values—zeros, infinities, NaNs, and subnormals—and ties it back to practical debugging, including how the way we print numbers can hide what’s truly stored. Why it matters: floating-point misunderstandings show up as flaky tests, finance bugs, graphics artifacts, and incorrect comparisons. Better mental models here pay dividends across almost every domain of software. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Quantum crypto demo debunked - A GitHub patch shows a quantum ECDLP “key recovery” demo still works when quantum output is replaced with /dev/urandom, highlighting how classical verification plus enough shots can mimic progress. Google’s conditional Anthropic investment - Google is reportedly committing $10B now, with up to $40B total tied to performance targets, underscoring milestone-based AI funding, compute access, and the soaring cost of frontier models. FPS counters that tell truth - A game-dev write-up argues for a rolling time-window FPS metric, using frame timestamps to avoid noisy per-frame spikes and misleading moving averages—better observability for real-time performance. Plain-text diagram tools resurgence - ASCII-style mockups and plain-text diagrams are gaining traction because they’re durable, version-control friendly, and constraint-driven—useful for humans and increasingly legible to AI workflows. USB-C 10 GbE reality check - New Realtek RTL8159 USB-C 10 GbE adapters run cooler and smaller than many Thunderbolt dongles, but real-world throughput depends heavily on confusing USB port bandwidth and driver support. Humpback whale super-groups surge - Scientists saw unusually large humpback whale feeding aggregations off South Africa—hundreds of whales—signaling conservation recovery and shifting ocean conditions in the Benguela upwelling system. Francis Bacon and tech ideology - A 400th-anniversary essay revisits Francis Bacon’s legacy, linking “knowledge is power” to modern technocracy, scientism critiques, and today’s AI-era drive to instrumentalize culture. Paraloid B-72 in conservation - Paraloid B-72, an acrylic resin used by conservators, stands out for clarity and aging stability; its solvent-dependent handling shows why materials science matters in restoration and museums. - Realtek RTL8159 USB-C 10GbE adapters shrink size and heat but depend on 20Gbps USB ports - Google Plans Up to $40 Billion Investment in Anthropic, Tied to Performance Targets - Use a Rolling One-Second Window for Accurate FPS Counters - Clad.you Builds 3D Body Models from an Eight-Question Survey Using a Physics-Aware MLP - Nex CRM open-sources WUPHF, a Slack-like office for coordinating AI agents with shared memory - Essay Argues Francis Bacon’s Legacy Fueled Modern Technocracy—and Today’s AI Moment - Humpback Whale ‘Super-Groups’ Swell Off South Africa as Populations Rebound - Plain-Text Diagramming Tools Make a Modern Comeback - Paraloid B-72: A widely used acrylic resin adhesive in conservation and restoration - Patch Shows Alleged Quantum ECDLP Key Recovery Works With /dev/urandom Instead of IBM QPU Episode Transcript Quantum crypto demo debunked First up, a reality check on a quantum-themed claim making the rounds. A GitHub document describes a patch to a quantum ECDLP key-recovery demo that replaces the IBM Quantum backend with random bytes from /dev/urandom. And here’s the kicker: the rest of the pipeline still “recovers” the same small challenge keys, and even succeeds on the headline larger challenge at a similar rate. The critique argues the demo is effectively generating random candidate keys and then using a classical check to accept a winner when enough attempts are made. Why it matters: it’s a reminder that verification can create the illusion of signal, and that we need rigorous baselines—especially when “quantum advantage” is the headline. Google’s conditional Anthropic investment Staying in the world of high-stakes computation, Google is reportedly planning to invest up to forty billion dollars in Anthropic, with ten billion committed immediately and the rest conditional on performance targets. Alongside the money, there’s also support for expanding compute capacity—still the chokepoint for training and serving top-tier AI models. The interesting angle here isn’t just the size; it’s the structure. Big tech is increasingly buying optionality: pay now for access, then pay more only if milestones are hit. Why it matters: it reshapes the competitive landscape into one where capital, chips, and contracts matter as much as research talent—and it signals that AI’s price tag keeps climbing. FPS counters that tell truth Now for a developer-facing piece that’s deceptively practical: how to build an FPS counter that doesn’t lie to you. The post argues that common approaches either jitter wildly or create distorted history because the averaging window effectively changes as performance changes. The recommended fix is to treat frames as events and compute FPS over a fixed rolling time window, so the number stays responsive without becoming noise. Why it matters: good metrics drive good decisions. Whether you’re optimizing a game, a UI, or a real-time system, measurement that’s stable and interpretable saves time—and prevents you from chasing phantom regressions. Plain-text diagram tools resurgence In a related “tools that fit the brain” theme, there’s an essay noting the small resurgence of plain-text diagramming and UI mockup tools—think ASCII-style layouts that live right alongside code. The argument is that constraints are a feature, not a bug: fewer visual choices can mean faster communication, easier version control, and more durable artifacts. There’s also an AI-era twist: plain text is easy to paste, diff, search, and feed into models without losing structure. Why it matters: as software teams juggle more automation, the formats that stay simple and portable often win—especially when they reduce friction in collaboration. USB-C 10 GbE reality check Let’s switch to hardware—specifically, the ongoing headache of “do I actually have the right USB port?” New USB-C 10 GbE adapters based on Realtek’s RTL8159 are showing up as smaller and cooler alternatives to chunky, hot Thunderbolt 10 GbE dongles. Jeff Geerling tested one across Macs and PCs and found the big limiter isn’t the adapter so much as the host port: near-full 10 GbE speeds showed up only on a machine with a fast enough USB 3.2 configuration, while many systems topped out well below that. Macs were largely plug-and-play, Windows needed a driver, and thermals looked notably improved. Why it matters: these adapters can be great—if your laptop’s port can actually feed them. The messy USB labeling problem is still very real for buyers. Humpback whale super-groups surge On the science and environment side, photographers off South Africa’s west coast documented unusually large “super-groups” of humpback whales—hundreds of animals packed together over just two days. Researchers link these gatherings to seasonal upwelling in the Benguela system, which concentrates food near the surface and triggers intense feeding behavior. The surge in sightings may also reflect a broader recovery since the global whaling moratorium, with lots of young whales showing up in identification data. Why it matters: it’s a conservation success story, but also a signal that ecosystems are changing. Recovery doesn’t mean risk-free—ship strikes, entanglement, noise, and warming waters still shape what comes next. Francis Bacon and tech ideology For something more reflective, an essay marking the four-hundredth anniversary of Francis Bacon’s death argues that Bacon didn’t just help inspire modern science—he also helped cement an ideology where knowledge becomes power and nature becomes something to dominate. The author threads this through a historical parable involving wealth, technological faith, and the Titanic, and then jumps forward to modern critiques of “only measurable things matter.” It even frames today’s AI moment as a continuation of that drive to turn culture and knowledge into instrument and leverage. Why it matters: these aren’t abstract debates anymore. How we talk about progress shapes what we build, what we fund, and what we ignore. Paraloid B-72 in conservation And finally, a surprising materials-science detour into museums and restoration: Paraloid B-72, an acrylic resin that’s become a staple for conservators. It’s valued for staying clear, resisting yellowing, and offering a balance of strength and flexibility—useful for ceramics, glass, fossils, and even labeling objects. The practical catch is that how it behaves depends heavily on solvents and handling, which turns “just glue it” into careful craft. Why it matters: preservation is engineering too. The long tail of technology isn’t only new inventions—it’s the quiet chemistry that keeps artifacts, specimens, and cultural memory intact. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: UK Biobank data on sale - UK Biobank says detailed volunteer health records were spotted for sale on Alibaba, raising urgent questions about data leakage, re-identification risk, and research trust. AI fake photo derails search - South Korean police arrested a man for spreading an AI-generated wolf photo that diverted an emergency search—an example of synthetic misinformation disrupting public safety. How distrust fuels polarization - A satirical 'anti-social guide' spotlights confirmation bias, defensiveness, and distrust—behaviors that accelerate polarization and degrade everyday communication. Ruby compiled to native binaries - Spinel compiles Ruby ahead-of-time into standalone native executables, showing how trading some dynamic features can yield major speed and deployment benefits. Tarballs as WebAssembly filesystems - A new approach mounts a .tar.gz as a virtual filesystem in the browser via WebAssembly, cutting memory overhead and load times for package-heavy apps like WebR. How LLMs are actually built - A visual explainer walks through LLM training, post-training, hallucinations, and RAG—useful context for anyone deploying AI assistants and evaluating reliability. Orwell on motives for writing - George Orwell’s 'Why I Write' outlines ego, aesthetics, history, and politics as core writing motives, connecting personal experience to craft and public purpose. - UK Biobank Says Health Data on 500,000 Volunteers Was Listed for Sale on Alibaba - South Korea arrests man over AI-generated photo that misled search for escaped zoo wolf - Satirical Guide Highlights Behaviors That Fuel Social Isolation and Conflict - DeepSeek Docs Outline OpenAI/Anthropic-Compatible API and July 2026 Model Deprecations - Spinel: Self-Hosting Ruby AOT Compiler Generates Standalone Native Binaries - ‘Endless Toil’ Plugin Adds Groaning Audio Feedback for Codex and Claude Coding Agents - Nowhere pitches websites encoded in URL fragments to resist hosting-based censorship - Indexing tar.gz files to mount them as a zero-copy filesystem in WebAssembly - Visual Guide Explains How Large Language Models Are Trained, Aligned, and Grounded - Orwell Explains the Four Motives Behind His Writing and Turn to Political Art Episode Transcript UK Biobank data on sale First up: UK Biobank says it found medical data tied to its 500,000 volunteers being offered for sale on Alibaba. The concerning part isn’t just that “data was leaked” in the abstract—these listings reportedly included the kind of rich, multi-dimensional details that make modern health research powerful: demographics, lifestyle, cognitive measures, lab results, and coded health outcomes, including cancer diagnoses and dates. If you’re wondering why this hits so hard, it’s because UK Biobank is one of the most important datasets in medical research. When participants feel that data may escape controlled research channels, trust erodes. And without trust, recruitment drops, studies slow, and the public becomes less willing to support the kind of longitudinal research that drives better treatments and earlier detection. It’s also a reminder that “de-identified” can be fragile when a dataset is detailed enough to be re-identified or misused once it spreads. AI fake photo derails search Staying with the theme of information causing real-world consequences: South Korean police arrested a man accused of disrupting the search for a missing zoo wolf by circulating an AI-generated image that appeared to show the animal near a road intersection. The image spread quickly online, authorities reportedly shifted resources, and residents even received an emergency warning—before investigators concluded the photo was fake after checking CCTV and usage records. The wolf was eventually captured, but the case puts a bright spotlight on something emergency management is now forced to treat as routine: synthetic media can be convincing enough to redirect public resources, amplify fear, and muddy situational awareness exactly when clarity matters most. How distrust fuels polarization One of the more reflective reads making the rounds today is a deliberately cynical “guide” to being anti-social—basically a checklist of how to escalate conflict and mistrust in everyday life. It urges people to assume bad intent, treat gut feelings as unquestionable facts, and recruit friends to reinforce a preferred narrative instead of testing assumptions. The point, of course, is critique. It’s a mirror held up to patterns many of us recognize online and offline: confirmation bias, defensiveness, and the tendency to interpret disagreement as an attack. In a world where algorithms already reward outrage, the reminder is timely—healthy communities rely on a little humility, a little patience, and a willingness to revise your story when the evidence changes. Ruby compiled to native binaries On the developer side, an interesting experiment in language performance: Spinel is an ahead-of-time compiler that turns Ruby into standalone native executables. The headline claim is big speedups compared to traditional Ruby execution, especially on compute-heavy workloads. What’s notable here is the tradeoff. Ruby is famously dynamic, and Spinel gets much of its performance by narrowing the set of features it supports—favoring code that can be analyzed and optimized ahead of time. That’s a recurring pattern in programming languages right now: many teams want the ergonomics of a high-level language, but they also want predictable deployment and performance. Projects like this test how much “dynamic magic” people are willing to give up to get there. Tarballs as WebAssembly filesystems Another clever engineering idea: using a .tar.gz archive directly as a virtual filesystem in WebAssembly—without the usual extract-and-copy step. Instead of unpacking everything into memory, the system indexes where files live inside the archive and reads just the bytes it needs on demand. Why care? Because in the browser, memory and startup time are precious. If you’re running serious tooling—like data science environments or package-heavy apps—avoiding a big up-front extraction can make the difference between “this feels instant” and “this feels broken.” It’s a good example of how small shifts in packaging and I/O strategy can unlock much smoother experiences for complex web apps. How LLMs are actually built If you’ve ever wished for a clean, non-mystical explanation of how LLMs get built, there’s a visual end-to-end walkthrough gaining attention today. It traces the pipeline from web-scale data collection and aggressive filtering, through tokenization, training a base model to predict text, and then post-training steps that shape it into something assistant-like. The practical value is that it clarifies why these systems behave the way they do—why hallucinations happen, why a “context window” isn’t long-term memory, and why knowledge can feel outdated. It also frames retrieval-augmented generation, or RAG, as a pragmatic fix: don’t just rely on the model’s internalized patterns—bring relevant documents into the conversation so answers can be grounded in current sources. Orwell on motives for writing Finally, a classic that still lands: George Orwell’s essay “Why I Write.” Orwell lays out writing motives—ego, aesthetic joy, a desire to record truth, and political purpose—and explains how his own experiences pushed him toward explicitly political work. It’s striking reading this alongside today’s debates about persuasion, propaganda, and platform-driven narratives. Orwell isn’t arguing that art must be political—he’s arguing that for him, the times made it unavoidable, and the craft mattered because it determined whether truth could compete with easy slogans. Even in 2026, that tension feels familiar. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: iOS notifications leaking deleted chats - Apple patched iOS so notification caches can’t retain deleted or disappearing message text from apps like Signal—closing a major privacy gap for seized devices. Firefox IndexedDB fingerprinting identifier - A Firefox IndexedDB bug created a stable cross-site fingerprint via `indexedDB.databases()` ordering (CVE-2026-6770), impacting Private Browsing and even Tor Browser until fixes land. Telecom SS7 and Diameter tracking - Citizen Lab reports covert location surveillance using telecom signaling abuse—SS7 and misconfigured Diameter—via “ghost” carriers and repeated transit providers. SQLite gets durable pubsub - Honker adds Postgres-style LISTEN/NOTIFY semantics to SQLite with durable queues and event streams, reducing reliance on separate brokers while keeping transactions atomic. Zig-built C compiler journey - A developer chronicled building a C compiler in Zig, offering a practical learning log that tracks progress from parsing to linking and highlights Zig’s growing systems-language ecosystem. Color-coded hex editing usability - A proposal argues hex editors should color bytes by default—like syntax highlighting—so humans can spot structure, anomalies, and boundaries faster in binary data. AI-shaped comment spam conversations - Bloggers are seeing AI-like comment spam that mimics real conversation threads, slipping links into plausible back-and-forth to bypass moderation and social trust cues. Cloud primitives and cost friction - A cloud critique claims today’s hyperscaler primitives push awkward constraints and high network costs; a new platform pitch argues AI coding agents will make that friction more painful. Repairable tractors without electronics - Ursa Ag is drawing interest for tractors built around remanufactured 1990s diesel engines with minimal electronics—betting farmers value repairability over locked-down software. - David Crawshaw Launches exe.dev to Rebuild Cloud Computing Primitives - Ursa Ag Bets on Electronics-Free Tractors to Win Over Repair-Weary Farmers - Honker Adds Postgres-Style NOTIFY/LISTEN, Queues, and Streams to SQLite via WAL-Based Push Notifications - Blog Post Calls for Default Byte Color-Coding in Hex Editors - Apple Patches iOS Bug That Let Forensic Tools Retrieve Deleted Messages From Notification Cache - Bloggers Warn of Conversation-Style Comment Spam Hiding Links - Jiga pitches AI-driven manufacturing sourcing platform and outlines remote-first, transparent culture amid hiring push - Citizen Lab Finds Covert Vendors Exploiting Telecom Signaling to Track Phone Locations - Developer Indexes a Zig-Based C Compiler Tutorial Series - Firefox and Tor Browser Bug Enabled Cross-Site Tracking via IndexedDB Result Ordering Episode Transcript iOS notifications leaking deleted chats First up: Apple has pushed an iPhone and iPad update to fix a privacy issue where notification content could linger on-device longer than users—and some apps—intended. The core problem was that iOS could retain notifications marked for deletion, meaning message text from apps like Signal might remain recoverable in a system database for weeks. The story matters because disappearing messages only work as well as the layers beneath them, and device seizures are exactly the scenario where those guarantees are supposed to hold. Firefox IndexedDB fingerprinting identifier Sticking with privacy, Mozilla fixed a subtle but serious fingerprinting vector in Firefox tied to IndexedDB. Researchers found that the order returned by the `indexedDB.databases()` API could act like a stable identifier across unrelated websites within the same browser process. That’s especially uncomfortable because it could persist in Private Browsing as long as the process stays running—and in Tor Browser, it could even undermine “New Identity” in that same session. The takeaway is a bit humbling: even something as mundane as result ordering can become a high-entropy tracking signal when it’s influenced by global internal state. Telecom SS7 and Diameter tracking And zooming out from browsers to networks: Citizen Lab says it uncovered two covert surveillance campaigns abusing telecom signaling systems to track phone locations. This is the long-running SS7 story, but with a modern twist—attackers can also exploit Diameter when carriers don’t enforce protections correctly, and some setups effectively fall back to SS7-like weakness. The report points to “ghost” companies posing as legitimate operators and to a few providers showing up repeatedly as entry or transit points. Why it matters is simple: location is one of the most sensitive data types, and the infrastructure that routes calls and texts still offers too many ways to quietly query it. SQLite gets durable pubsub On the developer tooling side, an experimental project called Honker is trying to give SQLite something it’s famously missing: Postgres-style NOTIFY and LISTEN, plus durable queues and event streams—without adding a separate broker. The interesting angle is that jobs and events live as rows in the same SQLite database, so publishing a message can commit atomically with your application data. For teams shipping single-machine apps—or embedded systems that still need background work—this could shrink operational complexity while keeping reliability characteristics people normally reach for Redis or a message bus to get. Zig-built C compiler journey If you like learning-by-building, there’s also a collected set of posts on writing a C compiler in Zig, called “paella.” It’s based on a well-known compiler-writing guide, but the value here is the journal-like progression: it’s the messy, practical record of getting from “hello, parser” to producing linkable outputs. These kinds of write-ups matter because they’re often the on-ramp for the next wave of systems programmers—especially as Zig keeps attracting folks who want low-level control without the full pain of older toolchains. Color-coded hex editing usability A smaller, but oddly compelling usability argument: one developer is making the case that hex editors and hexdump tools should use richer color by default. The point is that monochrome byte grids hide patterns your eyes are actually good at spotting—boundaries, repetition, or the single weird byte that shouldn’t be there. Think of it as syntax highlighting for binary data: a low-cost change that can make debugging files, formats, and corruption issues faster and less disorienting. AI-shaped comment spam conversations Now for the unglamorous reality of running a website in 2026: comment spam is evolving. One blogger described a new flavor that shows up as a short, believable conversation—multiple replies posted minutes apart—so it feels like genuine engagement. The trick was that a casino link was tucked into the middle comment in a way that blended into normal text. The broader point is that AI-generated “plausible filler” doesn’t need to be great to be effective; it just needs to look socially real long enough to slip past moderation and capture a click. Cloud primitives and cost friction On infrastructure, developer David Crawshaw published a sharp critique of modern cloud primitives—alongside news that he’s building a new cloud platform. His argument is that the big clouds are fundamentally shaped around awkward constraints: fixed VM instance types, storage trade-offs that push you toward remote services, and network pricing that turns data movement into a tax. He also predicts AI coding agents will increase the amount of software we produce, meaning cloud friction and cost won’t feel like an inconvenience—it’ll become a bottleneck. Whether or not his new approach wins, the critique resonates because it challenges a quiet assumption: that higher-level platforms can fully paper over the economics and ergonomics of the underlying cloud building blocks. Repairable tractors without electronics And finally, a very different kind of backlash against complexity: a Canadian startup, Ursa Ag, says it’s seeing strong interest from U.S. farmers for tractors built around remanufactured 1990s-era diesel engines—and intentionally light on modern electronics. The appeal is repairability: fewer locked-down components, less dependence on dealer software, and a better chance of fixing problems during the narrow windows that matter in planting and harvest. The bigger signal here is market pressure. If enough buyers prioritize maintainability over extra features, major manufacturers may have to rethink where “smart” becomes “fragile,” and where control becomes a liability. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Linux inside Windows 95 - WSL9x runs a modern Linux kernel alongside the Windows 9x kernel without virtualization, hinting at new life for legacy x86 and retrocomputing experimentation. GPS timing, relativity, accuracy - A GPS explainer shows how trilateration depends on precise timing, why a fourth satellite fixes clock drift, and how relativity prevents kilometer-scale navigation error. ChatGPT Images 2.0 upgrade - OpenAI’s ChatGPT Images 2.0 focuses on controllable layouts and reliable typography, pushing text-to-image toward practical design workflows like posters and infographics. Async AI agents need messaging - AI agents are moving from single chat sessions to long-running tasks; the debate shifts to durable state plus durable transport, with realtime push and reconnect built in. Workplace monitoring for AI training - Meta staff reportedly object to keystroke logging and screenshots for training computer-using AI, spotlighting employee privacy, consent, and reputational risk. MuJoCo physics simulation updates - DeepMind’s open-source MuJoCo keeps evolving with tooling and build improvements, reinforcing its role as a key robotics and reinforcement-learning simulation backbone. Smart contact lens for glaucoma - A polymer, electronics-free contact lens monitors intraocular pressure and releases glaucoma medication automatically, aiming to improve adherence and at-home care. Memory-safe garbage collection in Rust - The `safe-gc` Rust crate demonstrates garbage collection with zero unsafe code, trading some ergonomics for strong memory safety and easier auditing. Software engineering “laws” reference - A curated catalog of software engineering principles—like Conway’s Law and leaky abstractions—offers shared vocabulary for discussing scale, quality, and trade-offs. - Hailey releases WSL9x, a cooperative Linux kernel hack for Windows 95/98 without virtualization - Interactive Guide Explains How GPS Uses Timing, Trilateration, and Relativity - OpenAI Launches ChatGPT Images 2.0 With Improved Control, Typography, and Multilingual Rendering - Async AI Agents Expose the Limits of HTTP Chat Transports - Meta staff object to new workplace tool that logs keystrokes, mouse activity, and screenshots to train AI - DeepMind’s MuJoCo Physics Simulator Adds Ongoing Updates Across APIs and Platform Support - Electronics-Free Microfluidic Contact Lens Monitors Eye Pressure and Doses Glaucoma Drugs - Website Catalogs 56 Core “Laws” and Principles of Software Engineering - Nick Fitzgerald Releases a Fully Safe Rust Garbage Collector Crate Episode Transcript Linux inside Windows 95 Let’s start with that retrocomputing curveball: an experimental project called “Windows 9x Subsystem for Linux,” or WSL9x. The idea is to run a modern Linux kernel in a cooperative, ring-0 setup alongside the Windows 9x kernel—skipping hardware virtualization entirely. The author says it didn’t require weeks of spelunking through Windows 95 internals either; just leaning on a handful of VMM services for basics like threads and memory context. Why it matters: beyond the sheer hack value, it’s a reminder that “subsystems” don’t always need a hypervisor. If this direction holds up, it could reopen interesting paths for running meaningful Linux tooling in extremely constrained legacy environments—something we haven’t talked about much since the Cooperative Linux era. GPS timing, relativity, accuracy Staying in the “how computers see the world” lane, there’s a clear interactive explainer on GPS that’s worth your time, even if you think you already get it. It frames GPS as a time problem first: satellites send timestamps, your receiver turns time-of-flight into distance, and multiple satellites narrow that down into a real location. The twist that still surprises people is how central relativity is. Satellite clocks don’t tick at the same rate as clocks on Earth, and without correcting for those effects you’d rack up huge location errors fast—on the order of kilometers per day. The practical takeaway: modern positioning isn’t just clever geometry; it’s physics, statistics, and messy real-world signal problems like reflections in cities all layered together. ChatGPT Images 2.0 upgrade Now to AI on the creative side: OpenAI announced ChatGPT Images 2.0, pitching it as a step toward image generation you can actually direct—especially when text and layout are involved. The headline improvements are better prompt adherence, more dependable typography, and stronger multilingual text rendering. Why it matters: reliable text inside images is a big deal because it’s the difference between “cool demo” and “usable artifact.” If a model can consistently produce clean labels, panels, and structured compositions, you’re closer to end-to-end design output—things like posters, infographics, or editorial graphics—without a human spending half the time fixing broken lettering. Async AI agents need messaging Another AI theme today is less flashy but arguably more important: agents are shifting from chat sessions to long-running background work—triggered by schedules, webhooks, and remote control across devices. One post argues that the usual HTTP request plus streaming approach just doesn’t map to that world. The core point is simple: if the agent’s work outlives your connection, you need a communications layer that survives disconnects, supports push updates, and works when multiple people—or multiple devices—are involved. The author calls this missing piece “durable transport,” and positions it as the next step beyond just “durable state.” If you’re building agent products, this is a good lens for why so many prototypes feel brittle in real life. Workplace monitoring for AI training On the human side of AI development, Reuters and Business Insider report internal pushback at Meta over a new workplace monitoring tool. The initiative is said to record keystrokes and mouse movements, and to capture occasional screenshots during work app usage, with the stated goal of training AI systems to better understand how people use computers. Why it matters: this is exactly the kind of dataset that could accelerate “computer-using” agents, but it sits right on top of consent, privacy, and trust. Even if the intent is research, the optics are rough—and for Meta, privacy controversies are not a side story. Expect more companies to wrestle with this tension as agent ambitions collide with the reality that useful behavioral data often comes from watching humans work. MuJoCo physics simulation updates Switching gears to robotics and simulation: Google DeepMind’s MuJoCo continues to be actively maintained, with ongoing releases and improvements across tooling and platform builds. MuJoCo is one of those unglamorous foundations—fast, accurate physics simulation—that a lot of robotics and reinforcement learning work quietly depends on. Why it matters: when a simulator becomes a standard benchmark environment, improvements affect reproducibility and accessibility. Better builds and broader integrations mean more people can run the same experiments, compare results, and iterate faster—especially as more tooling moves toward browsers and lightweight environments. Smart contact lens for glaucoma In biomedical tech, researchers at the Terasaki Institute described an electronics-free smart contact lens aimed at glaucoma care. It monitors intraocular pressure and can automatically deliver medication when pressure rises, using microfluidic channels and pressure-triggered reservoirs. A smartphone app reads a visible indicator and uses a neural network to interpret it. Why it matters: glaucoma management has two stubborn problems—pressure checks are typically occasional, and adherence to daily drops is imperfect. A lens that both monitors and doses could reduce the gap between “what the treatment plan says” and “what happens at home.” It’s early, with rabbit testing and open questions like overnight monitoring and long-term comfort, but the direction is compelling: closed-loop therapy without packing electronics into your eye. Memory-safe garbage collection in Rust For programming language folks, Nick Fitzgerald introduced a proof-of-concept Rust garbage collector called `safe-gc` that forbids unsafe code entirely. The big design move is refusing to let you dereference GC pointers directly. Instead, you access objects through a heap handle, which keeps Rust’s borrowing rules in play. Why it matters: garbage collection in Rust usually forces someone to write unsafe code somewhere—either inside the library or in user-defined tracing. This project shows a different trade: you can keep memory safety airtight, even if users make mistakes, at the cost of some ergonomics and likely performance. It’s a practical example of Rust’s philosophy: you can often move problems from “memory corruption” into “explicit constraints and manageable failure modes.” Software engineering “laws” reference And finally, a lighter but useful item: a website curating dozens of widely cited “laws” and heuristics in software engineering—things like Conway’s Law, Brooks’s Law, leaky abstractions, and Goodhart’s Law. It’s organized as a reference and a shared vocabulary. Why it matters: these aren’t literal laws, but they’re recurring patterns. Having them in one place helps teams communicate about trade-offs—especially when the tricky parts aren’t just code, but planning, coordination, and the gap between how systems look on paper and how they behave under stress. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Apple CEO succession in 2026 - Apple announced Tim Cook will become executive chairman on Sept. 1, 2026, with hardware chief John Ternus taking over as CEO—major continuity signals for products, services, and regulation. EU DMA interoperability vs Apple - A FSFE report says Apple’s DMA interoperability request process has produced zero concrete new access so far, raising questions about enforcement, open standards, and developer choice on iOS. Catalog of software engineering laws - A new “Laws of Software Engineering” reference curates 56 principles like Conway’s Law, Brooks’s Law, CAP, DRY, and Goodhart’s Law—useful vocabulary for trade-offs and failure modes. Living with an open-hardware laptop - A personal long-term field report on the MNT Reform open-hardware laptop highlights repairability, real-world quirks, and community-driven fixes—what ownership looks like beyond the spec sheet. Flipper research on shelf labels - TagTinker is an open-source Flipper Zero app for authorized infrared ESL protocol research, lowering the barrier for analysis and replay testing while stressing legal and ethical boundaries. Tabletop dice pioneer Louis Zocchi - Game industry figure Louis Zocchi, founder of Gamescience and an early champion of polyhedral dice, has died at 91—his standards and distribution helped shape modern tabletop gaming. - Website Catalogs 56 Core “Laws” and Principles of Software Engineering - Apple Names John Ternus CEO as Tim Cook Becomes Executive Chairman - Codemix introduces @codemix/graph with type-safe traversals, Yjs sync, and Cypher queries - OpenClaw Updates Anthropic Integration with API Key, Claude CLI Reuse, Caching, and 1M Context Options - Owner Log Details MNT Reform Open-Hardware Laptop Issues, Mods, and Linux Workarounds - TagTinker brings infrared ESL protocol research tools to Flipper Zero - FSFE Report Says Apple’s DMA Interoperability Process Has Delivered No Solutions for Developers - Louis Zocchi, Pioneering Dice Designer and Gamescience Founder, Dies at 91 Episode Transcript Apple CEO succession in 2026 Apple is planning a major leadership transition effective September 1st, 2026. Tim Cook will move from CEO to executive chairman, while John Ternus—Apple’s senior VP of Hardware Engineering—will become the next CEO. Apple is presenting this as the outcome of long-term succession planning, with Cook staying on through the summer to help with the handover. Why it matters: CEO changes at Apple are rare, and they tend to shape expectations for years—even when the company says it’s “continuity.” Putting a hardware leader in the top job is also a statement about where Apple thinks its leverage remains: devices, silicon, and the tight integration that comes with them. And with regulators increasingly scrutinizing platform control, Apple also made a point of keeping Cook involved as executive chairman, including global policy engagement. EU DMA interoperability vs Apple Staying with Apple, a new report from Free Software Foundation Europe argues that Apple’s Digital Markets Act interoperability process isn’t delivering practical results. Using Apple’s own public tracker, the report says that as of late March, none of the formal interoperability requests had produced a new solution made available to developers. Many requests, it claims, were rejected as out of scope, too technical, or already covered by existing options. Why it matters: the DMA is supposed to make gatekeepers open up key platform capabilities so smaller developers can compete and users get real choice. If the process turns into paperwork with long timelines and ambiguous rejections, the law’s intent gets diluted—without a clear “no.” This also signals where the next phase may head: fewer polite requests, more enforcement actions, and more pressure for open standards rather than case-by-case exceptions. Catalog of software engineering laws On the software engineering side, a site called “Laws of Software Engineering” is making the rounds with a curated catalog of dozens of widely cited principles—everything from Conway’s and Brooks’s Laws to ideas like the Law of Leaky Abstractions, CAP, DRY, YAGNI, and Goodhart’s Law. The pitch is simple: these patterns keep showing up in real projects, and having them in one place gives teams a shared language for trade-offs. Why it matters: most software failures aren’t caused by one bad line of code—they’re caused by mismatched incentives, hidden complexity, and planning assumptions that don’t survive contact with reality. A reference like this is useful not because it’s “law” in the physics sense, but because it helps engineers and managers name the trap they’re walking into—like over-optimizing a metric, or assuming abstraction layers won’t leak at the worst moment. Living with an open-hardware laptop If you want a grounded look at the difference between “repairable” in theory and “repairable” in daily life, there’s also a long-running personal field report on the MNT Reform—an open-hardware laptop designed and assembled in Berlin. The author documents owning multiple units over several years, swapping devices, and even turning one into a community loaner. It’s full of practical details: physical wear issues, a trackball that can mark the screen when closed, and experiments like repositioning antennas to improve Wi‑Fi. Why it matters: open hardware often sells a promise—control, longevity, and the ability to fix what breaks. Reports like this show the real cost and real payoff: you may spend more time tuning and tinkering, but you also get a path to parts, modifications, and community knowledge that’s hard to find in sealed consumer laptops. Flipper research on shelf labels In security and hardware research news, an open-source project called TagTinker has been released for the Flipper Zero. It focuses on researching infrared electronic shelf-label protocols—observing signals, analyzing them, and doing controlled replay-style experiments on hardware you own or are authorized to test. The maintainer is explicit about boundaries, repeatedly warning against using it on deployed retail systems or anything that could alter pricing or interfere with operations. Why it matters: electronic shelf labels are everywhere, and they sit at an awkward intersection of retail operations and wireless security. Tools that make research easier can help defenders, auditors, and interoperability work—provided the culture around them stays responsible. This is the same story we’ve seen with many dual-use tools: the technology isn’t going away, so the best outcome is transparency, clear ethics, and better security design upstream. Tabletop dice pioneer Louis Zocchi Finally, a notable loss from the broader tech-adjacent world of hobby gaming: Louis Zocchi—often called the “Godfather of Dice”—has died at 91. Beyond game design and publishing work, his biggest legacy came through Gamescience, where he helped popularize polyhedral dice in the U.S. and pushed unusual formats that became part of tabletop culture. Why it matters: this is a reminder that “technology culture” isn’t only servers and code. Manufacturing, distribution, and standards—even for something as humble as dice—shape whole ecosystems. Zocchi’s impact is visible every time a tabletop game pulls out a bag of polyhedrals and treats it as normal. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Fake GitHub stars distort discovery - A new ICSE 2026 analysis estimates millions of suspected fake GitHub stars, pushing repos onto Trending and skewing discovery, funding, and VC sourcing signals. AI agents reshape SaaS moats - Figma is framed as the latest SaaS incumbent pressured by agent-first AI workflows, where LLM-generated assets erode collaboration-driven growth and weaken platform lock-in. Anthropic tools inside US government - Reuters cites reports that the NSA is using Anthropic’s Mythos tool despite supply-chain risk concerns, spotlighting the tension between rapid AI adoption and vendor governance. Vercel breach via OAuth chain - Vercel confirmed a security incident tied to a compromised Google Workspace OAuth app and third-party AI tooling, highlighting SaaS integration risk and secrets management pitfalls. Tokenizer shifts change AI costs - Claude Opus tokenizer changes can inflate token counts and effective spend, reminding developers that model updates can alter budgets even when sticker pricing stays the same. Listening failures in product teams - A product critique argues teams overuse frameworks to avoid real listening, leading to misread requirements, avoidable technical debt, and weaker customer outcomes. Microwave curing for printed electronics - Rice University demonstrated precise microwave-based curing of conductive inks on delicate surfaces, enabling new printed electronics use cases in medical and bio-integrated devices. Japan offshore earthquake update - A magnitude 7.4 quake off northeastern Japan underscores ongoing Japan Trench seismic hazard, with aftershocks and secondary risks remaining key concerns. - Investigation Finds a Growing Market for Fake GitHub Stars and VC Incentives Driving It - Magnitude 7.4 Offshore Earthquake Hits Near Miyako, Japan - Anthropic’s Claude Design Raises New Competitive Pressure on Figma - NSA Reportedly Using Anthropic’s Mythos Tool Despite Pentagon Supply-Chain Risk Label - Rice University’s focused-microwave tool enables 3D-printed circuits on delicate and living surfaces - SDF Promotes Free Public UNIX Shell Accounts and Community Services - Vercel Confirms Breach Linked to Third-Party OAuth App, Environment Variables Accessed - Claude Token Counter Adds Cross-Model Comparisons, Reveals Opus 4.7 Token Inflation - Ashley Rolfmore: Software teams can’t framework their way around listening Episode Transcript Fake GitHub stars distort discovery Let’s start with the story that quietly rewrites how projects get noticed: buying GitHub stars. An investigation, backed by a peer-reviewed ICSE 2026 study, argues there’s now a mature market for fake stars—and it’s big enough to distort discovery, credibility, and even financing. Using a tool called StarScout, researchers estimate roughly 6 million suspected fake stars spread across more than eighteen thousand repositories, with activity surging in 2024. The article claims some campaigns were effective enough to push dozens of repositories onto GitHub Trending, which is basically the front page for developer attention. What makes this more than a vanity problem is the downstream impact: the piece argues that some VC firms and scouts increasingly scrape GitHub metrics to source deals, which turns star counts into a cheap, gameable signal of “traction.” It also points out that many AI and LLM-related repos may be non-malicious recipients—meaning hype and opportunism can lift a project even if the maintainers didn’t start the campaign. The authors suggest a few blunt red flags—like unusually high numbers of stargazers with zero followers, or star counts that don’t match forks and watchers. And they raise a sharper consequence: if buying influence is treated like deceptive marketing, regulators could get involved, and startups inflating traction could invite serious scrutiny during fundraising. AI agents reshape SaaS moats Staying in the AI-adjacent world, there’s a growing debate about whether classic SaaS advantages still hold up when “agents” can do the work. One essay uses Figma as the example. Figma won by putting serious design tooling in the browser and turning design into a shared workspace for developers, PMs, and executives. The argument now is that this broad, cross-team adoption is becoming a weakness: LLMs are increasingly “good enough” for lots of the non-core tasks those extra users came to Figma for—things like quick mockups, slides, and on-brand collateral. The piece says Figma’s own AI push, described as “Figma Make,” feels less convincing than what’s coming out of frontier labs—specifically pointing to Anthropic’s Claude Design, which can ingest a company’s design system and generate usable assets fast. The strategic sting is that SaaS companies may buy inference from the very vendors who can turn around and compete with them—while the AI lab has better model economics and faster shipping velocity. Big picture: it’s a warning that collaboration features, plugins, and platform ecosystems matter less if a small team can build an agent that outputs the finished work. Anthropic tools inside US government That leads into another Anthropic-related story—this one about government adoption and vendor risk. Reuters reports that the U.S. National Security Agency is using Anthropic’s “Mythos Preview” tool, even though the Pentagon has reportedly designated Anthropic as a supply-chain risk, citing an Axios report. Reuters notes it couldn’t independently verify broader use across the Defense Department, and the NSA, DoD, and Anthropic didn’t immediately comment. Why it matters is the contradiction: governments want cutting-edge AI for analysis and operations, but they also have to manage risk around vendors, procurement rules, and potential misuse. And the report highlights a specific fear: as models get better at coding and agentic behavior, they may also make it easier to find and exploit software vulnerabilities. So the story isn’t just “AI in government”—it’s the tension between speed and control when the tech is powerful, opaque, and strategically sensitive. Vercel breach via OAuth chain If you’re building with AI day-to-day, here’s a smaller but practical development: tokenization changes that can quietly shift your bill. Simon Willison updated his Claude Token Counter so developers can compare token counts across Claude model IDs. The key finding: Claude Opus 4.7 appears to be the first in that line with a changed tokenizer, and in at least one test—counting the model’s own system prompt—Willison observed about a 1.46x increase versus Opus 4.6. That’s a real impact if your workloads are near context limits or if you’re cost-sensitive. The nuance is important: initial image tests looked like a huge jump, but it turned out the new model accepted much higher image resolutions; at comparable resolutions, token counts were similar. In other words, the cost risk isn’t “everything got 3x worse,” it’s that model updates can change the accounting underneath you. For teams budgeting AI features, that means you can’t just watch per-token pricing—you need to watch how your inputs are being counted. Tokenizer shifts change AI costs Now to security, and a reminder that the weakest link is often an integration you didn’t think twice about. Vercel confirmed a security incident after threat actors claimed to have breached the platform and sell stolen data. Vercel says the impact was limited to certain internal systems and a subset of customers, and that core services stayed operational while incident response and law enforcement got involved. What’s especially notable is the entry point Vercel later disclosed: a compromised Google Workspace OAuth app tied to a third-party AI tool, and an employee account compromise linked to a breach at Context.ai. From there, attackers escalated access and reached customer environment variables that weren’t marked as sensitive—so they weren’t encrypted at rest—and used that to go further. The takeaway is less about one vendor and more about modern supply chains: OAuth apps, SaaS integrations, and “helpful” AI add-ons can become high-impact pathways. And secrets handling is still make-or-break—classification and encryption aren’t paperwork, they’re blast-radius control. Listening failures in product teams Shifting from incidents to org behavior, there’s a thoughtful piece arguing that many teams are trying to systematize their way out of listening. The claim is simple: teams lean on engineering-friendly frameworks to avoid the messier work of understanding users and stakeholders. The post lists common failure modes—treating a request as a literal requirement, assuming everyone shares the same context, oversimplifying people into “technical” versus “non-technical,” and generalizing from one conversation to an entire customer base. Why it matters is that bad listening doesn’t just create a bad feature—it can calcify misunderstandings into the product and into the codebase. Over time, that becomes technical debt, slower delivery, and missed revenue. It’s a good reminder that “process” isn’t the same thing as “insight.” Microwave curing for printed electronics On the research front, one of the more exciting engineering stories today comes from printed electronics. Engineers at Rice University report a technique to “cure” freshly printed conductive inks without overheating the surface underneath—a longstanding bottleneck when you want electronics on delicate materials. Their approach uses a metamaterial-inspired near-field structure to focus microwave energy into a tiny area, heating the ink itself while keeping the surrounding substrate relatively cool. Why it’s interesting is what it enables: printing functional conductive structures on irregular or sensitive surfaces—things like silicone, paper, plastic, and even biological materials—without destroying them in the process. The team demonstrated a wireless strain sensor on bovine bone, and the broader implication is clear: if you can reliably put circuitry onto soft, biocompatible, or even living materials, you open doors to new medical devices, organ-interfacing sensors, and soft robotics that don’t require rigid boards. Japan offshore earthquake update Finally, a quick real-world update outside the usual tech lane, but relevant for infrastructure and risk planning. The USGS reports a magnitude 7.4 earthquake offshore northeastern Japan, about 100 kilometers east-northeast of Miyako, at 07:53 UTC on April 20, 2026. It occurred at roughly 35 kilometers depth and was driven by thrust faulting near the subduction boundary along the Japan Trench—one of the most seismically hazardous regions on Earth, and north of the 2011 Tohoku disaster zone. Early USGS assessments suggested overall impacts were likely to be limited, but offshore quakes can still bring secondary hazards, including aftershocks and localized coastal effects. It’s another reminder that resilience planning—communications, power, logistics, and monitoring—matters just as much as the headline magnitude. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Headphones turned into microphones - A USENIX WOOT paper shows “jack retasking” can flip some headphone ports into inputs, enabling stealth eavesdropping. Key keywords: malware, privacy, HD Audio, Realtek, headphones-as-mic. Claude Opus 4.7 changes - Anthropic’s published Claude system prompt for Opus 4.7 changed in tone and policy emphasis, while community data suggests higher token usage on the same prompts. Key keywords: system prompt, policy, token costs, Claude Opus 4.7, developer impact. Typewriters return in classrooms - A Cornell language instructor uses manual typewriter assignments to reduce AI-assisted writing and better measure real proficiency. Key keywords: education, assessment, academic integrity, typewriters, generative AI. Photonics chips with many lasers - NIST reports integrated photonics that can generate many laser wavelengths on a single chip, potentially shrinking lab-scale optics for quantum and sensing. Key keywords: photonics, lasers, lithium niobate, quantum, optical clocks. Skiplists reborn for SQL trees - Antithesis explains a “skiptree” approach that makes ancestor queries feasible in SQL analytics by borrowing skiplist ideas. Key keywords: data structures, SQL, BigQuery, fuzzing, timeline queries. Why game pause is hard - Game developers explain that ‘pause’ often means carefully managing simulation time, menus, and platform overlays—not simply stopping everything. Key keywords: game engines, timescale, performance, certification bugs, pause systems. Early computing culture preserved - The Internet Archive published BYTE Magazine’s first issue, capturing the DIY microcomputer mindset, alongside a look at pre-GPS navigation with electromechanical computing. Key keywords: computing history, BYTE 1975, preservation, analog computer, navigation. Landing first consulting clients - An HN thread on solo consulting highlights that first contracts often come from relationships, a clear niche, and visible credibility rather than cold outreach. Key keywords: freelancing, consulting, networking, positioning, SMEs. - Internet Archive Posts BYTE Magazine’s Inaugural September 1975 Issue - Researchers Show Malware Can Turn Headphones Into Eavesdropping Microphones via Audio Jack Retasking - Game Developers Reveal the Surprisingly Hacky Ways Games Implement Pause - Antithesis’ “Skiptree” Uses Skiplist Ideas to Make Tree Queries Work in BigQuery - NIST Demonstrates Multilayer Photonics Chips That Generate Many Laser Wavelengths - Community Data Suggests Opus 4.7 Uses About 38% More Tokens Than Opus 4.6 on Average - Cornell instructor uses typewriters to deter AI-written assignments - Hacker News Discusses How Solo Consultants Get Their First Clients - Inside the B-52’s Electromechanical Star-Tracking Navigation Computer - Claude Opus 4.7 System Prompt Adds Expanded Safety Rules, Tool Use Guidance, and New Tool Mentions Episode Transcript Headphones turned into microphones Security and privacy first. Researchers presented a technique dubbed “SPEAKE(a)R” that flips a basic assumption: that headphones are only for output. On many modern PCs, the audio hardware and drivers support “jack retasking,” meaning software can reconfigure an audio port on the fly. The paper demonstrates malware that quietly retasks a headphone or speaker connection into an input and records intelligible speech from across a room—then switches back when you start playing audio so you’re less likely to notice. Why it matters: physical mic covers, mute buttons, and even unplugging a microphone don’t necessarily end eavesdropping if a headset is still connected and the system allows that kind of port remapping. Claude Opus 4.7 changes In AI land, two related signals worth watching: behavior and cost. First, Anthropic updated the published Claude.ai system prompt for Opus 4.7, and analysis of the diff suggests a model that’s being nudged to be more decisive, to finish tasks instead of stalling, and to keep responses shorter and less overwhelming—along with expanded safety guidance in sensitive areas. Second, a community-run calculator aggregated anonymous real-world comparisons that suggest Opus 4.7 can consume noticeably more tokens than Opus 4.6 on the same requests, which translates directly into higher spend for teams running at scale. The takeaway isn’t that anything is “wrong,” but that model upgrades can quietly change your unit economics—so measuring before and after matters just as much as reading release notes. Typewriters return in classrooms That AI spillover shows up in classrooms too. A Cornell instructor teaching German has students do one writing assignment per term on manual typewriters—no screens, no spellcheck, no quick translate-and-polish workflow. The goal is simple: produce writing that reflects what students can actually do, not what a tool can tidy up. Students reportedly slow down, plan more, and collaborate with peers instead of defaulting to online fixes. Why it matters: universities are actively redesigning assessments toward formats that better capture real skill in a world where “perfect prose” is increasingly cheap to generate. Photonics chips with many lasers Now to hardware that sounds like science fiction but is very real lab progress. NIST and collaborators published work on integrated photonics chips that can produce many laser wavelengths on a small device by stacking materials with complementary strengths. The punchline is practical: a lot of quantum systems, optical clocks, and precision sensors need very specific laser colors, and today that often means bulky, power-hungry setups. If this kind of multi-wavelength photonics scales, it could help move quantum-grade instrumentation out of specialized labs and toward portable systems—plus improve optical links that are increasingly relevant as AI-era hardware looks for faster ways to move data between chips. Skiplists reborn for SQL trees On the data-engineering side, there’s a nice reminder that ‘old’ data structures still earn their keep. Antithesis described how skiplists inspired a “skiptree” to answer ancestor-style queries over huge branching timelines from fuzzing runs. The problem: analytics warehouses like BigQuery are great at scanning large tables, but they’re awkward when you need lots of iterative parent-pointer lookups. Their workaround stores a sparse hierarchy that lets you jump through ancestry with a fixed set of SQL joins, keeping the workload closer to what columnar analytics systems are good at. The broader point is economic: sometimes you don’t pick the perfect database—you adapt your representation so the database you have stops fighting the question you’re asking. Why game pause is hard For game development, a deceptively simple question got a lot of honest answers: what actually happens when you press pause? Developers say it’s rarely a literal freeze. Many engines pause by manipulating time—sometimes even using near-zero values to dodge edge cases—while also juggling menus, controller disconnect states, and console overlays that can collide in messy ways. Some teams even “pause” by capturing a still frame, then rearranging what’s rendered behind the menu to reduce complexity. Why it matters: pausing touches performance, user experience, and platform compliance, and it’s one of those features that looks trivial until late-stage bugs turn it into a release blocker. Early computing culture preserved Time for a double shot of computing history. The Internet Archive posted a digitized copy of BYTE Magazine’s very first issue from September 1975—titled “Computers—the World’s Greatest Toy.” It’s packed with the early personal-computing mindset: practical interfaces, homebrew assembly tooling, and making the most of surplus parts. This isn’t nostalgia for nostalgia’s sake—having primary sources searchable and downloadable helps historians and working engineers understand what constraints and assumptions shaped the microcomputer boom. And in a different corner of history, Ken Shirriff looked at the “Angle Computer,” an electromechanical analog computer used in B-52 celestial navigation gear before GPS. It’s a striking example of how mechanical linkages, servos, and early electronics worked together to solve real navigation problems with serious accuracy—right at the moment before digital approaches took over. Together, these stories are a reminder that today’s ‘obvious’ solutions were once wildly non-obvious, and ingenuity filled the gaps. Landing first consulting clients Finally, a career note from Hacker News: a new solo consultant asked how people land their first paid projects, especially when helping small and mid-sized businesses untangle workflows and integrations. The consensus was blunt: the market is crowded, and cold outreach is tough. Many first contracts still come from existing relationships—former coworkers, referrals, communities where you’ve been consistently useful—and from having a clear niche rather than “I can do anything in software.” The useful takeaway is that credibility is often built in public and cashed in through trust, not through perfect pitch decks. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Claude 4.7 tokenizer cost shock - A developer found Claude Opus 4.7’s new tokenizer can inflate English-and-code prompts up to ~1.47x, impacting context window, caching, and effective cost despite unchanged pricing. Floating-point equality and epsilons - A critique of “just use an epsilon” explains why fuzzy float comparisons can break transitivity and algorithms, and why exact checks are sometimes the correct choice under IEEE-754. Interval unions for uncertainty math - A new open-source interval-union calculator models uncertainty with unions of ranges, preserving the inclusion property even through tricky operations like division by intervals containing zero. Emacs 30 file trust usability - An Emacs package smooths Emacs 30’s new file-trust security model with just-in-time prompts and project-scoped trust, reducing risky workarounds while keeping protections. Kdenlive 2026 stability and AI - Kdenlive’s state-of-the-project report emphasizes stability, workflow polish, and practical AI-assisted masking/segmentation, alongside interoperability and performance gains. Michael Oser Rabin remembered - Michael Oser Rabin, Turing Award winner, pioneered nondeterminism, randomized algorithms, and cryptography-adjacent tools like Miller–Rabin—core building blocks of modern computing. Category theory meets order theory - A “Category Theory Illustrated” chapter reframes orders as relations, connects preorders to thin categories, and links joins/meets to categorical coproducts/products for clearer intuition. Amiga art preservation and attribution - The Amiga Graphics Archive added newly found early work by Jo-Anne Park, highlighting the ongoing challenge of provenance and credit in retrocomputing digital preservation. Lunar dust health risks for astronauts - ESA is investigating whether sharp, electrostatically active lunar dust could cause long-term respiratory harm, a key unknown for sustained Moon bases and repeated missions. - Category Theory Illustrated Explains Orders, Posets, and Preorders as Thin Categories - Turing Award Winner and Cryptography Pioneer Michael O. Rabin Dies at 94 - Amiga Graphics Archive Adds Newly Found Early Works by Jo-Anne Park - Anthropic Launches Claude Design to Generate and Iterate on Prototypes and Visual Assets - Why Epsilon Comparisons Often Harm Floating-Point Code - Open-Source Interval Calculator Brings Interval-Union Arithmetic and Guaranteed Bounds to the Browser - Kdenlive’s 2026 Project Update: Stability Push, OpenTimelineIO Rewrite, and Dopesheet Plans - Tests Suggest Claude Opus 4.7’s Tokenizer Increases Token Counts for English and Code, With Modest Instruction-Following Gains - trust-manager Introduces Just-in-Time Trust Prompts for Emacs 30 - ESA Investigates Health Risks of Lunar Dust Ahead of Sustained Moon Missions Episode Transcript Claude 4.7 tokenizer cost shock Let’s start in AI tooling, with a story that’s less about shiny new features and more about the economics of everyday usage. A developer compared token counts between Claude Opus 4.6 and 4.7 and found that the new tokenizer can noticeably increase token usage for English and code-heavy text—sometimes well beyond the “roughly up to 1.35x” expectation. That matters because pricing and quotas didn’t change, so your effective context window shrinks and long sessions can hit limits sooner. The author did see a modest improvement in strict instruction-following on a small benchmark sample, but the broader point is: model updates can change the practical cost profile even when the sticker price stays the same. Floating-point equality and epsilons Staying in the “things that bite you slowly” category, there’s a strong essay pushing back on the reflexive advice: “never compare floating-point numbers for equality.” The argument isn’t that floating-point is magically safe—it's that the blanket fix people reach for, epsilon comparisons, can be worse. Epsilons are often arbitrary, and more importantly they can break transitivity, which quietly poisons sorting, deduping, geometry code, and anything that assumes equality behaves consistently. The takeaway is to choose comparisons based on the property you actually need—sometimes exact equality is precisely the right guardrail, like detecting true zero cases, and sometimes the real solution is a more robust algorithm instead of a wider tolerance. Interval unions for uncertainty math Related, but from a more constructive angle, an open-source “Interval Calculator” takes uncertainty seriously by computing with unions of intervals rather than single numbers. The headline benefit is that it stays honest through operations that normally get awkward—like dividing by a range that includes zero—by returning disjoint ranges when that’s the mathematically correct answer. Why it matters: if you’re doing worst-case analysis, validation, or safety bounds, the promise you want is “the true value is definitely inside this result,” not “it’s probably close.” It’s a reminder that for some domains, the best way to handle numerical fragility is to change the model you compute with, not just tweak the rounding at the edges. Emacs 30 file trust usability On the developer-workflow front, Emacs got a notable security shift in Emacs 30: files are no longer treated as automatically trusted. That’s good for reducing the blast radius of malicious content, especially after real-world concerns like arbitrary code execution vulnerabilities. But the default friction is also real—features can quietly stop working in untrusted buffers, and users get tempted into blunt, risky workarounds. A new package called trust-manager tries to thread that needle by prompting you at the moment it matters, remembering your decision per project, and making trust status visible and reversible. The broader theme here is important: security controls only work long-term if they fit how people actually work. Kdenlive 2026 stability and AI In open-source creative tools, Kdenlive published a 2026 state-of-the-project update that’s refreshingly focused on stability and polish rather than chasing novelty. The project still shipped meaningful upgrades—like improved masking and object segmentation for background work—while also investing in performance and crash fixes. Interoperability got attention too, with work around OpenTimelineIO to make it easier to move projects between editing ecosystems. Why it matters: for creator tools, reliability is a feature, and incremental workflow improvements often beat flashy additions when you’re editing on deadlines. Michael Oser Rabin remembered A major figure in computer science passed away this week: Michael Oser Rabin, who died on April 14th at age 94. Rabin’s fingerprints are all over modern computing—from foundational theory like nondeterministic automata, to practical algorithms like Rabin–Karp, and especially to cryptography through the Miller–Rabin primality test. If you’ve ever used secure communication that depends on generating and checking large primes efficiently, you’ve benefited from his work. The significance isn’t just historical; it’s a reminder that a lot of today’s “normal” infrastructure rests on ideas that once looked purely theoretical. Category theory meets order theory For the more mathematically inclined, there’s a new installment of “Category Theory Illustrated” that uses order theory as a bridge into categorical thinking. Instead of treating order as a scoring system for objects, it focuses on relations and the laws they obey—showing how partial orders, preorders, and equivalence classes fit together. The key category-theory punchline is that preorders behave like “thin” categories, where the usual category notions—composition and identity—show up as transitivity and reflexivity. And joins and meets line up with coproducts and products. Why it matters: it’s a clean mental model for moving from everyday “less than” intuition to the diagram-driven way category theory lets you reason about structure. Amiga art preservation and attribution In retrocomputing and digital preservation, the Amiga Graphics Archive posted an update that sounds small but is actually the heart of archival work: they uncovered very early images by artist Jo-Anne Park, including versions from the Commodore 64 era, helping confirm attribution and show her artistic progression. Preservation isn’t just about storing files—it’s about provenance, credit, and context. When the original source files are missing and work survives only through scattered scans and secondhand copies, every confirmed link between an artifact and its creator becomes part of rebuilding the record of an era. Lunar dust health risks for astronauts And finally, from space to the lungs: ESA highlighted a lingering problem from Apollo that’s getting new attention as agencies plan longer stays on the Moon. Astronauts reported irritation and allergy-like symptoms from lunar dust—fine, abrasive particles that stuck to suits and got everywhere. The open question is the long-term risk: could inhaled lunar dust lead to serious respiratory damage over time? The Moon’s environment makes the problem nastier—low gravity keeps dust suspended longer, and electrostatic charging can help it cling and infiltrate habitats. This matters because “dust control” isn’t a housekeeping issue for future bases; it’s a health requirement and a systems reliability issue rolled into one. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: AI coding agents go desktop - OpenAI’s Codex expands into an agent that can operate desktop workflows, speeding up testing and iteration where no clean API exists. Claude Opus 4.7 safeguards - Anthropic’s Claude Opus 4.7 improves long-running coding reliability and ships new cybersecurity request blocking, highlighting safety controls for more capable AI models. AI in hardware verification loops - A hands-on workflow pairs Claude Code with SPICE simulation and oscilloscope captures to automate waveform comparison and reduce human “eyeballing” in electronics debugging. Ada anticipated modern safety - An essay revisits Ada’s packages, strong typing, concurrency, and later contracts, arguing today’s “safe software” movement echoes ideas standardized decades ago. HPC stuck with old models - A retrospective on HPC argues hardware raced ahead—GPUs, NUMA, huge parallelism—while programming models like MPI/OpenMP stayed dominant, largely due to institutional inertia. Python bytecode in 500 lines - The Byterun chapter demystifies how Python executes bytecode via frames and a stack VM, giving developers practical insight into debugging and performance limits. Keyboard-first image viewer updates - FIM, a Vim-inspired image viewer, ships fixes and broader platform and format support, keeping lightweight and terminal-centric workflows viable. Code-first parametric CAD workflows - CadQuery promotes reproducible, version-controlled 3D modeling by describing CAD as Python code, aligning mechanical design with software practices. AI-optimized Chinese flashcards - A language learner builds an AI-assisted browser injection to compress lookup time inside flashcards, pushing vocabulary acquisition by removing interface friction. - Essay Argues Ada Anticipated Many Features of Modern Programming Languages - Anthropic Launches Claude Opus 4.7 with Stronger Coding, Higher-Resolution Vision, and Cyber Safeguards - FIM Image Viewer Releases 0.7.1 Snapshot With GTK/SDL Fixes and Updated Documentation - OpenAI Expands Codex With Computer Control, Plugins, Memory, and Long-Running Automations - Learner Builds AI-Powered Flashcard Overlay to Speed Up Chinese Character Acquisition - Rawquery Argues LLMs Make ‘Average’ Analytics Work Cheap—and That’s Powerful - CadQuery promotes code-based, parametric 3D CAD modeling in Python - Byterun Shows How a Python Bytecode Interpreter Works in About 500 Lines - HPC Hardware Leaps Forward While Programming Languages Largely Stand Still, Chapel Author Argues - Developer Demonstrates Claude-Assisted Hardware Verification Using SPICE and Oscilloscope Data Episode Transcript AI coding agents go desktop Let’s start with AI for software development, because two big releases are pushing the same theme: agents that act more independently, and the guardrails needed to keep that power from going sideways. OpenAI rolled out a major update to Codex, aiming to make it less of a code-completion tool and more of a full development partner. The standout change is “background computer use” in the desktop app—so the agent can interact with your screen, click around, type, and run workflows that don’t have neat APIs. That matters because a lot of real engineering work is still glued together by GUIs, local dev servers, and ad-hoc testing steps. If an agent can handle that reliably, it’s not just writing code—it’s shortening the entire loop from change to validation. Claude Opus 4.7 safeguards In the same neighborhood, Anthropic released Claude Opus 4.7 as a general upgrade, with an emphasis on harder software engineering tasks and long, multi-step work. The practical angle here is reliability: better instruction-following, more consistent completion of extended tasks, and more self-checking before it reports results. The bigger story is safety. Opus 4.7 is the first Anthropic model shipped with new automated safeguards that detect and block high-risk cybersecurity requests. That’s notable because as these models become more capable, the security posture can’t be an afterthought. Anthropic is also opening a verification program for vetted security pros, trying to draw a clearer line between legitimate testing and misuse. AI in hardware verification loops Taken together, these updates underline a shift: the competition isn’t only about raw model intelligence anymore. It’s about operational control—how you budget longer runs, how you review outputs, and how you keep agentic behavior aligned with policy and intent. Even small changes, like new tokenization or higher-effort reasoning modes, can ripple into cost, latency, and developer expectations, so teams will have to measure impact rather than assuming a drop-in swap. Ada anticipated modern safety Staying with AI, but moving from software to electronics: one write-up describes using Claude Code as part of a hardware verification loop—pairing it with a SPICE simulator and an actual oscilloscope. The key point isn’t that AI can magically design complex circuits from a prompt. The author says that only works for trivial cases. The win is in the tedious verification work: comparing simulated and measured waveforms, normalizing time axes, and catching mismatches that engineers too often eyeball. It’s a good example of where AI helps most today—turning repetitive, error-prone checks into something more systematic, while still relying on real measurements and careful constraints. HPC stuck with old models Now for a sharp pivot into programming language history—because one essay argues we’ve been re-discovering “modern” language safety ideas that Ada baked in decades ago. The piece retells why Ada existed in the first place: the U.S. Department of Defense had a chaos of hundreds of languages and dialects across critical systems, so they pushed for standardization, culminating in the Steelman requirements and Ada’s release in the early 1980s. The author’s claim is that Ada anticipated a lot of today’s safety playbook: clear separation between interface and implementation, strong typing that prevents entire categories of bugs, standardized generics, and concurrency constructs designed to reduce race conditions and deadlocks. What makes it relevant now is the broader industry push toward verifiable software—contracts, stricter null handling, and even formal proofs through tools like SPARK. The essay’s punchline is cultural: Ada’s reputation for verbosity and its government-procurement vibe may have kept it out of the spotlight, even as its ideas quietly won in safety-critical domains where failures are, ideally, invisible. Python bytecode in 500 lines That leads neatly into a broader complaint from the high-performance computing world: hardware has evolved at a breathtaking pace, but programming models haven’t kept up. A retrospective contrasts decades of gains—massive core counts, GPUs, complicated memory hierarchies—with the familiar reality of Fortran, C and C++, plus MPI and OpenMP. The author argues programming has arguably gotten harder because modern machines force you to manage more forms of parallelism and think constantly about data movement. And importantly, they don’t chalk this up to a lack of good ideas. They point at institutional barriers: legacy codebases, funding that prioritizes hardware over software, and weak pathways for research languages to become dependable tools. If you’ve ever wondered why “better languages for parallel computing” don’t take over, this is the uncomfortable answer: it’s less about syntax, more about economics and risk tolerance. Keyboard-first image viewer updates For developer education, another item worth your time comes from Allison Kaptur’s “500 Lines or Less” series, with a chapter on Byterun—a Python bytecode interpreter written in Python. Why it matters is clarity. By shrinking the core structure of a CPython-like virtual machine into a readable codebase, it helps developers understand what actually happens between source code and execution: bytecode, frames, stacks, jumps, and why dynamic typing forces so much to be decided at runtime. Even if you never write an interpreter, this kind of mental model pays off when you’re debugging weird control flow, performance surprises, or generator behavior. Code-first parametric CAD workflows On the lighter tooling side, FIM—the keyboard-driven image viewer inspired by Vim—posted updated release notes and a fresh snapshot. This is the kind of project that quietly keeps power-user workflows alive: fast browsing of large photo sets, terminal and lightweight GUI options, and better stability across environments. The update is mostly about fixes and refinements, plus broader platform and format compatibility. It’s not flashy, but it’s the sort of incremental maintenance that makes small utilities dependable enough to stick around for years. AI-optimized Chinese flashcards If you like “infrastructure for making things,” CadQuery also showed up as a reminder that CAD can be treated like code. CadQuery is a Python library for parametric 3D modeling, built around the idea that designs should be scriptable, version-controlled, and easy to regenerate with new parameters. The broader significance is reproducibility: when the model is code, you can review changes, automate variations, and integrate design work into the same disciplined workflows software teams take for granted. Story 10 Finally, a language-learning post makes a compelling argument that the bottleneck for intermediate reading—at least in Chinese—isn’t motivation, it’s coverage. The author says that even understanding most tokens still leaves too many high-importance gaps, forcing constant lookups that break comprehension. Their response is aggressive memorization to push coverage higher, but they hit a different wall: tool friction. So they built a browser injection to pull dictionary context, character breakdowns, and quick explanations directly into the flashcard screen, cutting lookup time from tens of seconds to under one. It’s a neat case study in what AI-assisted coding enables: not just new apps, but personal “micro-automation” that removes tiny delays until the whole workflow feels different. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: IPv6 adoption hits new highs - Google’s IPv6 stats show 45.54% of user traffic over native IPv6, highlighting real-world internet scaling beyond IPv4 and where reliability still lags. AI models raising security stakes - Researchers show an AI assistant can meaningfully accelerate exploit discovery on embedded devices, while Anthropic’s unreleased Mythos model tested ahead on simulated cyberattacks—raising governance and defense-cost questions. Windows Defender bug claim resurfaces - A new proof-of-concept alleges Windows Defender remediation behavior could be abused for local privilege escalation, turning a protective feature into a potential file-write risk if confirmed. Spam reporting meets platform walls - A fediverse post about escalating Gmail spam reports underscores the friction of automated abuse pipelines at large platforms versus community moderation approaches in decentralized networks. Ancient DNA reveals recent selection - A Nature study using 15,836 genomes reports hundreds of selection signals across 10,000 years in West Eurasia, especially in immunity, and suggests polygenic shifts tied to changing environments. AI-powered return to paper workflows - James Somers argues AI could enable a “paper computer,” translating pen-and-paper organization into digital action to reduce distraction and restore focus—an alternative to screen-first productivity. Why XOR swap is trivia - A fresh look at the XOR swap trick finds modern compilers negate its benefits, and in real code it can be slower or unsafe—useful mostly as a cautionary tale about “clever” code. - Google Data Shows IPv6 Usage Reaches About 46% of User Connections - Researchers Show Codex Escalating a Samsung Smart TV Browser Shell to Root via Physical-Memory Mapping Bug - Large ancient DNA study finds hundreds of recent selection signals across West Eurasia - Darkbloom Launches Decentralized, Hardware-Verified Private AI Inference on Idle Apple Silicon Macs - Mastodon user seeks direct Gmail contact to report large-scale spam, cites ineffective abuse reports - Mythos Benchmarks Suggest Cybersecurity Is Becoming a Token-Budget Arms Race - PoC ‘RedSun’ claims Windows Defender cloud-tag handling can enable admin privilege escalation - James Somers’ Vision for a ‘Paper Computer’ That Uses AI to Escape Screens - Why the XOR Swap Trick Rarely Helps in Modern Code Episode Transcript IPv6 adoption hits new highs First up, a quiet milestone for the internet’s plumbing. Google’s IPv6 statistics page now shows that 45.54% of its users were reaching Google over IPv6 as of April 13th. The striking detail is that almost all of it is native IPv6, not older transition tricks. Why you should care: IPv6 adoption is one of the cleanest signals that the internet is actually expanding beyond IPv4 address scarcity. And Google’s regional views add an important reality check—deployment isn’t the same as quality. Some places may “have IPv6,” yet still suffer higher latency or reliability problems when hitting IPv6-enabled sites, which is exactly the kind of nuance operators and policymakers need to see. AI models raising security stakes Now to security, where two stories rhyme: AI is becoming less of a coding helper and more of an operational force multiplier. In one experiment, researchers gave OpenAI’s Codex a post-exploitation environment on a Samsung Smart TV and asked it to go from a limited foothold to root. The takeaway wasn’t a single magic trick—it was the workflow. With source access, logs, and the ability to compile and iterate, the model helped identify a dangerously permissive kernel interface and turned it into a practical escalation. The bigger point is about capability: when an AI can combine code reading, system probing, and rapid iteration, the barrier shifts from “can I find the bug” to “do I have access and time,” and both of those are becoming cheaper. Windows Defender bug claim resurfaces In a similar vein, Anthropic previewed a model called Mythos that they say was strong enough at cybersecurity tasks that they chose not to release it broadly. The UK AI Security Institute’s evaluation largely backed the claim: on a demanding simulated corporate network attack, Mythos outperformed other frontier models and was the only one to complete the full chain in repeated trials. What’s interesting here is less the leaderboard and more the economics. Evaluators observed that performance improved with bigger token budgets, hinting that brute-force exploration—powered by more compute—can keep buying results. If that pattern holds, security starts to look like a budget contest: defenders may need to spend heavily to find and fix weaknesses before attackers spend less to exploit them. That pushes organizations toward more rigorous hardening phases, and it also strengthens the case for widely-audited components where many parties can fund scrutiny. Spam reporting meets platform walls Staying with security, there’s a new GitHub proof-of-concept repository claiming a Windows Defender behavior could be leveraged for privilege escalation. The allegation is that under certain conditions, Defender’s handling of a detected file could result in a rewrite to the original location in a way an attacker might abuse to overwrite protected files. If true, that’s the kind of bug that’s especially unsettling because it flips a safety mechanism into a write primitive—exactly the sort of thing local attackers look for. At this stage, the important note is that it’s a public claim with a PoC, not a fully adjudicated incident report, so the right posture is cautious attention: watch for vendor confirmation, patches, and independent validation rather than assuming either doom or dismissal. Ancient DNA reveals recent selection On the platform side, one smaller thread captured a familiar frustration: a fediverse user asked if anyone could connect them to a human on the Gmail team to report what they described as serious, actionable spam activity. They claimed a spammer sent over ten thousand messages through Gmail in a week, and that standard abuse forms led nowhere. The replies quickly turned into a mini-demonstration of decentralized norms—people pointing out spammy behavior in the conversation itself and suggesting local server moderation. The broader theme is accountability and access. Centralized platforms scale with automation, but that can make escalation feel impossible even when someone believes they’ve found something important. Decentralized systems don’t magically solve spam, but they do offer more visible, local levers—sometimes for better, sometimes for chaos. AI-powered return to paper workflows Switching gears to science, a Nature study analyzed 15,836 ancient and recent genomes from West Eurasia and introduced a time-series method designed to detect directional natural selection while accounting for migration and population structure. Using that approach, the authors report hundreds of independent selection signals over about ten thousand years—far more than earlier ancient-DNA scans typically surfaced. Many are tied to immune and inflammatory pathways, with signals intensifying around the Bronze Age, consistent with changing disease pressures as populations densified and lifestyles shifted. The careful, important nuance: the paper also discusses polygenic shifts connected to modern trait predictors, while warning that today’s labels don’t necessarily map neatly onto ancient adaptive realities. Still, the resource and method matter because they let researchers track evolutionary change through time rather than inferring it from a single snapshot. Why XOR swap is trivia From biology to interface design, James Somers makes a provocative productivity argument: AI could enable a kind of “paper computer,” where the calm, spatial advantages of paper—markups, note cards, physical sorting—are translated back into digital actions without forcing you to live inside a screen. The pitch isn’t nostalgia for messy workflows; it’s a critique of modern computing’s default mode: distractible, multitasking, and notification-driven. Somers imagines systems that respect single-purpose modes and turn physical interaction into structured digital updates, giving you the convenience of syncing without the constant attention tax. Whether or not the exact vision lands, it’s a useful reframing: AI doesn’t have to mean more screen time; it could be used to make computing quieter. Story 8 Finally, a bit of programming folklore gets a reality check. A post revisits the classic XOR swap trick—swapping two values without a temporary variable—and asks whether it ever helps. In modern compiled code, it basically doesn’t. Optimizing compilers already handle swaps efficiently, and the XOR version often generates extra work, while also introducing footguns like breaking when both pointers refer to the same place. The lesson is broader than one trick: “clever” code that once made sense in constrained environments frequently becomes slower, riskier, and harder to optimize today. It’s a reminder that readability and correctness usually win, and that the compiler is almost always better at this particular game than we are. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Learn compilers without the pain - A compiler-education critique argues beginners get buried in theory, and points to Crenshaw plus a nanopass framework to build real compilers sooner. Keywords: compiler education, AST, nanopass, transformations, Dragon Book. Sleep science and learning outcomes - Dr. Piotr Wozniak’s sleep synthesis connects circadian alignment to memory, mood, safety, and productivity—calling out alarms, light, and early school schedules. Keywords: circadian rhythm, learning, free-running sleep, naps, delayed sleep phase. A sea slug that photosynthesizes - The “leaf sheep” sea slug can temporarily photosynthesize by stealing chloroplasts from algae, offering a rare window into animal-plant cellular borrowing. Keywords: kleptoplasty, chloroplasts, evolution, sea slug, photosynthesis. On-device AI with Gemma 4 - Google’s Gemma 4 models can run fully offline on iPhones via an app, highlighting a shift toward private, reliable on-device inference. Keywords: on-device AI, iPhone GPU, offline inference, privacy, LLM. Claude Code routines for automation - Anthropic adds “routines” to Claude Code—saved, triggerable autonomous runs tied to repos and connectors—powerful, but with real security implications. Keywords: automation, GitHub triggers, HTTP API, autonomous agents, permissions. Legacy X11 bug freezes desktops - A rare edge case in Enlightenment E16 could hard-freeze an X11 session due to a non-terminating title-truncation loop, fixed with simple guardrails. Keywords: X11, window manager, infinite loop, robustness, defensive programming. Weird-shaped Windows apps return - A post argues Windows apps feel bland partly because web wrappers displaced native Win32 creativity, and shows why custom-shaped windows are both possible and risky. Keywords: Win32, Electron, UI identity, reliability, desktop apps. WWII radar labs and reverse engineering - MIT’s Rad Lab history shows how coordinated research and the cavity magnetron accelerated radar in WWII, while a reverse-engineering essay highlights the value of recovering intent from artifacts. Keywords: radar, magnetron, MIT Rad Lab, reverse engineering, benchmarks. - Two Recommended Readings to Demystify Writing a Compiler - Essay Links Sleep Quality to Learning and Warns Modern Schedules Disrupt Circadian Health - Leaf Sheep Sea Slug Uses Stolen Chloroplasts to Photosynthesize - Enlightenment E16 Patch Fixes 20-Year-Old Infinite Loop Triggered by Long Window Titles - Google’s Gemma 4 LLM Now Runs Offline on iPhones via AI Edge Gallery - YC Startup Proliferate Seeks Founding Engineer to Build Human-AI Agent Engineering Workspace - Claude Code adds cloud-run “routines” with scheduled, API, and GitHub triggers - How MIT’s Radiation Laboratory Accelerated Radar Development in World War II - Why Weird-Shaped Windows Faded—and How Win32 Still Enables Them - How Bulgaria’s Apple II Clone and a 1999 Study Revealed the Purpose of Key ISCAS-85 Circuits Episode Transcript Learn compilers without the pain Let’s start with learning and building. One write-up takes aim at the way compiler textbooks often greet beginners: with sweeping theory, lots of terminology, and the vibe that compilers are mythical beasts best approached with protective gear. The author argues that this is exactly how we keep the “compilers are impossibly hard” legend alive. Instead, they recommend Jack Crenshaw’s classic “Let’s Build a Compiler!” approach: build something small, single-pass, and understandable, and get a working result early. The caveat is important, though—Crenshaw mostly sidesteps using an internal representation like an AST, which limits how easily you can evolve the compiler. To bridge that gap, the author points to a “nanopass” framework for teaching compilers: many tiny, readable transformations over an internal program structure, rather than a few enormous, mysterious phases. Why it matters: it turns compilers from a monolith into a chain of manageable steps—easier to test, easier to extend, and far less intimidating for learners who want to build, not just memorize. Sleep science and learning outcomes Staying with brains, but moving from software to biology: a long-running article by Dr. Piotr Wozniak synthesizes sleep research with a blunt practical message—modern schedules and habits often fight our circadian biology, and we pay for it in learning, mood, and health. The piece argues that sleep quality isn’t just about hours; it’s about timing—matching real sleep pressure with circadian rhythms. It also pushes the idea of “free-running sleep,” meaning you sleep when you’re genuinely sleepy and wake naturally, as a way many people can re-align when life has shoved them out of sync. Why this lands on Hacker News: it connects sleep loss not only to personal productivity, but to societal-scale outcomes—accidents, chronic disease risk, and large economic costs. It also calls out early school schedules for teenagers as a recurring mismatch with adolescent biology. Even if you don’t buy every claim, the takeaway is clear: if you’re optimizing your stack, sleep is still a foundational dependency. A sea slug that photosynthesizes Now for the surprise creature of the day: Costasiella kuroshimae—the “leaf sheep” sea slug. This tiny, shell-less sea slug can photosynthesize, sort of. It eats particular algae and steals their chloroplasts, keeping them functional inside its own cells for a while. That trick, called kleptoplasty, gives it a short-term energy supplement and a backup plan when food is scarce. Why it matters beyond trivia: it’s a rare example of an animal temporarily integrating plant-like machinery without fully becoming plant-like. For researchers, that’s a natural experiment in what evolution can get away with, and how far biological systems can stretch the usual boundaries between categories we take for granted. On-device AI with Gemma 4 Switching to AI and devices: Google’s open-source Gemma 4 model family can now run natively on iPhones with fully offline inference, using an app called Google AI Edge Gallery. The headline isn’t just “a model runs on a phone”—we’ve been inching toward that for years. The bigger story is that consumer devices are increasingly capable of useful local AI workloads, with no API calls and no network dependency. That changes the privacy equation, the reliability story, and the feasibility of AI in environments where cloud access is restricted—think field work, regulated industries, or just bad connectivity. It also nudges the ecosystem: once offline inference is practical, developers can design features that assume local intelligence is always available, rather than treating AI as a remote service you rent per request. Claude Code routines for automation Related, but more on the workflow side: Anthropic introduced “routines” in Claude Code. The idea is to save a setup—your prompt, repos, connectors, environment—and have it run automatically on Anthropic-managed infrastructure. It can be triggered on a schedule, by an HTTP endpoint, or by GitHub events. The appeal is obvious: you can offload repetitive engineering chores like review assistance, documentation checks, or routine verification tasks. But the interesting part is the tradeoff. These runs can be fully autonomous, and actions in GitHub can appear as you. That turns this from a convenience feature into something that demands real operational discipline: scoped permissions, careful connector access, and an assumption that automation is only as safe as the boundaries you set. It’s a glimpse of where “AI coding” is headed—less chat, more background processes that quietly do work on your behalf. Legacy X11 bug freezes desktops Now, a cautionary tale from the land of desktops and edge cases. A long-standing bug in the Enlightenment E16 window manager could freeze an entire X11 session when opening a specific PDF—because the window title was too long to fit in the decoration. Debugging traced it to a title-truncation routine that tries to insert a middle ellipsis, using an iterative search that can fail to converge and instead oscillate forever. Why this matters: it’s a reminder that seemingly clever math-y iteration inside UI code can become catastrophic when there’s no escape hatch. The fix wasn’t glamorous—caps on iterations, sensible bounds, and safer estimates—but that’s the point. Robustness often comes from boring guardrails that prevent rare input from turning into total failure. Weird-shaped Windows apps return On the topic of desktops, another post argues that modern Windows apps feel slower, heavier, and visually same-y partly because so many are effectively web apps in a wrapper. The author revisits an older Windows culture where apps had distinctive shapes and playful custom UI, and points out that Win32 can still do that today. The real reason it faded isn’t just technical capability—it’s the hidden cost. Once you abandon standard window frames, you inherit a pile of tricky behaviors users now expect to “just work,” from resizing to hit-testing to high-DPI quirks. The broader takeaway: the industry didn’t only choose uniformity out of laziness. It also chose predictability. But as toolkits improve and hardware gets faster, there’s still room for identity—if developers are willing to pay the maintenance bill. WWII radar labs and reverse engineering Finally, a quick double-feature on tech history and the value of understanding what already exists. MIT Lincoln Laboratory’s history page revisits the WWII-era MIT Radiation Laboratory, where coordinated research—boosted by the British cavity magnetron—accelerated practical microwave radar. It’s a case study in how quickly capabilities can advance when government, academia, and industry align around urgent goals. And in a separate reflection on reverse engineering, Alexander Feldman highlights two stories: Bulgaria’s Pravetz machines, which cloned the Apple II to spread computing under constrained access, and the ISCAS-85 circuit benchmarks that researchers studied for years without knowing their purpose—until a team reverse-engineered them and uncovered real functional blocks. Why these belong together: both stories argue that progress isn’t only about inventing the new. Sometimes the real leap comes from recovering intent—figuring out what a system is trying to do—so you can evaluate it honestly, improve it meaningfully, and build on it instead of guessing. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: WordPress plugin supply-chain backdoor - A major WordPress supply-chain attack abused trusted plugin updates to plant a backdoor across 30+ plugins, even using an Ethereum smart contract for C2 discovery—highlighting plugin ownership-transfer risk and SEO spam payloads. Google targets back-button hijacking - Google Search added an explicit spam policy against back-button hijacking, warning of manual actions and ranking demotions as it cracks down on deceptive navigation and malicious ad-tech behavior. GitHub adds stacked pull requests - GitHub rolled out native stacked pull requests plus a gh CLI extension, making it easier to review large changes as ordered, smaller PRs with safer CI and branch-protection behavior. Jujutsu version control gains attention - The jj CLI for the Jujutsu VCS is positioned as a simpler, more composable alternative to Git, with Git-compatible storage that lets developers experiment without team-wide migration. Diffusion LMs catch up to AR - Researchers introduced I-DLM, a diffusion-based language model approach that aims to match autoregressive quality while improving throughput via parallel token generation and introspective consistency. OpenDuck brings DuckDB to cloud - OpenDuck is an open-source effort to make DuckDB work transparently with remote databases using a minimal protocol and hybrid local-remote execution—an extensible alternative to proprietary ‘DuckDB in the cloud.’ Backblaze backup exclusions raise trust - A long-time Backblaze user reports silent exclusions of .git and common cloud-sync folders like OneDrive/Dropbox, raising concerns about backup transparency, retention assumptions, and user notification. Franklin’s Apple II clone era - A look back at Franklin Computer’s early-1980s Apple II compatibles shows how aggressive cloning and flashy marketing collided with evolving IP enforcement and landmark legal pressure. NimConf 2026 community timeline - NimConf 2026 announced its online schedule and CFP dates, setting the community’s rhythm for sharing Nim projects, libraries, and real-world use cases ahead of the June event. - DaVinci Resolve 21 Adds Dedicated Photo Page for Still-Image Grading and Workflow - NimConf 2026 Set for June 20 With Call for Talk Proposals - Google Updates Search Spam Policies to Ban Back Button Hijacking - Jujutsu’s `jj` CLI: A Simpler, More Powerful Alternative to Git - Backdoor Found Across 30+ WordPress Plugins After Portfolio Sale - I-DLM claims diffusion language models can match autoregressive quality while decoding faster - User Claims Backblaze Quietly Excluded OneDrive, Dropbox, and Git Folders from Backups - GitHub Adds Native Stacked Pull Requests with gh-stack CLI - Franklin’s Apple II Clone Ads and the Benjamin Franklin Impersonator Behind Them - OpenDuck launches as an open-source, self-hosted alternative for hybrid DuckDB cloud execution Episode Transcript WordPress plugin supply-chain backdoor First up, a serious WordPress supply-chain incident that’s a reminder of how fragile “trusted updates” can be. A security researcher says an attacker bought a portfolio of popular WordPress plugins and later pushed updates that planted a backdoor across more than 30 plugins. The fallout wasn’t just theoretical: analysis suggests some sites had code injected into wp-config.php, and the malicious behavior selectively served SEO spam and redirects primarily to Googlebot—so site owners might not notice right away. The particularly eyebrow-raising detail is the reported use of an Ethereum smart contract to help resolve the attacker’s control infrastructure. That’s not the mainstream path defenders plan for, and it can complicate the usual playbook of domain takedowns. WordPress.org closed a large set of affected plugins and pushed an update to disable the phone-home behavior, but the bigger lesson is about governance: when plugin ownership changes hands, users often don’t get a clear, prominent warning—and that can hand an attacker a distribution channel with built-in trust. Google targets back-button hijacking Staying with web integrity, Google Search is tightening its spam policies around something many people have experienced but may not have named: “back button hijacking.” That’s when a site manipulates browser history so the back button doesn’t take you where you expect—sometimes bouncing you to ads or pages you never chose. Google says the tactic already violated its broader rules, but now it’s explicitly categorized under malicious practices, with enforcement slated to begin mid-June. Why it matters: this is one of those user-hostile tricks that can quietly spread through third-party scripts—ad tech, widgets, analytics helpers—and Google is signaling that “it wasn’t our code” won’t be a safe excuse if it degrades basic navigation and trust. GitHub adds stacked pull requests On the developer workflow front, GitHub is moving closer to what many teams already try to do informally: splitting big changes into reviewable slices without losing the thread. GitHub now has native support for stacked pull requests, plus a companion CLI extension called gh stack. The key point isn’t the mechanics—it’s the social and operational win. Smaller PRs tend to get reviewed faster, reduce merge conflicts, and make it easier to spot risky changes. GitHub is also trying to make stacks behave predictably with protections and CI that reflect what will happen when everything ultimately lands on the main branch. If your team struggles with “mega-PRs,” this is GitHub acknowledging that the platform should help enforce incremental delivery, not just host it. Jujutsu version control gains attention Related, there’s a thoughtful look at jj, the command-line interface for Jujutsu, a distributed version control system that’s aiming at a familiar audience: people who know Git, but don’t necessarily love it. The pitch is interesting because it attacks a long-running assumption in dev tooling—that power requires complexity. Jujutsu tries to keep the workflow approachable while still enabling advanced operations that can be awkward in Git. And the practical hook is compatibility: it can sit on a Git backend, meaning an individual can try it without forcing a team migration or rewriting history. Tools that offer “opt-in adoption” tend to get real-world experimentation, which is often how ecosystems shift over time. Diffusion LMs catch up to AR In AI research, a new paper on “Introspective Diffusion Language Models” is taking aim at a problem that’s limited diffusion-style language models: quality. Diffusion models are attractive because they can generate in a more parallel way, which hints at speed and throughput advantages—especially at high concurrency—but they’ve typically lagged behind standard autoregressive models on output quality. The researchers claim their approach, I-DLM, closes that gap by training for what they call introspective consistency—basically making the model’s internal scoring line up with the text it’s producing. They also describe an inference method that generates multiple tokens while verifying earlier ones in the same pass, and they emphasize deployment on common serving stacks rather than exotic infrastructure. If the results hold up broadly, this is one of the more credible signals that “faster LLMs” might not have to mean “worse LLMs,” which is the tradeoff the industry keeps running into. OpenDuck brings DuckDB to cloud For data and analytics folks, OpenDuck is an open-source project that’s trying to bring “DuckDB, but seamlessly remote” to a self-hostable world. The idea is that you can attach a remote database and treat those tables as first-class citizens alongside local data, with the system splitting work between local execution and remote DuckDB workers. Why it’s notable is the direction, not the branding: a lot of teams want the simplicity of local analytics with the reach of cloud storage and compute, but they don’t want to lock into a proprietary interface. OpenDuck is pitching a minimal, swappable protocol and an architecture that’s meant to make hybrid local-remote queries feel normal. If it matures, it could broaden the “portable analytics” story beyond a single vendor’s platform. Backblaze backup exclusions raise trust Now a cautionary tale about backups—because “set it and forget it” only works if the defaults are truly protective. A long-time Backblaze user says the service has been quietly excluding certain folders from backup, including .git directories and common cloud-sync folders like OneDrive and Dropbox. Backblaze’s rationale, as described in release notes, is performance and avoiding unintended uploads from sync caches or mount points, and that’s understandable. But the complaint is about trust and communication: if users believe “everything important is backed up,” silent exclusions can turn into a nasty surprise during a restore—exactly when you least want ambiguity. The larger takeaway is one worth repeating: sync is not backup, and backup software needs to be aggressively transparent when it decides something is out of scope. Franklin’s Apple II clone era A quick detour into computing history: a retro look at Franklin Computer Corporation’s early-1980s ad campaigns for its Apple II–compatible machines. The ads were memorable, sometimes flamboyant, and the products were competitive on price and features—but the story is inseparable from the cloning controversy. Franklin’s machines were described as extremely close to Apple II designs, and Apple ultimately prevailed in a legal fight that helped define how software and hardware IP would be treated in this era. It’s a snapshot of an industry that was still figuring out where innovation ended and copying began—and how marketing could sprint ahead of the legal system until it couldn’t. NimConf 2026 community timeline Finally, a community calendar note: the Nim team announced NimConf 2026, an online event planned for June with talks premiered on YouTube and live Q&A in chat. The bigger reason it matters isn’t the date—it’s that it sets a deadline-driven rhythm for the ecosystem. For smaller language communities, conferences can act like a forcing function: polishing libraries, writing up real-world case studies, and sharing what actually worked in production. If you track Nim at all, this announcement effectively starts the “what will we have to show by June?” clock. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: One operator for all math - An arXiv paper claims a single primitive operator, eml(x,y)=exp(x)−ln(y), plus the constant 1 can generate exp, ln, arithmetic, and famous constants like e and π—suggesting a new universal expression format for elementary math. Symbolic regression with EML trees - The same EML representation is used as a differentiable “circuit,” trained with gradient methods like Adam to recover exact closed-form formulas from data—pointing to a new, structured search space for symbolic regression and interpretable models. Engineering ROI and build economics - A software economics essay argues teams often ship features without quantifying value; it connects platform-team impact to hours saved per engineer and product impact to churn, activation, and conversion—highlighting ROI as a competitive edge in the AI era. AMD ROCm versus Nvidia CUDA - AMD says ROCm is becoming a cohesive, faster-shipping AI stack, betting that better portability and “just works” reliability can reduce CUDA lock-in and influence data-center GPU buying decisions. Faster constant integer division - A new compiler optimization for division by constants targets 64-bit CPUs more directly than the classic Granlund–Montgomery approach; LLVM has already merged the change, promising real speedups in hot code paths. Lean as a perfectable language - An essay argues Lean is special because you can express and prove properties about programs inside the language itself, combining dependent types, theorem proving, and metaprogramming—pushing safer refactoring and optimization. Why web UI feels inconsistent - A critique of modern web design says we lost shared interface idioms—checkboxes, menus, keyboard shortcuts—due to touch-first compromises and framework-heavy UI, urging a return to standard HTML and predictable browser behaviors. Hacker News builders: agents and privacy - The April 2026 “What Are You Working On?” thread shows builders focusing on AI coding agents, sandboxing and verification, plus local-first and privacy-oriented tools—capturing shifting priorities toward control and reliability. Open-source homemade soft drinks - A long-running DIY cola project documents reproducible food-chemistry tricks—emulsifying essential oils, tuning acids and sweeteners—and publishes versioned recipes on GitHub, showcasing open experimentation beyond software. - Paper Proposes a Single Binary Operator That Can Generate All Elementary Functions - Why Software Teams Often Lack Clear ROI—and Why AI Raises the Stakes - AMD Bets on ROCm Upgrades and Triton Support to Challenge Nvidia’s CUDA Moat - Blinry Shares Open-Source DIY Cola, Orange, and Almond Soda Recipes After Years of Iteration - Essay Calls for a Return to Consistent, Idiomatic Interface Design on the Web - boringBar Replaces the macOS Dock with a Window-Focused Taskbar for Spaces and Multi-Monitor Setups - Hacker News April 2026 Thread Highlights Surge in AI Agent Tools and Local-First Apps - New Compiler Technique Speeds Up 32-bit Constant Division on 64-bit CPUs - Essay argues Lean is a “perfectable” language combining dependent types, theorem proving, and metaprogramming Episode Transcript One operator for all math First up, a new arXiv paper that tries to make “elementary math” feel almost embarrassingly uniform. The claim is bold: with a single operator, eml(x,y)=exp(x) minus ln(y), plus the constant 1, you can construct the usual scientific-calculator toolkit—exp, ln, arithmetic, exponentiation—and even build constants like e, π, and i. The practical takeaway isn’t that you should write math this way tomorrow, but that it offers a surprisingly simple grammar for representing formulas: everything becomes the same binary-tree shape. If that holds up broadly, it could simplify how symbolic systems store, transform, and search over expressions, because you’re no longer juggling dozens of primitive node types—just one. Symbolic regression with EML trees What makes it more than a curiosity is the machine-learning angle: the paper treats these EML expression trees as differentiable circuits, then trains them with standard optimizers to fit numerical data and recover exact closed-form functions at relatively shallow depths. That’s interesting because symbolic regression usually fights two enemies at once: a gigantic search space and results that are hard to interpret. A constrained, uniform representation can act like a funnel—still expressive, but more structured—potentially making it easier to land on formulas humans can read, especially when the “true law” is genuinely elementary. Engineering ROI and build economics Shifting from math to management, there’s a piece arguing many engineering orgs build day to day without a clear view of the economics behind those choices. The author puts real numbers on the intuition: an eight-person team in Western Europe can cost on the order of tens of thousands of euros per month, and that implies an internal platform team needs to reliably save multiple hours per week per supported engineer just to break even—and more than that once you price in maintenance and the fact that not every initiative works. The point isn’t that platform teams are bad; it’s that “we shipped it” isn’t the same as “it paid off.” In a world where AI tools keep compressing development time, the essay argues headcount and sprawling codebases stop looking like moats and start looking like liabilities unless you can prove they’re buying you measurable outcomes—churn, conversion, activation, or hard cost savings. AMD ROCm versus Nvidia CUDA In AI infrastructure news, AMD is making the case that ROCm—its CUDA alternative—is graduating from a loosely connected toolkit into something closer to a cohesive product. In an interview, AMD’s AI software leadership talked about tightening release cadence, smoothing developer experience, and unifying acceleration across AMD hardware under a “OneROCm” umbrella. Why this matters is pretty simple: in data centers, software friction often decides hardware deals. If a stack “just works,” it reduces lock-in and makes price-performance shopping easier. AMD is also leaning on more open development practices and investing in higher-level tooling that can make portability less painful, with the big strategic goal being to turn GPU choice into a competitive market again, not a foregone conclusion. Faster constant integer division Now to compilers: an arXiv paper proposes a new way to optimize 32-bit unsigned integer division by constants on 64-bit CPUs. This is one of those unglamorous optimizations that quietly affects a lot of real software, because compilers constantly lower divisions like “x divided by 7” into faster sequences. The authors argue the classic approach commonly used in compilers doesn’t fully exploit 64-bit hardware, and they report meaningful microbenchmark speedups on both a high-end Intel Xeon and Apple’s M-series silicon. The most immediate reason to care: patches exist for both LLVM and GCC, and the LLVM change has already landed in mainline—so developers may simply get faster code out of future compiler releases without touching their source. Lean as a perfectable language On the programming-languages front, there’s an essay making the case that Lean stands out because it’s “perfectable”: you can write a program, state a property about it, and prove that property—inside the same environment, with machine checking. That matters less as a party trick and more as a vision for how software changes safely. If you can prove two pieces of code are equivalent, refactoring becomes less of a leap of faith, and optimization can be more aggressive without turning into a bug farm. The essay also highlights Lean’s metaprogramming and syntax extension as unusually practical, pointing toward a future where proving and programming blur together—especially as more developers look for stronger guarantees than tests alone can provide. Why web UI feels inconsistent There’s also a thoughtful critique of modern web UI: the argument is that we’ve lost the “idiomatic” consistency people took for granted in desktop software—predictable controls, standard menus, reliable shortcuts, and interfaces that don’t reinvent basic interactions. On the web, you see a different date picker, form pattern, and keyboard model everywhere, which forces constant relearning and breaks flow. The author pins the drift on mixed touch-and-desktop priorities, heavy component reuse that can spread bad patterns, and framework-driven front ends that bypass browser conventions. The practical message is refreshing: lean on standard HTML elements, keep labels clear, respect expected browser behavior, and build trust through predictability—because novelty in UI is often just friction with better marketing. Hacker News builders: agents and privacy From the community corner, Hacker News’ April 2026 “What Are You Working On?” thread is a snapshot of what builders are actually doing when no one’s writing a press release. Two themes stood out: AI-assisted development, especially agent workflows with guardrails like sandboxing and verification, and a strong tilt toward privacy and local-first tools—offline translation, on-device transcription, self-hosting, and systems that avoid constant cloud dependence. You also see a practical security streak: people exploring new ways to do remote access and networking under restrictive corporate environments. The broader signal is that AI is pushing productivity up, but it’s also raising the premium on reliability, safety, and keeping humans in control of their data and tools. Open-source homemade soft drinks Finally, a fun bit of open experimentation: a multi-year blog project documenting homemade soft drinks, including a sugar-free, caffeine-free cola, with iterative tweaks and recipes tracked in a public GitHub repo. It’s essentially “reverse engineering,” but in food chemistry—balancing flavor oils, keeping them mixed, dialing acidity, and comparing against commercial benchmarks. Why it fits in this feed is the ethos: reproducibility, versioning, and community improvement, applied outside software. It’s also a reminder that a lot of what we call engineering—measurement, iteration, and careful documentation—transfers cleanly to the physical world. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: AI benchmarks gamed by exploits - UC Berkeley researchers show popular AI agent benchmarks can be reward-hacked via environment leakage and weak validators, undermining leaderboard trust and safety claims. iPhone passcode broken by update - A student reports an iOS update blocks a Czech diacritic in a lock-screen passcode, highlighting fragile input methods, encryption constraints, and data-recovery pitfalls. JVM flags database for OpenJDK - An updated VM Options Explorer catalogs OpenJDK 11 HotSpot flags with defaults, deprecations, and vendor differences—useful for performance tuning and upgrade planning. Reversible computing and energy limits - A piece connects Landauer’s principle to reversible computation, explaining why reducing information erasure could lower energy use even if today’s hardware is far from the limit. Hard tech bets after Intel - Pat Gelsinger, now backing hard tech startups, outlines looming compute bottlenecks—memory, networking, energy—and why heterogeneous systems may shape the next AI era. Design for skimmers, not readers - The “Miller Principle” argues people rarely read docs, UI text, or long messages, pushing teams toward resilient product design and communication that survives skimming. Lean software ops without VC - A developer argues for ultra-lean infrastructure—simple deployments, low burn, pragmatic tooling—so profitable software can scale without constant fundraising pressure. Maintenance as engine of progress - Stewart Brand reframes maintenance, repair, and precision as drivers of scientific and industrial advances, shaping how institutions and cultures sustain innovation. Debating history’s biggest ideas - A blog’s chronological ‘greatest intellectual achievements’ list—Shannon, Darwin, computation, and more—sparks debate about what truly counts as a field-defining breakthrough. - VM Options Explorer updates searchable catalog of OpenJDK 11 HotSpot JVM flags - Blogger Compiles and Debates a Canon of History’s Biggest Intellectual Breakthroughs - Pat Gelsinger on Post-Intel Investing, 10,000x Inference Gains, and the Next Phase of Moore’s Law - Alex Miller’s ‘Miller Principle’: Assume People Won’t Read Your Text - Phyphox app showcases smartphones as tools for physics experiments - How Toffoli Gates Enable Universal Reversible Computing - iOS update blocks Czech háček in passcodes, locking some iPhone users out - Berkeley Researchers Show Top AI Agent Benchmarks Can Be Gamed for Near-Perfect Scores - Bootstrapped Founder Details a $20/Month Stack for Running Profitable SaaS Apps - Stewart Brand Argues Maintenance and Precision Drive Technological Progress Episode Transcript AI benchmarks gamed by exploits First up: a group at UC Berkeley says several widely used AI agent benchmarks can be “reward-hacked” to score near the top without actually doing the intended work. Their point isn’t that researchers are dumb—it’s that many eval setups accidentally leak answers, blur the boundary between the agent and the grader, or rely on brittle validation. That matters because benchmark numbers drive everything from model selection to funding to safety narratives. If scores can be gamed, the incentives drift toward manipulating measurement instead of improving real capability, and the public story about progress gets distorted. iPhone passcode broken by update Staying with software that behaves differently than expected: a student says an iOS update locked him out of his iPhone because the lock-screen passcode keyboard stopped accepting a specific Czech character. The key still appears, but the phone won’t actually input it during the “before first unlock” passcode entry. And because he didn’t have a cloud backup, the official recovery path—restore the device—means losing the photos and data he cares about most. The broader lesson is uncomfortable: security features like strong encryption make recovery genuinely hard, so small input-method changes can turn into catastrophic access failures for anyone using uncommon characters to strengthen passcodes. JVM flags database for OpenJDK On the developer-ops side, Chris Whocodes published a refreshed “VM Options Explorer” for OpenJDK 11 HotSpot—a searchable, normalized catalog of JVM flags with context like defaults, deprecations, and where each option lives in the code. The interesting part isn’t just that there are a lot of knobs; it’s that the page helps you see how flags evolve across JDK releases and across vendor builds. If you operate JVM services, this is exactly the kind of detail that can prevent a painful upgrade, where an old tuning flag suddenly turns into a warning—or worse, a startup failure—and you’re left wondering what changed and when. Reversible computing and energy limits Now to computing fundamentals: a readable explainer revisits Landauer’s principle—the idea that erasing information has an unavoidable energy cost—and contrasts it with reversible computation, which in theory can avoid that particular penalty. Even though modern hardware burns far more energy than the theoretical minimum, the argument is that “reversible” thinking can still guide practical efficiency gains. The piece also highlights the tradeoff: you often need extra scratch space and additional outputs to keep computations reversible. Why this matters right now is simple: as compute demand keeps climbing, energy efficiency is turning from a nice-to-have into a core constraint. Hard tech bets after Intel Speaking of constraints, former Intel CEO Pat Gelsinger is now backing hard-tech startups and used a recent interview to lay out where he thinks the next computing bottlenecks are forming. He’s betting on a heterogeneous future—systems mixing classic CPUs with AI accelerators and, eventually, quantum components—while warning that today’s AI growth is running into very real limits around memory, interconnects, and cluster reliability. He also frames energy supply as a strategic resource, not just an operating cost, and ties it to geopolitics and supply-chain resilience. Whether you agree with every prediction or not, it’s a useful map of where an industry veteran expects money, engineering talent, and policy attention to converge. Design for skimmers, not readers A lighter read with a serious takeaway: Alex Miller’s “Miller Principle” claims, bluntly, that no one reads anything—docs, UI text, long emails, even code comments. It’s tongue-in-cheek, but the product lesson is real: if your system only works when users carefully absorb instructions, it probably won’t work. Good design assumes skimming, distraction, and time pressure—and tries to make the correct action the easiest action. Lean software ops without VC On building sustainably, a developer wrote about getting turned down at a pitch night because investors couldn’t see why funding was needed—his products already make recurring revenue with very low infrastructure spend. The essay’s broader theme is anti-glamour: keep deployments simple, keep costs predictable, and avoid architectural choices that drag you into constant operational overhead. The reason this resonates is that it’s less about any single tool and more about a strategy: reducing burn rate buys you optionality—more time to learn what customers actually want, and less pressure to chase growth-at-all-costs. Maintenance as engine of progress Two culture-and-ideas stories to close. First, Stewart Brand argues in a new project that maintenance—repair, calibration, and upkeep—isn’t a footnote to progress; it’s one of the engines of progress. His through-line connects precision manufacturing, interchangeable parts, and the institutions that preserve practical know-how. It’s a useful reframing for tech culture, which often celebrates invention while undervaluing the steady work that keeps systems safe, reliable, and improvable. Debating history’s biggest ideas And finally, a blog post proposes an informal, chronological list of world-changing intellectual breakthroughs—using Claude Shannon’s information theory as an example of a foundational idea that most people benefit from without ever hearing about. The author’s real goal seems to be sparking debate about what counts as a true intellectual revolution, and what gets left out of the usual canon. In a moment when AI and computing dominate headlines, it’s a reminder that today’s “obvious” technologies often rest on yesterday’s quiet, abstract insights. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Firefox add-ons: 84k scrape - A developer scraped nearly all Firefox extensions and found phishing, mass spam, and permission bloat—plus nasty performance limits when add-ons scale. Keywords: Mozilla Add-ons, phishing extensions, PUA, ecosystem health, performance bottlenecks. Linux kernel rules for AI - The Linux kernel published guidance for AI-assisted patches, stressing GPL-2.0-only compliance, human responsibility, and transparency via an “Assisted-by” tag. Keywords: Linux kernel, AI coding assistants, licensing, SPDX, Developer Certificate of Origin. Searchable U.S. pardon database - Pardonned.com turns DOJ clemency pages into a searchable, open-source database, making it easier to audit claims and compare pardon patterns across administrations. Keywords: presidential pardons, clemency, scraping, SQLite, public accountability. DIY MacBook ergonomics mod - A blogger literally filed down sharp MacBook corners for wrist comfort, reflecting a growing “modify your own tech” culture centered on ergonomics and user agency. Keywords: MacBook unibody, ergonomics, hardware mod, customization, repair mindset. Connect Four strategy compression - WeakC4 claims a compact, search-free path to perfect play in standard Connect Four by compressing winning knowledge into reusable patterns. Keywords: Connect Four, weak solution, human-interpretable strategy, pattern library, game solving. One-dimensional chess versus AI - An online “1D-Chess” variant shows how strategic tension survives on a single line of squares, letting players test ideas against an AI. Keywords: chess variant, minimal rules, emergent complexity, puzzles, browser game. Internet Archive concert preservation - Volunteers are digitizing Aadam Jacobs’ massive concert-tape archive for the Internet Archive, preserving fragile pre-digital recordings of punk and indie history. Keywords: Internet Archive, tape preservation, live bootlegs, metadata, cultural history. Chimpanzee community breaks into war - Researchers report Uganda’s largest known wild chimp community split into lethal factions, offering unsettling clues about how group violence can emerge without human institutions. Keywords: Ngogo chimps, Kibale, coalition violence, social fragmentation, Science study. Artemis II splashdown and validation - NASA’s Artemis II splashed down safely after a lunar flyby, validating Orion’s reentry, heat shield, parachutes, and recovery workflow for future Moon missions. Keywords: Artemis II, Orion capsule, reentry blackout, splashdown, lunar program. - Blogger Files Down MacBook Corners for Comfort, Urges Tool Customization - WeakC4 Publishes a Search-Free, Compressed Winning Strategy for Connect Four - Starfling Landing Page Teases Tap-and-Release Mobile Game With Ad-Based Revives - Online 1D-Chess Game Revives Martin Gardner’s One-Dimensional Variant - Volunteers Digitize Aadam Jacobs’ 10,000-Concert Tapes Into an Online Archive - Pardonned.com Launches Searchable Database of U.S. Presidential Pardons - Developer Scrapes and Installs Nearly All Firefox Extensions, Exposing Add-on Ecosystem Risks and Performance Limits - Ngogo chimpanzees split into deadly eight-year conflict in Uganda, study finds - Artemis II Astronauts Return to Earth After Orion Splashdown Near San Diego - Linux Kernel Publishes Rules for AI-Assisted Contributions Episode Transcript Firefox add-ons: 84k scrape Let’s start in the browser ecosystem, where a developer set out to scrape, analyze, and even attempt installing nearly every Firefox extension listed on Mozilla Add-ons—around eighty-four thousand of them. The dataset alone is enormous, but the real story is what it revealed: everything from suspiciously permission-hungry add-ons to mass-published low-quality extensions, plus outright phishing attempts like lookalike wallet tools using sneaky characters in their names. The author reports that Mozilla removed several of the worst offenders quickly once flagged, which is good news—but it also underlines how easily scams can slip into busy marketplaces and still rack up real users. And then there’s the stress test: at extreme numbers, Firefox’s add-on management wasn’t just slow—it became borderline unusable, with constant rewrites to extension metadata, long freezes, and crashes. Most people will never install anything close to that, but these extremes act like a spotlight: they expose where platforms strain, and where moderation and ecosystem hygiene still need sharper tools. Linux kernel rules for AI Staying with software governance, the Linux kernel project has added formal documentation for how contributors should use AI coding assistants. The tone is basically: use whatever tools you want, but the rules don’t bend. If you submit code, you’re still accountable for it—especially around licensing. The guidance emphasizes GPL-2.0-only compatibility, correct SPDX identifiers, and an important detail: AI tools can’t “sign off” on patches. Only a human can certify the Developer Certificate of Origin. The interesting shift here is the push for transparency. Contributors are asked to label AI help with an “Assisted-by” tag that includes the tool and model version. It’s not anti-AI; it’s pro-traceability. For one of the most influential open-source projects on the planet, this is Linux drawing a bright line between assistance and responsibility—and making the paper trail clearer for everyone downstream. Searchable U.S. pardon database Next up, a civic-data project that’s getting attention for a simple reason: it makes a politically sensitive power easier to inspect. A Hacker News user launched Pardonned.com, a searchable database of U.S. presidential pardons and clemency actions compiled from the Department of Justice’s public pages. The motivation is straightforward: people argue about who got pardoned, for what, and when—and it’s surprisingly tedious to verify. By turning scattered pages into a searchable, browsable set—while publishing the code as open source—the project lowers the friction to audit claims and spot patterns. Commenters also pointed out the messy reality of public records: when the source text is vague, a database can’t conjure missing details like fines or time served. Even so, this is a meaningful step toward accountability-by-default, and it hints at what’s possible once comparison and visualization tools mature. DIY MacBook ergonomics mod On the hardware side, one personal blog post sparked a very particular kind of debate: should you physically modify your expensive laptop for comfort? Kent Walters says yes—and he proved it by filing down the sharp bottom corners of his MacBook to make it easier on his wrists. He argues that Apple’s aluminum unibody design makes crisp edges possible, but that users shouldn’t treat factory aesthetics as sacred if the device doesn’t fit their body. What makes this notable isn’t just the bravery of taking a file to aluminum. It’s the mindset: tools should adapt to humans, not the other way around. As more people embrace repair, customization, and ergonomics—whether it’s keyboard switches, trackball angles, or now laptop corners—this kind of “ownership means agency” attitude is becoming a real cultural current in tech. Connect Four strategy compression Now for game-solving—where the theme today is compressing complexity into something a human can actually hold in their head. A project called WeakC4 published a compact, “search-free” weak solution for standard seven-by-six Connect Four. In plain terms, it claims a first-player win if you follow its prescribed lines, without needing a giant database of every possible board position or a runtime engine doing deep search. The clever bit is packaging: instead of exhaustive lookup, it distills perfect-play knowledge into a small opening tree that feeds into reusable pattern-based continuations. Why it matters: this is a concrete example of turning heavy computation into a lightweight, interpretable artifact—something you can visualize, navigate, and maybe even memorize. It’s a reminder that “solved” doesn’t have to mean “inaccessible.” One-dimensional chess versus AI In a similar spirit—simple rules, surprising depth—there’s an online playable version of “1D-Chess,” a one-dimensional chess variant where the board is just a line. You play as White against an AI using only kings, knights, and rooks. It’s intentionally minimalist, but it raises a fun question: does reducing the board’s dimension reduce strategic complexity, or does it just concentrate it into different tactics? Projects like this matter because they’re approachable. They let you test ideas quickly, explore endgame-like situations without a full chessboard, and still run into that familiar feeling of “wait, this is trickier than it looks.” It’s a neat reminder that complexity isn’t only about size—it’s often about constraints. Internet Archive concert preservation Switching gears to digital preservation: a remarkable collection of live concert recordings is being cataloged and digitized for the Internet Archive. The tapes come from Aadam Jacobs, a Chicago music fan who secretly recorded Nirvana’s first Chicago show in 1989 and then went on to tape more than ten thousand concerts over four decades. Volunteers are now racing against the clock—because cassette and DAT tapes don’t last forever. The work is painstaking: real-time transfers on repaired legacy equipment, careful cleanup, metadata, and setlist detective work. This story matters because it captures a pre-digital cultural record that official channels often never preserved, especially for smaller venues and early-era performances. It also sits in that complicated zone where history, fandom, and copyright intersect—yet the noncommercial preservation angle is exactly what institutions like the Archive were built to support. Chimpanzee community breaks into war On the science front, researchers are reporting a grim development in the world’s largest known wild chimpanzee community at Ngogo in Uganda’s Kibale National Park: the group has split into factions and descended into a prolonged period of lethal violence. Since the final break in 2018, scientists documented dozens of targeted attacks, including multiple adult male deaths and many infant killings. Why it’s drawing attention is the implication that intense, warfare-like conflict can emerge from shifting social boundaries alone—without human-style institutions like politics, religion, or formal ideology. The study, published in Science, emphasizes pressures like resource competition and reproductive rivalry, plus disruptions that may have removed key social stabilizers. It’s not a neat analogy for humans—and the researchers are careful about that—but it is a stark window into how fragile social cohesion can be in complex communities. Artemis II splashdown and validation Finally, a major milestone in spaceflight: NASA’s Artemis II mission has returned, with the Orion capsule splashing down in the Pacific off San Diego after a roughly ten-day lunar flyby. The crew endured the expected high-speed reentry and a communications blackout as plasma built up around the heat shield, then hit a clean parachute sequence for what mission control described as a precise landing. The significance here is validation. Artemis II isn’t just a scenic lap around the Moon—it’s a full end-to-end test of systems you need to trust for future missions: heat shielding, parachutes, recovery procedures, and the operational rhythm of bringing a crewed spacecraft home. Each successful return shrinks the unknowns for what comes next in NASA’s Moon plans—and beyond. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Signal et notifications iOS - Une affaire fédérale montre que des notifications iOS peuvent conserver du contenu de messages Signal, même après suppression de l’app. Mots-clés: Signal, iPhone, notifications, forensique, confidentialité. Souveraineté numérique de l’État - La France accélère la réduction des dépendances extra-européennes avec des plans ministériels, des communs numériques et des migrations vers Linux et des outils interministériels. Mots-clés: DINUM, ANSSI, Linux, commande publique, souveraineté. Microsoft et pilotes Windows open source - Microsoft a suspendu des comptes de signature de pilotes utilisés par des projets open source, ralentissant potentiellement correctifs et mises à jour de sécurité. Mots-clés: Windows drivers, verification, WireGuard, VeraCrypt, supply chain. Standards d’intégration pour agents IA - Un débat oppose les “Skills” basés sur des CLIs et le Model Context Protocol (MCP) pour des intégrations IA plus portables, maintenables et sécurisées. Mots-clés: LLM, MCP, OAuth, outils, standard. Orion Artemis II: informatique résiliente - NASA décrit l’architecture “fail-silent” d’Orion pour Artemis II, conçue pour rester sûre malgré radiations, erreurs et pannes, sans possibilité de réparation. Mots-clés: Artemis II, Orion, tolérance aux pannes, logiciel critique, redondance. Porte quantique stable à atomes neutres - ETH Zurich annonce une porte quantique de type swap très stable sur des atomes de potassium, une avancée vers des calculateurs quantiques plus fiables à grande échelle. Mots-clés: qubits, atomes neutres, fidélité, bruit, scalabilité. Spécifications formelles pour jeux complexes - Un moteur de combat de Dungeons & Dragons a été spécifié formellement et testé par fuzzing, révélant des bugs subtils que les tests manuels ratent souvent. Mots-clés: Quint, model-based testing, state machine, règles, vérification. - La France lance un plan interministériel pour réduire les dépendances numériques extra-européennes - Microsoft suspends key open-source Windows developer accounts over verification policy - Inside Orion’s Fail-Silent, Eight-CPU Flight Computer for Artemis II - ETH Zurich Demonstrates Noise-Resistant Swap Gate for 17,000 Neutral-Atom Qubits - Keeper: Go-based encrypted secret store with bucketed policies, audit chains, and key rotation - Formal D&D Combat Modeling Uses Model-Based Testing to Catch Hidden Rules and State Bugs - David Mohl Says MCP Beats Skills for Real LLM Service Integrations - InstantSpaceSwitcher Offers Near-Instant macOS Space Switching Without Disabling SIP - GitButler raises $17M Series A to reinvent collaboration beyond Git - FBI reportedly recovered deleted Signal messages from iPhone notification storage Episode Transcript Signal et notifications iOS On commence par cette histoire qui fait grincer des dents côté vie privée. Dans un dossier judiciaire lié à des faits de vandalisme au Texas, un témoignage indique que le FBI aurait retrouvé du contenu de messages Signal supprimés, non pas en “cassant” Signal, mais en exploitant des données stockées par iOS autour des notifications. L’idée clé, c’est que même si l’app disparaît, certaines traces de notifications peuvent survivre — et si le contenu des notifications n’est pas masqué, cela peut devenir une source de reconstitution partielle. Pourquoi c’est important ? Parce que ça rappelle une réalité souvent oubliée: le maillon faible n’est pas toujours le chiffrement, mais l’écosystème autour — réglages de notifications, caches, et habitudes d’usage. En clair: la confidentialité dépend aussi de la surface “système”, pas seulement de l’app. Souveraineté numérique de l’État Autre sujet souveraineté et contrôle, mais cette fois côté secteur public français. Le gouvernement a réuni le 8 avril un séminaire interministériel, piloté notamment par la DINUM, pour accélérer la réduction des dépendances numériques de l’État à des solutions extra-européennes. Le signal politique est assez net: reprendre la main sur les choix technologiques, les risques, et — point crucial — la commande publique. Parmi les mesures évoquées, on parle d’une sortie progressive de Windows au profit de postes Linux, et de migrations d’agents vers des outils interministériels comme Tchap, Visio et FranceTransfert. L’enjeu dépasse le symbole: si l’État fixe un cap, des standards d’interopérabilité et un calendrier, il donne de la visibilité à l’écosystème européen, tout en réduisant les verrouillages et la dépendance aux décisions de fournisseurs externes. Microsoft et pilotes Windows open source Dans la même veine, mais côté industrie logicielle: Microsoft a suspendu des comptes développeurs utilisés pour signer et publier des pilotes Windows et des builds, touchant temporairement des projets open source très connus, y compris dans la sécurité. Des mainteneurs ont expliqué n’avoir eu ni avertissement clair, ni explication exploitable, et surtout la difficulté d’obtenir une réponse humaine. Microsoft a ensuite indiqué qu’il s’agissait d’une suspension automatique liée à une procédure de vérification obligatoire, et a promis de revoir la communication. Pourquoi ça compte ? Parce que l’open source, surtout quand il est au cœur de la sécurité, dépend de chaînes de distribution et de conformité qui peuvent s’enrayer d’un coup. Quand l’application des règles est automatisée et que le support est difficile à atteindre, le risque n’est pas théorique: un correctif critique peut se retrouver bloqué au pire moment. C’est un rappel brutal de la fragilité “administrative” de la supply chain logicielle. Standards d’intégration pour agents IA On passe à l’écosystème IA, où une bataille plus silencieuse se joue: comment standardiser l’intégration des services et outils dans les assistants et agents basés sur des LLM. Un développeur défend l’idée que la mode des “Skills” pour donner des capacités aux modèles est souvent mal orientée dès qu’elle repose sur des CLIs à installer et exécuter. Le problème, c’est la réalité des clients IA: sur mobile, dans certains environnements d’entreprise, ou dans des apps verrouillées, exécuter une CLI locale est impraticable. Et derrière, ça complique tout: mises à jour, secrets, métadonnées, et même la taille du contexte. À l’inverse, le Model Context Protocol — MCP — est présenté comme une approche plus robuste pour de vraies intégrations: une interface outillée, souvent distante, avec des mises à jour centralisées et une authentification plus propre. En toile de fond, c’est un choix d’architecture qui peut décider si l’IA “outillée” sera portable et maintenable, ou une collection fragile de scripts et de docs. Orion Artemis II: informatique résiliente Direction l’espace, où la fiabilité est une obligation, pas une préférence. NASA a détaillé l’architecture informatique de la capsule Orion pour Artemis II, en insistant sur un principe “fail-silent”: quand un composant part en vrille, il ne doit pas produire une sortie potentiellement dangereuse — il doit se taire. C’est une philosophie qui colle aux systèmes critiques où le logiciel pilote presque tout: contrôle du véhicule, routage des communications, et fonctions vitales. Ce qui rend le sujet intéressant, c’est le contexte deep space: radiation, erreurs aléatoires, et zéro possibilité de réparation. NASA mise donc sur la redondance, sur des exécutions strictement déterministes, et même sur un logiciel de secours “différent” pour limiter les pannes communes dues à un bug partagé. Pourquoi ça nous concerne au-delà de la Lune ? Parce que ces méthodes se transfèrent: aviation, médical, infrastructures — partout où “ça a redémarré” n’est pas une stratégie acceptable. Porte quantique stable à atomes neutres Côté quantique, ETH Zurich a présenté une porte de type “swap” particulièrement stable avec des qubits encodés dans des atomes neutres piégés. Le détail à retenir n’est pas la recette expérimentale, mais le résultat: une opération à deux qubits plus robuste au bruit, et donc plus fiable, tout en pouvant être exécutée en parallèle sur un grand nombre de paires. C’est précisément le goulet d’étranglement de beaucoup de plateformes quantiques: on sait aligner beaucoup de qubits, mais obtenir des opérations à deux qubits suffisamment propres, répétables, et industrialisables reste difficile. Si cette approche tient la route en s’intégrant à des systèmes de contrôle plus fins, elle renforce l’idée que les architectures à atomes neutres peuvent viser l’échelle sans se faire punir par des erreurs de contrôle à chaque étape. Spécifications formelles pour jeux complexes On termine sur un exemple inattendu de rigueur logicielle appliquée… à Dungeons & Dragons. Un développeur a étendu un modèle formel Quint jusqu’à un moteur de combat complet, avec toutes ces chaînes d’interruptions que les joueurs connaissent: réactions, contre-sorts, attaques d’opportunité, et retours au bon endroit dans le tour. Ce qui est fascinant, c’est la méthode: utiliser le modèle comme “oracle”, puis faire du model-based testing et du fuzzing contre une implémentation TypeScript pour trouver des divergences. Résultat: des bugs subtils et franchement non évidents sont ressortis tôt, là où des tests écrits à la main passent souvent à côté. Et la cerise sur le gâteau: transformer des milliers de questions-réponses de règles en assertions vérifiables, avec l’aide de LLMs pour générer des tests typés. Pourquoi c’est intéressant au-delà du jeu ? Parce que c’est une démonstration très concrète: quand les règles sont complexes et l’espace d’états explose, les specs formelles et les tests génératifs peuvent devenir un avantage décisif — y compris pour valider du code écrit par des agents. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Mac OS X on Wii - A developer got Mac OS X 10.0 “Cheetah” booting natively on a Nintendo Wii, highlighting retro PowerPC hacking, drivers, and creative bootstrapping. Linux app-level network monitoring - Little Snitch arrived on Linux with per-app outbound connection visibility and block rules, using eBPF for privacy-oriented traffic monitoring. Open-source supply-chain hardening - Astral shared a practical CI/CD defense-in-depth playbook—pinned GitHub Actions, least privilege, protected releases, and Sigstore—to reduce supply-chain risk. Thunderbird funding and sustainability - Thunderbird says fewer than 3% of users fund development, framing donations as key to keeping a privacy-first email client maintained without ads or data sales. Modern C# in Unity - Unity’s move toward modern .NET makes newer C# features more viable; the discussion covers pragmatic use of properties, tuples, records, and LINQ performance tradeoffs. Dr. Dobb’s archives preserved - An Internet Archive ISO preserves decades of Dr. Dobb’s and related programming journals, a searchable time capsule of software engineering history and source code. The Fifth Element VFX craft - A retrospective explains how The Fifth Element’s “futurescape” blended miniatures, CG, and matte work—an influential late-1990s VFX workflow snapshot. Old English dual pronouns - A linguistics feature revisits Old English dual pronouns like “wit” and why they faded, connecting pronoun change to clarity, simplification, and social history. - Objective Development Releases Little Snitch for Linux Using eBPF to Monitor and Block App Connections - Thunderbird Appeals for User Donations to Sustain Development - Developer Boots Mac OS X 10.0 Natively on a Nintendo Wii - Astral details its defense-in-depth approach to securing open-source builds and releases - How Digital Domain Built The Fifth Element’s Comic-Book Future with Miniatures and Early CG Pipelines - Old English’s lost dual pronouns and how English simplified them away - CSS Studio launches early access for visual CSS editing synced to source code via AI agent - RavensBlight Catalogs a Large Collection of Gothic Printable Paper Toys - Unity Developers Urged to Adopt Modern C# Features as .NET Support Improves - Internet Archive Hosts Dr. Dobb’s Developer Library DVD 6 Compilation Episode Transcript Mac OS X on Wii First up in privacy and visibility: Objective Development has released Little Snitch for Linux. If you’ve used the Mac version, you know the appeal—seeing which apps are reaching out to the internet, and choosing what gets through. On Linux, it brings that same “show me the outbound connections” mindset, with per-app views, traffic history, and rules that can block by domain or network. What makes this interesting isn’t just another firewall. It’s the app-level angle—useful for spotting surprising background chatter, misbehaving plugins, or that one tool that phones home more than you expected. The project is also partly open source, which will matter to folks who want to audit the UI and the eBPF pieces, even if the core daemon stays proprietary. The authors are candid about limits too: under heavy traffic, attribution can get fuzzy, and it’s framed more as privacy monitoring than a hard security boundary. Linux app-level network monitoring Staying on security, but shifting from endpoints to the build pipeline: Astral published a detailed look at how they harden the development and release process for their widely used open-source tools. The subtext here is the modern reality—software supply-chain attacks aren’t theoretical anymore, and CI systems can be an uncomfortable blend of powerful and easy to misconfigure. Astral’s write-up is valuable because it’s concrete: reduce risky workflow triggers, pin automation to known versions, keep permissions tight, and treat releases like a gated process rather than a casual “tag and ship.” They also lean into stronger provenance signals—things like attestations—and operational guardrails like two-person approvals for sensitive steps. Even if you don’t copy their exact setup, it’s a good snapshot of what “responsible maintainer” looks like in 2026, and where the ecosystem still makes secure defaults harder than they should be. Open-source supply-chain hardening Now, a story about sustainability in open source that’s less about exploits and more about economics: the Thunderbird team is asking users to donate, pointing out that fewer than 3% of users financially support the project. The pitch is straightforward: no ads, no selling user data, no corporate funding—just community money paying for servers, fixes, and the engineers who keep the lights on. Whether you personally use Thunderbird or not, the bigger takeaway is familiar: widely used, privacy-respecting software often survives on a surprisingly thin layer of support. When that layer shrinks, roadmaps shrink with it. Thunderbird funding and sustainability Alright—let’s get to the one that made a lot of people do a double-take. Developer Bryan Keller says he’s ported Mac OS X 10.0 “Cheetah” to run natively on a Nintendo Wii. This isn’t just “it boots to a logo.” It’s a long chain of unglamorous wins: a custom boot path, kernel patching to survive early startup, building enough hardware knowledge to provide the OS what it expects, and then writing drivers so the machine can actually find storage, show a usable display, and accept input. What makes it matter is the demonstration of portability—how far a legacy OS can be pushed when the CPU family lines up, and how much modern computing lore is still embedded in device trees, driver stacks, and assumptions about hardware. It’s also a reminder that “obsolete” platforms keep teaching new lessons—especially about what we’ve standardized, and what we still haven’t. Modern C# in Unity On the day-to-day developer side, there’s a thoughtful piece making the case that many Unity projects still look like they’re written for an older C# era—partly because Unity’s runtime history pushed people toward older patterns, and partly because tutorials tend to fossilize best practices. The interesting point is not “use every new language feature.” It’s that Unity’s move toward a more modern .NET stack changes the cost-benefit math. Features like properties, tuples, and records can reduce boilerplate and clarify intent, while performance-sensitive choices—like when to avoid LINQ—still matter in gameplay loops. If you’ve inherited a Unity codebase that feels stuck in time, this is a useful lens: modernize selectively, and do it for readability and maintainability first, with performance measurements where it counts. Dr. Dobb’s archives preserved For the archivists—and anyone who learned to code by reading old magazine articles—there’s a gem on the Internet Archive: an ISO image of “Dr. Dobb’s Developer Library DVD 6,” bundling decades of programming writing across multiple publications. Why it matters is simple: a lot of practical engineering knowledge lives in formats that quietly disappear—dead websites, lost attachments, link rot, broken search. Having a browsable, searchable snapshot of that era preserves more than nostalgia. It preserves context: how people explained C and C++, Unix and Windows, systems administration, Perl, and the everyday tradeoffs that shaped modern software culture. The Fifth Element VFX craft Two lighter, but still very Hacker News, reads to wrap up. One is a retrospective on how Digital Domain built the look of The Fifth Element. It’s a reminder of a particular late-1990s sweet spot: miniatures, early CG, matte work, and a production pipeline that was inventive because it had to be. The result still stands out—bright, layered, and oddly optimistic compared to the darker “future city” template that dominated the genre. If you like the craft side of tech—tools serving taste—this one’s worth your time. Old English dual pronouns And finally, a linguistics detour: a piece on Old English “dual” pronouns—words specifically for “we two” or “you two”—and how that category basically vanished by the 13th century. The fun part isn’t just trivia. It’s seeing language as an interface that gets simplified when the extra precision isn’t worth the complexity, and how social forces also shape what survives. Pronouns feel stable until you zoom out a few centuries—then they look a lot like versioned software with breaking changes. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: VeraCrypt blocked by code-signing - VeraCrypt maintainer Mounir Idrassi says Microsoft terminated his long-used Windows driver signing account with no warning or appeal, freezing Windows releases. Keywords: VeraCrypt, code signing, Windows drivers, Secure Boot, supply chain trust. Anthropic Glasswing AI security push - Anthropic’s Project Glasswing uses an unreleased Claude Mythos 2 Preview model with partners to find and help patch high-severity vulnerabilities faster than most humans. Keywords: AI security, zero-days, vulnerability research, critical infrastructure, disclosure. Read Git history before code - A developer argues the fastest way to assess a new codebase is to start with Git history to spot churn, ownership concentration, bug hotspots, and risky revert patterns. Keywords: Git, churn, hotspots, maintainability, engineering risk. Markdown notes as context graph - A plain Markdown knowledge base with Obsidian-style links can behave like a lightweight graph database that improves AI outputs by preserving project decisions and context. Keywords: knowledge base, Markdown, wikilinks, context engineering, LLM. Side projects fuel engineering growth - An essay connects enterprise “skyscraper” engineering discipline with “shed” side projects, arguing personal building time prevents burnout and sharpens judgment. Keywords: side projects, burnout, systems thinking, experimentation, career. Open-source agent shifts ownership - Mario Zechner moved the open-source coding agent ‘pi’ under Earendil, keeping an MIT core while signaling potential paid or proprietary add-ons. Keywords: open source, governance, licensing, MIT, sustainability. DIY robot vacuum learns poorly - Two roommates built a low-cost robot vacuum and trained a behavior-cloned CNN, but real-world navigation stayed inconsistent—highlighting how data quality beats model tweaks. Keywords: robotics, behavior cloning, dataset quality, autonomy, edge cases. Music-reactive LEDs meet mel scale - A decade-long open-source LED music visualizer improved by mapping audio to the mel scale, making limited LED strips feel more ‘musical’ to humans. Keywords: DSP, mel scale, visualization, perception, LEDs. Revision 2026 demoscene highlights - A full recording of Revision 2026’s PC Demo compo preserves the demoscene’s latest real-time graphics craft and a much-discussed finale for wider viewing. Keywords: demoscene, real-time rendering, PC demo, archive, creative coding. Artemis II far-side Moon photos - NASA released Artemis II lunar flyby images, including a striking in-space solar eclipse and far-side views, underscoring Orion’s crewed deep-space readiness. Keywords: NASA, Artemis II, Orion, Moon far side, eclipse imagery. - Five Git History Checks to Spot Codebase Risk Before Reading the Code - VeraCrypt Developer Says Microsoft Terminated Signing Account, Blocking Windows Releases - Mario Zechner Joins Earendil and Transfers pi Coding Agent Project - Revision 2026 PC Demo Compo Video Archives Full Competition Lineup - Anthropic Launches Project Glasswing to Use Frontier AI for Defensive Software Security - How a Markdown Vault and Wikilinks Turn Your File System into a Graph Knowledge Base - NASA Releases First Artemis II Photos From Lunar Flyby, Including In-Space Eclipse - DIY Robot Vacuum Project Struggles with Behavior-Cloning Navigation Reliability - Why Building a Good Audio-Reactive LED Strip Requires Perceptual Audio and Lighting Models - Dylan Butler’s Case for Protecting Personal Side Projects in Software Engineering Episode Transcript VeraCrypt blocked by code-signing Let’s start with VeraCrypt. Its developer, Mounir Idrassi, says he’s been knocked off course for months, and the biggest reason is blunt: Microsoft terminated the account he used to sign Windows drivers and the bootloader. No warning, no explanation, and—according to him—no appeal. That’s not just an admin headache; it effectively blocks shipping new signed Windows builds, and Windows is where most VeraCrypt users are. The bigger takeaway is how fragile critical open-source infrastructure can be when distribution depends on a single vendor-controlled trust gate. Even if the code is fine, the ability to deliver updates can disappear overnight. Anthropic Glasswing AI security push Staying in security, Anthropic announced Project Glasswing, a partnership program that uses an unreleased model—Claude Mythos 2 Preview—to hunt vulnerabilities in critical software. Anthropic claims the model is already turning up large numbers of high-severity issues across things like operating systems and widely used components, and they’re emphasizing delayed disclosure so fixes land before details go public. Why this matters: the cost of finding serious bugs is dropping, and the speed is going up—meaning defenders need to scale their response just to keep pace. Whether or not every claim holds up under scrutiny, the direction is clear: AI is accelerating both discovery and exploitation, and “secure by default” is no longer a slogan—it’s survival. Read Git history before code On the engineering-practice side, there’s a sharp suggestion for anyone onboarding to a new codebase: start with Git history before you even open files. The idea is to use commit patterns as a quick risk map—high-churn areas that keep changing, places where fixes cluster, and sections with concentrated ownership where one person carries the context. You’re not trying to learn everything from commits; you’re trying to figure out where the codebase is fragile, where delivery is chaotic, and where questions will cost you the most time. It’s a pragmatic reminder that software has a social history, and the repo already tells you who’s been firefighting. Markdown notes as context graph That theme—context as leverage—also showed up in a piece arguing that a personal knowledge base made of plain Markdown notes and simple wikilinks is basically a graph database you can carry around. The pitch isn’t “make a prettier wiki.” It’s “make better inputs” for future work, especially when you’re asking an AI assistant to draft a design doc, a handoff, or a recap. When notes are linked by people and projects, you stop relying on vague memory and start pulling in the actual trail of decisions. The unresolved hard part is still the boring one: consistently processing the inbox of stuff you capture so it doesn’t become a junk drawer. Side projects fuel engineering growth Another essay drew a clean line between big-company engineering and side projects—and then argued they’re not opposites at all. At work, you learn to build like a skyscraper: reviews, tests, and careful changes because failures are expensive. At home, you build like a shed: fast experiments, low stakes, and the freedom to try weird tools. The point is that each reinforces the other. Personal tinkering keeps curiosity alive and prevents you from becoming “just a ticket closer,” while professional discipline can gradually upgrade your hobby code from a pile of hacks into something resilient. It’s a quiet career strategy: protect your ability to build for yourself. Open-source agent shifts ownership In open-source governance news, Mario Zechner announced he’s bringing his coding agent project, pi, under a company umbrella—Earendil—while keeping the core MIT-licensed. The headline here isn’t a feature release; it’s expectations. Ownership moves, trademarks and repos shift, and the project’s future may include paid features or proprietary components alongside an open core. For users and contributors, these transitions can be either stabilizing or disruptive, depending on how transparent the roadmap stays and how well the community’s trust is handled. Sustainability is real—but so is the risk of a project’s center of gravity changing. DIY robot vacuum learns poorly On the DIY AI front, two roommates tried building a low-cost robot vacuum and trained it with behavior cloning—basically teaching by example from recorded teleoperation. The result worked, but not reliably: oscillations, random reverses, and occasional confidence in the wrong direction. Their takeaway is familiar to anyone who’s shipped ML into the physical world: model tweaks can’t rescue messy demonstrations and ambiguous training signals. Robotics is where “it kinda works” can still be frustrating, because your mistakes bump into furniture. Music-reactive LEDs meet mel scale For makers who want tech that feels good—not just accurate—there’s a long-running open-source LED music visualizer story that landed on a surprisingly human insight. With so few LEDs, you can’t afford a display that technically reflects the spectrum but looks meaningless. The author’s breakthrough was mapping sound using the mel scale, which better matches how people perceive pitch and musical change. The larger lesson is that perception is part of the spec. If your output is for humans, the right transformation is often the one that matches what we notice, not what the raw data says. Revision 2026 demoscene highlights On the creative computing side, a full recording of Revision 2026’s PC Demo competition went up, preserving the live screening and commentary. Even if you’re not deep in the demoscene, these archives are valuable: they’re a time capsule of what real-time graphics and audiovisual synchronization look like when they’re pushed as art, not product. And as always with demoparties, the conversation isn’t just “how did they do it,” but “what did it feel like,” especially when a finale lands as a shared cultural moment for the community. Artemis II far-side Moon photos And finally, space: NASA released the first images from the Artemis II crew’s lunar flyby, including far-side views and a striking in-space solar eclipse—where the Moon is backlit and you still see faint surface detail from Earthlight. Beyond the beauty shots, this is a signal that Orion and crewed deep-space operations are getting real-world validation. Artemis II is fundamentally a test mission, and these releases are the public-facing proof that the mission is executing the kind of navigation, imaging, and operations you need before attempting the next steps toward sustained lunar presence. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Apollo computer bug, found late - A newly spotted Apollo Guidance Computer IMU error-path bug could deadlock gyro operations—an eye-opener for software reliability, formal specs, and legacy code review. AI chatbots and cognitive diversity - USC researchers warn LLM chatbot ubiquity may homogenize writing and reasoning, narrowing linguistic variety and shaping opinions via biased “credible” norms and WEIRD-trained outputs. Console hacking and secure boot - A historical look at console security shows how secure boot, code signing, and defense-in-depth still get undermined by implementation flaws—NES lockouts to Switch BootROM and PS exploits. Email worms and accidental exposure - A developer explains how early-2000s mass-mailing worms scraped an email address from Counter-Strike map files, illustrating malware-era email fragility and metadata leakage. GPU history versus real usage - An interactive timeline charts GPUs that shaped PC graphics, then contrasts flagship hype with Steam survey reality—most gamers stick with midrange cards despite rapid silicon progress. Idiocracy satire as a barometer - The “Are We Idiocracy Yet?” site uses humor to map real events to the film’s themes—politics-as-spectacle, corporate creep, and algorithmic outrage—acting as a cultural trend gauge. Japan rice farming and policy pressure - A firsthand account of running a small rice farm near Shuzenji highlights water management, wildlife pressure, aging farmers, and the economic forces squeezing small-scale agriculture in Japan. Brutalist concrete laptop stand build - A maker’s brutalist concrete laptop stand blends utility with deliberate decay aesthetics, reflecting the crossover between functional desk gear and craft-driven industrial design. - Satirical ‘Idiocracy Proximity Index’ Says Reality Is Closing In on the Movie - USC researchers warn AI chatbots could homogenize human writing and reasoning - Researchers Uncover 57-Year-Old Lock Leak Bug in Apollo Guidance Computer Gyro Code - Maker Builds Brutalist Concrete Laptop Stand with Power, USB, and Plant Pot - Locker Launches as Open-Source, Self-Hosted Alternative to Dropbox and Google Drive - Interactive Timeline Maps 30 Years of Influential GPUs and What Gamers Actually Use - Counter-Strike Map Credits Helped Worms Overrun a BT Email Address - Working a Family Rice Farm in Rural Japan Reveals Growing Pressures on Small-Scale Farming - A Brief History of How Game Consoles’ Security Was Built—and Broken Episode Transcript Apollo computer bug, found late Let’s start with that Apollo surprise. A new write-up argues there’s a previously unnoticed bug in the Apollo Guidance Computer’s inertial measurement unit code: a resource lock can stay stuck if an error hits at just the wrong moment. The intriguing part isn’t just the hypothetical impact—it’s how it was found. The author used a spec-driven approach, with help from an LLM, to make “this must be released” obligations explicit. The broader lesson lands hard: even heavily scrutinized, mission-critical code can hide mundane failure modes like resource leaks, and you sometimes need new analysis lenses—not more eyeballs—to flush them out. AI chatbots and cognitive diversity Staying with AI, researchers at USC Dornsife are worried about a subtler long-term effect of ubiquitous chatbots: homogenization. Their argument is that when billions of people draft, polish, and even think through problems using the same small set of models, individual voice and creative ownership can blur, and shared norms of what sounds “correct” can narrow. They also raise a social concern: if models carry cultural and data skews, repeated interaction can gently steer users toward the model’s framing. It matters because the risk isn’t just wrong answers—it’s a slow compression of linguistic and cognitive diversity, especially in schools and workplaces that standardize around one tool. Console hacking and secure boot On the security front, there’s a solid history piece tracing how game consoles went from basically open boxes to locked-down computers—and how attackers kept pace anyway. Each era repeats the same rhythm: vendors add checks, researchers find a crack, and the next generation gets more complex. What stands out is the recurring moral: no single trick—lockout chips, disc checks, secure boot, code signing—stays magical. The real differentiator is disciplined security-by-design and defense-in-depth that matches an honest threat model, because implementation mistakes and overlooked edge cases keep reappearing. Email worms and accidental exposure Another throwback hits email, and it’s a great reminder of how messy the early internet was. A game developer recounts having to ask BT to effectively “blackhole” an email address back in 2002, because mass-mailing worms were flooding it nonstop. The twist is why that specific inbox got nuked: the address was embedded in Counter-Strike map files, so worms that scraped text-like files on PCs could harvest it at enormous scale. It’s a story about collateral damage—spoofing, bounces, abuse complaints—and how small design choices, like shipping an address in widely distributed content, can become an attack multiplier. GPU history versus real usage For hardware nostalgia with a reality check, an interactive project called “Every GPU That Mattered” walks through landmark graphics cards across about three decades. Beyond the timeline fun, the punchline is the adoption gap: the cards that dominate marketing aren’t necessarily what people run. Using Steam survey data, it underscores that mainstream, older, midrange GPUs still carry huge share while flagship parts remain comparatively rare. Why it matters: performance narratives often track the top end, but software ecosystems—and optimization decisions—live and die by what’s actually installed. Idiocracy satire as a barometer Switching to culture as signal, a satirical site titled “Are We Idiocracy Yet?” claims to track how closely current events resemble Mike Judge’s film, complete with an “Idiocracy Proximity Index.” It’s funny on the surface, but it’s really a curated critique: politics as spectacle, brands acting like aggressive trolls, corporate reach expanding into roles people expect from public institutions, and online incentives that reward outrage and extremes. Whether you buy the scoring or not, it’s a useful barometer for what people feel is slipping: institutional trust, shared reality, and the line between governance and entertainment. Japan rice farming and policy pressure One of the more grounded, non-software stories is a firsthand account of helping run a small rice farm in Japan for a season. It gets into the practical grind—water management, keeping wildlife out, fixing leaks in connected paddies—but then zooms out to the bigger picture: aging farmers, small plots, policy history, and the economics that make scaling hard. It’s relevant to tech people because it’s a reminder that “systems” aren’t only digital. Food production has bottlenecks, incentives, and legacy constraints too—and the failure modes are just less forgiving. Brutalist concrete laptop stand build And finally, a maker project that’s very Hacker News: a heavy concrete laptop stand inspired by brutalist architecture, complete with intentional “damage” and weathered details—yet still functional with built-in power and charging. The appeal here is the blend of utility and aesthetic contrarianism: designing with imperfections on purpose, and turning industrial materials into desk gear that feels like a tiny piece of urban infrastructure. It’s a small example of a bigger pattern: as commodity accessories get boring, craftsmanship and personality become the differentiator. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Kenya’s ant queen smuggling - Kenya is battling illegal exports of giant African harvester ant queens to overseas pet markets, raising biodiversity, invasive-species, and enforcement concerns. Keywords: wildlife trafficking, formicariums, CITES gap, ecosystem engineers. On-device AI on iPhone - Google’s AI Edge Gallery for iPhone highlights a shift toward offline, on-device LLM inference for privacy and lower latency, with a community angle for tools and add-ons. Keywords: edge AI, Gemma 4, offline inference, privacy. Self-hosted voice-and-vision assistant - Parlor shows real-time voice-and-vision chat with an AI running fully on a personal computer, signaling that local assistants are getting practical without cloud bills. Keywords: local multimodal, WebSockets, TTS streaming, privacy-preserving AI. Training a tiny LLM - GuppyLM is a reproducible, from-scratch training example for a tiny transformer, emphasizing transparency over fancy tricks and making LLM education more accessible. Keywords: open source, small model, synthetic data, tokenizer, transformer. Windows desktop UI strategy churn - Jeffrey Snover argues Microsoft’s desktop GUI story has fragmented for decades, eroding developer trust and leaving a confusing landscape of overlapping frameworks. Keywords: Win32, WPF, UWP, WinUI, platform churn. AR project rescue gone wrong - A contractor’s AR Beijing bus-tour sprint became a cautionary tale about broken engineering practices, demo-driven leadership, and the limits of contracts when payment enforcement is weak. Keywords: AR calibration, production chaos, unpaid invoice, client sign-off. France repatriates gold reserves - The Bank of France moved the last portion of its New York-held gold into Paris via replacement purchases, boosting profits amid high prices and reshaping reserve logistics. Keywords: central bank, bullion, repatriation, capital gain. Open-source moon-bounce radio - Moon RF is proposing an open-source path to ‘moon-bounce’ amateur radio using modern phased-array ideas, potentially lowering barriers to space-adjacent experiments. Keywords: moon-bounce, software-defined radio, phased array, open hardware. - Consultant Says AR Bus Project Collapse Ended in $35,000 Unpaid Invoice - GuppyLM open-sources a 9M-parameter “fish” language model built from scratch - Bank of France repatriates last New York-held gold, books €13 billion gain - Jeffrey Snover: Microsoft’s Windows GUI Frameworks Lost Coherence After the Petzold Era - Google launches AI Edge Gallery for iPhone with offline Gemma 4 support - Moon RF Rebrands and Details Open-Source Phased Arrays for Moon-Bounce Radio - PostHog announces broad hiring push across engineering, AI, and go-to-market roles - Kenya cracks down on smuggling of prized harvester ant queens - Parlor open-sources an on-device, real-time voice-and-vision AI assistant - Corsair Repositions Drop.com as Hub for Licensed Gaming Hardware Collaborations Episode Transcript Kenya’s ant queen smuggling First up, a story that sounds niche until you realize it’s a blueprint for how modern wildlife trafficking works. Kenya is seeing a surge in illegal exports of giant African harvester ant queens—collected during swarming season and sold abroad to hobbyists who keep ants as exotic pets. Officials have now tied major airport seizures to organized trafficking, and researchers warn that removing queens isn’t just “taking a few bugs.” It can knock out entire colonies that help shape grasslands by moving soil and spreading seeds. The other twist is regulatory: ants aren’t covered by CITES, so cross-border tracking is patchy, and online sales are hard to police. It’s a reminder that biodiversity risk isn’t only about elephants and ivory—tiny species can be lucrative, and the rules haven’t caught up. On-device AI on iPhone Now to the day’s biggest cluster: AI shifting from the cloud to your own device. Google has released an iPhone app called AI Edge Gallery that runs open-source models locally, and it just added support for Gemma 4. The headline isn’t that another app exists—it’s what it signals. On-device inference is becoming good enough that mainstream companies are treating it as a real platform, not a demo. The pitch is straightforward: faster responses, less reliance on connectivity, and more privacy because your prompts and images don’t have to leave the phone. If this keeps improving, it changes the economics too—fewer paid API calls, more experimentation, and more pressure on cloud-only AI products to justify why your data should travel. Self-hosted voice-and-vision assistant In the same on-device vein, an open-source project called Parlor is drawing attention for real-time voice-and-vision conversation with an AI that runs entirely on your own machine. What’s interesting here is not a new chatbot personality—it’s the interaction model. You can speak naturally, interrupt the assistant mid-sentence, and get streaming audio back quickly enough to feel like a conversation rather than a turn-based interface. The practical takeaway is that local assistants are starting to cross a threshold: they’re not just private, they’re usable. For developers, that means you can prototype voice experiences without signing up for cloud bills or building an entire backend. For users, it’s a glimpse of a future where “AI assistant” doesn’t automatically mean “upload everything to someone else’s servers.” Training a tiny LLM And if you want to understand how models like this get built—without the mystique—there’s a charming teaching project called GuppyLM. It’s a tiny, from-scratch language model that roleplays as a fish, speaking in short, lowercase lines about tank life. The fish gimmick is fun, but the real value is educational: the repository walks through the whole pipeline, from generating a dataset, to training a tokenizer, to running a basic transformer and chatting with it. The author intentionally avoids modern bells and whistles so the system stays readable and reproducible. In a world where “AI” often means “trust the black box,” small projects like this matter because they let more people learn the fundamentals end-to-end—and that ultimately makes the ecosystem healthier. Windows desktop UI strategy churn Switching gears to developer platforms, Jeffrey Snover—best known for PowerShell—has a blunt critique of Microsoft’s Windows desktop GUI story: it’s been decades of churn without a consistent path developers can trust. His argument is less about any one framework being bad and more about repeated strategic pivots. Each generation arrives with big promises, then gets deprioritized or replaced, leaving teams stuck maintaining long-lived apps while guessing what Microsoft will love next year. The result is today’s messy reality: multiple overlapping “official” options, plus a steady rise in third-party approaches like Electron because, for many teams, predictability beats elegance. The broader point is that platform success isn’t just APIs—it’s stability, migration paths, and not making developers feel like yesterday’s bet is today’s dead end. AR project rescue gone wrong Next, a painful real-world account from someone recruited to rescue an augmented-reality bus tour project in Beijing. He expected a month-long sprint. He walked into something closer to an emergency room: inexperienced developers pushing straight to production, shaky workflows, and core technical issues like AR calibration and unreliable location signals undermining the entire experience. He tried to push for foundational fixes—basic engineering discipline, realistic sign-off points, and stabilizing the pipeline—but leadership kept steering toward flashy demos instead of addressing the root problems. Then comes the part that lands hardest: after long days, using his own equipment, and covering expenses, he says he never received the remaining payment—tens of thousands of dollars—despite repeated acknowledgments that the debt existed. The lesson isn’t “contracts are useless,” but it is sobering: enforcement can be weak, especially across borders, and sometimes the real skill is recognizing early when a team doesn’t want help—they want a miracle. It’s also a reminder for clients: confidence is cheap, competence is rare, and the difference often shows up only after the bill arrives. France repatriates gold reserves On the macro side of trust and institutions, the Bank of France has moved the last portion of its gold that was still held in New York back into Paris—at least in practical terms. Instead of shipping and refining older bars, the bank effectively swapped: it sold what it had held in the U.S. and bought modern standard bullion in Europe, now stored in France’s vaults. Officials say it wasn’t political, more logistical and audit-driven, but the timing is notable: soaring gold prices translated into a major capital gain and helped flip the bank’s annual results. Why does this matter to a tech audience? Because it’s a reminder that “where assets live” is a strategic question—whether we’re talking about gold, data, or compute. Geography, standards, and control still matter. Open-source moon-bounce radio Finally, a space-adjacent project for the radio nerds—and I mean that as a compliment. Moon RF is proposing an open-source hardware-and-software approach to make “moon-bounce” communication more accessible. Traditionally, bouncing a signal off the Moon has been the domain of well-funded amateurs with big antennas and careful tracking. The pitch here is to use modern phased-array techniques and software-defined radio ideas to lower that barrier, so more people can experiment with long-distance links, space-style signal work, and even some radio astronomy-adjacent tasks. If it succeeds, it’s another example of a broader theme today: capabilities that used to require institutions and large budgets are gradually getting packaged into systems that curious individuals can build, study, and improve. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Rust tail calls beat assembly - Rust nightly’s new tail-call support helps an interpreter outrun hand-tuned ARM64 assembly on Apple Silicon, spotlighting real-world compiler progress and portability tradeoffs. AI shifts developer workflows - Developers are using AI agents for rapid prototyping, parser generation, and editor tooling, but stories this week emphasize human-led architecture, testing, and maintainability as the difference-maker. Token pricing for coding AI - OpenAI’s Codex moves toward clearer token-based metering, changing how teams forecast costs across input-heavy prompts, cached context, and output-heavy code generation. New languages and safer runtimes - From a Rust-like language that compiles to Go to a Lisp garbage collector that scans stacks and registers, projects are pushing safer abstractions while confronting messy platform realities. Email data sharing controversy - A blogger traces a unique email address back to an alleged sales-intel data pipeline, raising privacy, consent, and potential GDPR compliance questions for SaaS user ecosystems. Sauna heat and immune response - A controlled Finnish sauna session shows short-lived spikes in white blood cells without broad cytokine shifts, adding nuance to how heat stress may influence inflammation markers. Artemis II views Moon’s far side - NASA’s Artemis II crew reports their first direct view of the Moon’s far side, a high-visibility validation of Orion’s deep-lunar trajectory ahead of future landing ambitions. - Artemis II astronauts glimpse the Moon’s far side and photograph Orientale basin - Developer ships SQLite devtools after AI-assisted build—and warns about the design tradeoffs - Rust Nightly’s `become` Enables a Fast Tail-Call Interpreter for Uxn, With Mixed x86 and WASM Results - GitHub project adds ‘caveman mode’ to Claude Code to cut token usage - OpenAI shifts Codex pricing guidance to token-based credit rates during plan migration - Study finds sauna session boosts circulating immune cells more than cytokines - Lone Lisp Grows Up: Adding Conservative Stack and Register Scanning to Its Garbage Collector - Blogger alleges BrowserStack user emails surfaced in Apollo.io database - Nanocode project offers an end-to-end recipe for training a tool-using Claude Code–style model - Lisette introduces a Rust-inspired language that compiles to Go Episode Transcript Rust tail calls beat assembly Let’s start in systems land, because this one is genuinely eye-catching. Matt Keeter reports a new interpreter backend for his Uxn CPU emulator that leans on Rust nightly’s freshly added `become` keyword to guarantee proper tail calls. The headline result: on ARM64—specifically an M1 Mac—the new approach beats his earlier Rust interpreter and even edges out his hand-tuned ARM64 assembly on benchmarks like Fibonacci and Mandelbrot. Why it matters is less about one emulator and more about what it signals: safe Rust can sometimes reach “assembly-class” speed when the language and compiler finally get the control-flow guarantees that interpreters thrive on. It’s not a universal victory lap, though. On x86-64, performance is mixed, and on WebAssembly the approach falls flat across engines. That’s a useful reminder that a clever dispatch strategy can be very architecture- and runtime-dependent—and that brand-new compiler features often arrive before the ecosystem has fully learned how to optimize them everywhere. AI shifts developer workflows Staying with language internals, there’s a thoughtful write-up on evolving the “Baby’s First Garbage Collector” in a small Lisp implementation called Lone Lisp. The core problem is classic: once values cross the boundary into native code, the language runtime can lose track of what’s still “alive,” and the collector may reclaim objects that are actually still referenced—leading to crashes and truly weird corruption. The fix here is pragmatic rather than pristine: a conservative collector that scans the native stack for anything that looks like a pointer into the heap, and then goes a step further by spilling CPU registers so those hidden references become visible during collection. The tradeoff is that conservative GCs sometimes keep garbage longer than necessary, but the upside is robustness. If you’ve ever wondered why “just write a GC” becomes a multi-week detour, this is why: real programs don’t stay neatly inside the runtime’s walls. Token pricing for coding AI On the “new language” front, Lisette is an interesting attempt to bring Rust-like safety and expressiveness to developers who still want Go binaries and Go libraries. It compiles to Go, but tries to enforce a stricter style: immutability by default, algebraic data types, pattern matching, and an explicit push away from nil toward Option- and Result-like error handling. Why it’s notable is the bet it’s making: instead of trying to replace Go’s ecosystem, it tries to layer a different set of semantics on top of it. If it succeeds, it could appeal to teams that like Go’s deployment story but want more help from the type system to prevent the everyday foot-guns. New languages and safer runtimes Now to AI and developer tooling, where today’s theme is “speed, with a receipt.” Developer Lalit Maganti released syntaqlite, a parser-based foundation for SQLite tooling—formatters, linters, editor integrations—the unglamorous but essential stuff. The story isn’t just the release; it’s the process. Maganti argues AI agents are a major force multiplier for implementation work and the last-mile grind—tests, bindings, docs, playgrounds—while being genuinely risky for architecture and API design. One detail that lands: an early AI-driven “vibe-coding” attempt produced something that worked, but was fragile and hard to reason about, and it got scrapped. The rewrite kept tighter human control and added more checks. That’s a pattern we’re seeing repeatedly: AI accelerates output, but it doesn’t automatically create a system you can live with six months later. Email data sharing controversy If you’re more interested in training these coding agents than using them, there’s also a small JAX-based project called nanocode that tries to show an end-to-end path toward a tool-using coding assistant—think models that can read files, edit them, and run commands rather than only chat. It leans on synthetic tool-use conversations, supervised fine-tuning, and preference optimization to shape behavior. The significance here is accessibility. Even if you never train your own model, projects like this turn the “secret sauce” into something you can experiment with—dataset design, alignment choices, and how you structure tool calls so the model learns multi-step workflows instead of single-shot answers. Sauna heat and immune response A smaller but very practical idea making the rounds: an open-source “caveman” mode for Claude Code that forces ultra-compressed responses. The claim is big token savings—on the order of cutting output length dramatically—while preserving technical accuracy, code blocks, and exact error messages. Why this matters is that it treats style as a performance knob. For teams feeling latency or cost pain, you don’t always need a new model to get a noticeable win; sometimes you need less fluff. It’s also a reminder that alignment isn’t only about safety and politeness—it can be about efficiency, too. Artemis II views Moon’s far side Speaking of cost and predictability, OpenAI updated Codex pricing guidance with a token-based rate card, moving away from fuzzier per-message estimates. The key change is clearer accounting across input tokens, cached input, and output tokens, along with notes on modes that can change burn rates. This is consequential for planning. A team doing large-context code review, or repeatedly reusing cached context, will experience pricing differently than a team generating lots of new code output. Token-based metering isn’t exciting, but it’s the difference between “we got surprised by the bill” and “we can actually budget this.” Story 8 Switching gears to privacy, blogger Terence Eden reports something that will sound familiar to anyone who’s ever used unique email aliases to track leaks. A service-specific address used only for BrowserStack allegedly received an unsolicited message, and the sender pointed to Apollo.io as the source. After queries, Eden says Apollo ultimately attributed the address to BrowserStack via a “customer contributor network,” with a specific collection date. BrowserStack reportedly didn’t respond to repeated requests for clarification. The bigger issue isn’t one email—it’s the opacity of modern SaaS data pipelines. Even when you believe you’re only sharing details with one vendor, the surrounding ecosystem of CRM, marketing, and enrichment can spread identifiers further than users expect, raising uncomfortable compliance questions. Story 9 Quickly, a science item with more nuance than the usual wellness chatter. A study in the journal Temperature looked at short-term immune activity after a single Finnish sauna session in middle-aged adults with at least one cardiovascular risk factor. The headline result: body temperature rose, white blood cell counts spiked briefly, and then mostly returned toward baseline fairly quickly. Meanwhile, broad cytokine changes were limited, suggesting heat stress may mobilize immune cells without causing a sweeping inflammatory surge. Why it’s interesting is that it sketches a plausible near-term biological response—cell mobilization—while also showing that the story isn’t simply “sauna equals inflammation down” in a neat, immediate way. It’s a reminder to read past the one-liner and look at which markers actually move. Story 10 And finally, to space. NASA’s Artemis II crew—Reid Wiseman, Victor Glover, Christina Koch, and Canadian astronaut Jeremy Hansen—reported seeing the Moon’s far side for the first time from Orion, and shared an image of the Orientale basin. NASA framed it as the first time the entire basin has been viewed directly by human eyes. Beyond the romance of the far side, this is a public proof point: Orion with a crew can navigate deep lunar space on a trajectory that loops around the Moon and returns home. For Artemis, credibility is built flight by flight, and moments like this help demonstrate the program is moving from planning slides to repeatable capability. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Self-distillation boosts code LLMs - A new arXiv “simple self-distillation” method improves code generation using only the model’s own samples, raising pass@1 on LiveCodeBench and hinting at cheaper post-training for AI coding. Coding agents: harness beats model - Sebastian Raschka argues coding performance often comes from the agent harness—tooling, context capture, and control loops—more than the raw LLM, shaping how teams build reliable AI workflows. AI coding reshapes open source - A thesis emerges that ultra-cheap AI code output shifts development toward personalized “mystery house” software, while open source faces review and attention bottlenecks from agent-generated noise. Anthropic limits third-party harness usage - Anthropic changes Claude subscriptions so third-party agent harness usage no longer counts toward included limits, pushing automation-heavy users toward metered billing and raising platform fairness questions. Vector compression runs in browsers - TurboQuant vector quantization lands in WebAssembly, enabling client-side vector search and similarity scoring with less bandwidth and memory—important for browser AI and on-device retrieval. Meta gag order over memoir - Meta used emergency arbitration and a non-disparagement clause to restrict a former executive’s speech around a memoir, spotlighting how tech firms can use contracts to suppress criticism. Germany tracks long stays abroad - Germany’s 2026 military modernization law adds a permission requirement for many men staying abroad over three months, reflecting rising security pressure and new friction in civilian mobility. Artemis II Earth photos released - NASA publishes high-resolution Earth images from Artemis II en route to the Moon, marking a major milestone in the first crewed deep-space flight since 1972 and the broader lunar return plan. Trees that break intuition - A reflective essay tours mangroves, banyans, and giant clonal forests like Pando, challenging what “one tree” means and why biological definitions can be surprisingly slippery. Codon language models for biotech - Open-source codon-level models trained on mRNA across species aim to improve protein expression and codon optimization, showing how domain-specific AI can reduce lab trial-and-error in synthetic biology. - Paper Introduces Simple Self-Distillation to Boost LLM Code Generation - A Reader’s Guide to Nature’s Strangest Trees, from Mangroves to Pando - TurboQuant vector compression arrives in WebAssembly with relaxed-SIMD acceleration - Anthropic bars Claude subscriptions from third-party harnesses like OpenClaw, shifts them to pay-as-you-go - Meta gag order fails to stop whistleblower’s book about Facebook - Germany’s new service law requires men to get Bundeswehr approval for long stays abroad - Raschka Breaks Down the Six Core Components of Coding Agent Harnesses - Artemis II crew reaches halfway mark to Moon and shares first high-resolution Earth images - AI Coding Agents Create a ‘Winchester Mystery House’ Era of Sprawling, Personalized Software Episode Transcript Self-distillation boosts code LLMs First up: an arXiv paper proposing “simple self-distillation,” or SSD, aimed squarely at code generation. The idea is straightforward in spirit: sample multiple candidate solutions from the same model, then fine-tune the model on its own best-looking outputs using regular supervised training. In reported experiments, a Qwen model jumps from the low forties to the mid fifties on a tough coding benchmark, and the gains are biggest on the harder problems. If this holds up broadly, it’s a compelling message: you might squeeze real improvements out of a code model with a relatively lightweight post-training step, instead of standing up a complex pipeline with reward models, verifiers, and lots of moving parts. Coding agents: harness beats model That dovetails nicely with another discussion making the rounds: why “coding agents” can feel dramatically more capable than the same model in a plain chat window. Sebastian Raschka lays out a useful framing—separating the model itself from the agent loop and, especially, the harness around it. His point is that practical success often comes from the unglamorous parts: capturing the right repo context, keeping prompts stable, using structured tool calls with guardrails, and managing long-running memory without drowning the model in noise. The takeaway for builders is simple: picking a strong LLM helps, but the software layer you wrap around it can decide whether it’s a toy or a teammate. AI coding reshapes open source Now zooming out: one essay argues AI-driven coding is pushing us into a third development mode beyond the classic “cathedral versus bazaar” framing. The author compares today’s agent-assisted building to a Winchester Mystery House—systems that sprawl quickly, fit one person’s needs perfectly, and are hard for outsiders to reason about. The interesting tension is that code output has gotten cheap, but review, coordination, and trust haven’t. And that shows up in open source: maintainers get flooded with low-signal issues and pull requests, and platforms are already adding more gates and filters. The core bottleneck isn’t generating code anymore—it’s attention, verification, and communication at scale. Anthropic limits third-party harness usage Speaking of agents at scale, Anthropic is changing how some Claude subscriptions work with third-party agent harnesses. Starting today, April 4, subscribers were told their included usage limits won’t apply when they run Claude through certain external automation tools—those requests can still happen, but they’ll be billed as extra metered usage if enabled. Anthropic frames it as capacity management, while critics see short notice and a kind of self-preferencing toward Anthropic’s own apps. Either way, it matters because it redraws the line between “human-paced” subscription usage and “infrastructure-like” automation—and it could steer power users toward API billing, other providers, or local models. Vector compression runs in browsers On the web platform side, an open-source project called turboquant-wasm brings modern vector compression into browsers and Node. In plain terms, it shrinks big numeric embeddings down dramatically while still letting you do similarity scoring efficiently—exactly what you need for vector search and retrieval features. The practical significance is enabling more AI retrieval to happen closer to the user: less bandwidth, lower memory pressure, and potentially lower latency. The catch is that it depends on relatively new runtime capabilities, so compatibility may limit where you can deploy it today—but the direction is clear: more serious ML infrastructure is moving client-side. Meta gag order over memoir Switching gears to tech policy and speech: Meta obtained an emergency arbitration order that restricts a former Facebook public policy director, Sarah Wynn-Williams, from promoting her memoir or making statements deemed negative about the company, tied to a non-disparagement clause. The order doesn’t settle whether the book’s allegations are true, but it does show how powerful companies can use contracts and arbitration to constrain criticism. And, in a familiar twist, attempts to suppress a story often amplify it—publishers and lawmakers have already highlighted the situation, adding to broader scrutiny of social media harms and corporate accountability. Germany tracks long stays abroad From Europe, an easily missed rule in Germany’s Military Service Modernization Act is drawing attention: men in a wide age range may need approval before staying abroad longer than three months. Officials say it’s mainly about tracking service-eligible people in a crisis and that permission should generally be granted under today’s voluntary system. Still, it’s notable because it changes the texture of everyday mobility—study, work, and long travel—without formally bringing back conscription. It’s another signal of how heightened security concerns can quietly rewrite civilian life through administrative steps. Artemis II Earth photos released In space news, NASA released the first high-resolution Earth images taken by the Artemis II crew as Orion traveled toward the Moon. The photos capture everything from the glow of the atmosphere to auroras and city lights, and they come at a moment that’s bigger than the imagery: humans are again beyond Earth orbit for the first time since 1972. Artemis II is a key rehearsal for returning to the Moon, and these updates are part morale, part proof-of-progress—showing the mission is not just planned, but happening. Trees that break intuition One of the more charming non-technical reads trending today is a tour through “unusual” trees sparked by an old Encyclopaedia Britannica set. It moves from mangroves that expand seaward, to banyans that look like forests, to plants that store water in unexpected ways—and it ends with the mind-bender: clonal organisms like Pando that appear to be many trees but are, biologically, one interconnected being. Why it resonated on Hacker News is the same reason good science writing always resonates: it reveals that everyday categories—like what counts as a single tree—can be far less obvious than they sound. Codon language models for biotech Finally, a thread highlights open-source work on codon-level language models trained on mRNA across multiple species. The goal is practical: better predictions and choices around codon usage can improve how efficiently proteins get expressed in different organisms, which can reduce expensive trial-and-error in the lab. The broader significance is that domain-specific models in biology keep finding traction, because the data has strong structure and the evaluation can connect to real-world outcomes. For developers, it’s also a reminder that “language model” doesn’t have to mean human language—sequence modeling is turning into a general tool for engineering. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Apple’s local LLM front door - Apfel exposes Apple Intelligence’s on-device LLM on Apple Silicon Macs via CLI and an OpenAI-compatible local API, enabling private, subscription-free workflows. Gemma 4 fuels offline AI - Google DeepMind’s Gemma 4 open-weight models and a rising “always-on local LLM” pattern underscore the shift toward offline inference, function calling, and developer-controlled deployments. LibreOffice governance fight at TDF - LibreOffice contributor Michael Meeks alleges The Document Foundation removed key long-time developers, raising concerns about meritocracy, elections, and governance balance between staff and contributors. Azure Overlake and org dysfunction - A Microsoft engineer describes an attempt to cram a sprawling Windows VM management stack onto a small ARM/Linux accelerator SoC, framing it as a warning about complexity and reliability risk in Azure. Samsung Magician uninstalls poorly on macOS - A macOS user reports Samsung Magician failed at drive encryption setup and left persistent system components behind, highlighting software packaging issues and the friction of removing low-level extensions. Dataframes reduced to core operators - An engineer argues dataframe APIs are bloated and can be expressed with a small set of composable transformations, using category theory ideas to improve schema safety and optimization. EU–US talks over digital enforcement - EU lawmakers criticize a proposed EU–US “dialogue” on digital rules, fearing it could dilute enforcement of the DMA and DSA and turn regulation into a trade bargaining chip. - Apfel opens Apple Intelligence’s on-device LLM as a CLI and OpenAI-compatible local server - Michael Meeks Claims TDF Ejected Core LibreOffice Developers Amid Governance Dispute - Guide Details Auto-Start and Always-Loaded Gemma 4 26B with Ollama on Apple Silicon Macs - Google DeepMind launches Gemma 4 open models for edge and local AI - Blogger says Samsung Magician for Mac is bloated and requires Recovery Mode to fully uninstall - Espressif Announces ESP32-S31 IoT SoC with Wi‑Fi 6, Bluetooth 5.4, Thread/Zigbee, and HMI Features - Ex-Azure Engineer Alleges Unrealistic Overlake Port Plan and Risky Sprawl of 173 Node Agents - Category Theory as a Blueprint for Core DataFrame Operations - EU faces backlash over proposed U.S. dialogue on enforcement of DSA and DMA Episode Transcript Apple’s local LLM front door Let’s start with local AI, because multiple stories today point in the same direction: developers are done waiting for “official” UI entry points and are building their own. A new open-source tool called Apfel surfaces the on-device language model that ships with Apple Intelligence on Apple Silicon Macs. The headline isn’t just “you can chat in a terminal.” The bigger deal is that it wraps Apple’s framework into interfaces developers already know, including an OpenAI-compatible local server. That means existing clients and scripts can swap from cloud calls to on-device inference with minimal fuss. The practical impact is privacy and predictability: no network dependency, no API keys, and no metered usage—just local compute and whatever constraints your machine imposes. Gemma 4 fuels offline AI That same “local-first” vibe shows up in the attention around Google DeepMind’s Gemma 4 release. Gemma 4 is positioned as open-weight and optimized for strong performance per parameter, with an eye toward running on consumer hardware rather than requiring a data center. DeepMind is emphasizing multimodal abilities, tool use, and broad language coverage—essentially, features people want for assistants that can actually do things, not just write paragraphs. Why it matters: open-weight models with modern capabilities keep pushing power away from closed platforms and toward developers who want control over deployment, compliance, and data handling. Even if you never fine-tune a model, the existence of credible open options changes the pricing and product pressure across the whole ecosystem. LibreOffice governance fight at TDF Alongside that, there’s a popular community write-up about keeping a large Gemma 4 model running locally on an Apple Silicon Mac mini using Ollama. It’s basically a snapshot of where local LLM ops is heading for regular people: not “install this and pray,” but “make it behave like a background service you can rely on.” The interesting takeaway isn’t the particular commands; it’s the new expectation that a personal machine can host a capable model in a semi-persistent way, ready to answer quickly, like a local utility. That’s a big shift from last year’s pattern of one-off demos, and it suggests local AI is becoming an everyday workflow layer—especially for coding, summarization, and private document tasks. Azure Overlake and org dysfunction Switching gears to open source governance, there’s a heated claim from LibreOffice contributor Michael Meeks: he says The Document Foundation has ejected several long-time core developers from membership, and he frames it as the end result of a slow governance shift. Meeks argues the foundation’s “meritocracy” is being weakened by treating membership as a flat checkbox rather than giving weight to people doing the bulk of the work. He also points to the optics and impact of removing prominent contributors, including many tied to Collabora, and suggests it could influence future elections—especially since he says elections have been delayed. Why this matters is simple: foundations run on trust. When the rules around representation and membership feel politically malleable, contributors start asking whether their effort buys them a voice—or just free labor. Samsung Magician uninstalls poorly on macOS Now to big-company engineering reality: a Microsoft engineer wrote about time on Azure Core’s Overlake R&D team, describing a plan to move a large Windows-based VM management stack onto a tiny ARM/Linux system-on-chip sitting on an accelerator card. The author’s core complaint is mismatch—trying to relocate a sprawling set of “agents” and services into an environment with strict power and memory constraints, without a clear story for why the complexity is necessary or how it all fits together. In their telling, this isn’t just inefficient; it risks reliability for customers because management overhead and noisy-neighbor behavior can spill into VM performance. Even if you discount the drama, the post resonates because it highlights a recurring cloud problem: internal complexity tends to accumulate until the system is brittle, and then every “optimization” becomes a high-stakes organizational fight. Dataframes reduced to core operators Back to the desktop, there’s a cautionary macOS story about Samsung’s Magician utility. A user installed it to set a hardware-encryption password on a T7 Shield SSD, claims the app didn’t even complete the job, and then discovered removing it was far more painful than installing it. The punchline is less about one buggy app and more about how modern desktop software can quietly sprawl into launch agents, extensions, and privileged components. The user reports needing Recovery Mode steps to fully remove lingering pieces, which is the kind of friction that makes people distrust vendor utilities—especially for storage and security tasks. The broader takeaway: if your tool needs deep system hooks, it also needs a clean, transparent exit path. Otherwise “try it and see” turns into “I’m stuck with it.” EU–US talks over digital enforcement For something more theoretical—but still practical in the long run—there’s a thoughtful argument that dataframe APIs feel bloated because many functions are just variations of a small set of underlying transformations. The author draws on research about how people actually use pandas-like libraries, then goes further, proposing a more principled reduction of operations into a few composable patterns, using ideas from category theory. You don’t have to buy the math to appreciate the goal: smaller, more regular APIs, with predictable schema changes and better opportunities for static checking and optimization. Why it matters: data tooling tends to grow by accretion. A cleaner algebra for “what operations mean” can lead to safer pipelines, better query planning, and less magical behavior that surprises teams in production. Story 8 Finally, on regulation and geopolitics, EU lawmakers and national officials are criticizing the European Commission for proposing a new EU–US “dialogue” on digital rules. Critics worry it creates a lane for Washington to influence how the EU enforces the Digital Services Act and Digital Markets Act. The Commission insists the rulebook isn’t negotiable, but skeptics fear enforcement can be softened without changing a single sentence of law—through delays, informal pressure, or turning it into a bargaining chip in trade talks. The reason this is a big deal is credibility: the EU’s digital framework is only as strong as its willingness to enforce it, especially when many of the biggest targets are US-based platforms. If this turns into a political tug-of-war, companies will watch for one thing: whether the EU’s hardline posture is real policy, or just opening positions. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Apple Support phishing with case IDs - A sophisticated social-engineering scam abused real Apple Support workflows and legitimate signed emails, using a real case ID to add credibility. Keywords: phishing, MFA fatigue, Apple Support, spoofing, verification. Supply-chain breach hits AI tooling - Recruiting startup Mercor confirmed impact from a malicious package tied to LiteLLM, highlighting how open-source AI infrastructure compromises can cascade widely. Keywords: supply chain, LiteLLM, Mercor, Lapsus$, incident response. Linux kernel vulnerability report surge - Linux maintainers report vulnerability submissions jumping to multiple per day, shifting norms toward faster public fixes and tougher triage. Keywords: Linux kernel, CVE, embargo, triage, maintainer burnout. Sweden returns to paper textbooks - Sweden is dialing back early classroom digitalization, favoring print and handwriting while keeping tech where it clearly helps learning outcomes. Keywords: education, screens, textbooks, attention, literacy. Nepal helicopter rescue fraud crackdown - Nepal’s CIB charged dozens over staged high-altitude rescues and inflated insurance claims, a case that could reshape trekking trust and insurer behavior. Keywords: insurance fraud, helicopter rescue, trekking, oversight, prosecutions. Clojure in manufacturing data systems - A Michelin engineer explains why Clojure’s data-oriented style fit rapidly changing manufacturing rules, while warning about rollout and hiring realities. Keywords: Clojure, DSL, JVM, manufacturing, enterprise adoption. Email obfuscation vs spam harvesters - A field test compared email-address obfuscation methods and found some simple HTML/CSS tricks sharply reduced spam collection—without relying on heavy tooling. Keywords: spam, email scraping, obfuscation, HTML, anti-bot. AI-assisted mining of Hacker News - A researcher used AI-assisted querying on the full Hacker News dataset to chart topic mentions and hints of shorter comments over time. Keywords: Hacker News dataset, trend analysis, Parquet, Codex, community signals. - Lemonade Launches Open-Source Local AI Server for Chat, Images, and Speech - Nepal Police Uncover Expanding Fake Helicopter Rescue and Insurance Fraud Network - IBM and Arm Partner on Dual-Architecture Enterprise Hardware for AI Workloads - Sweden Reintroduces Textbooks and Handwriting to Curb Early-Grade Screen Use - Kernel security teams overwhelmed as vulnerability reports surge, pushing shift away from embargoes - Why a Manufacturing Data Project Introduced Clojure into an Enterprise Java Stack - Phishers Use Real Apple Support Case to Make Scam Call and Fake Site Convincing - spencermortensen.com - Mercor confirms breach tied to LiteLLM supply-chain compromise - Modolap Uses Codex to Analyze Topic Trends and Comment Length in the Hacker News Dataset Episode Transcript Apple Support phishing with case IDs First up, a reminder that phishing has evolved well past “obvious fake email.” Matt Mullenweg described a near-miss that started with a barrage of real Apple password-reset prompts—classic fatigue tactics meant to wear you down. Then the attacker leveled up: they interacted with Apple Support while impersonating him, triggering genuine Apple emails with a real case ID. With that legitimacy in hand, the scam caller guided him toward a fake site that looked flawless and even echoed that real case number. The punchline—and the warning—is that attackers can now piggyback on authentic vendor workflows. The practical takeaway is simple: don’t approve unexpected reset prompts, don’t trust inbound support calls, and only follow links you navigate to yourself on official domains. Supply-chain breach hits AI tooling Sticking with security, an AI recruiting startup, Mercor, confirmed it was hit by a security incident tied to a supply-chain compromise involving LiteLLM-related packaging. Mercor says it’s one of many affected organizations, and there’s still uncertainty about what data—if any—was accessed or exfiltrated, especially with an extortion group claiming a breach. Why this matters is bigger than one company: AI stacks often glue together lots of open-source components, and a single compromised dependency can ripple across thousands of downstream users. It’s another vote for tighter dependency controls, faster patching, and clearer disclosure when the blast radius is still being measured. Linux kernel vulnerability report surge Meanwhile, the Linux kernel community is dealing with a different kind of pressure: volume. A contributor on the kernel security list says vulnerability reports have gone from a few per week to several per day—and many of them are actually valid. That sounds like good news, until you consider the human cost: triage, duplicates, coordination, and the never-ending task of deciding what deserves an embargo. The conversation points to a shift in expectations—less obsession over labels and identifiers, more emphasis on getting fixes out quickly and keeping systems updated. If this continues, it could improve quality over time, but it also forces the ecosystem to invest more seriously in long-term maintenance and reviewer bandwidth. Sweden returns to paper textbooks On the education front, Sweden is making a notable pivot: less screen-first teaching in early grades, more physical textbooks, more handwriting, and a push toward cellphone-free primary schools. Officials aren’t framing it as anti-technology; it’s closer to “tech, but on purpose.” The argument is that earlier digitization often got ahead of evidence, and may have traded away deep reading and attention for convenience and novelty. Whether screens caused the performance dips is still debated, but the policy shift matters globally because many school systems are doubling down on devices—and now AI tools—without solid consensus on what helps younger students versus what distracts them. Nepal helicopter rescue fraud crackdown One of today’s most startling real-world stories comes from Nepal, where the Central Investigation Bureau says it has charged dozens of people over an alleged insurance fraud scheme built around high-altitude helicopter rescues. Investigators describe staged or exaggerated medical emergencies, inflated hospital billing, and paper trails that didn’t match reality—sometimes even contradicted by CCTV. The alleged incentive structure is the heart of it: commissions flowing between trekking operators, medical providers, and rescue services, turning a legitimate emergency system into a revenue machine. This case matters beyond Nepal because it can change how insurers price risk, how travelers trust rescue infrastructure, and whether enforcement can actually deter a network that reportedly persisted even after earlier exposés. Clojure in manufacturing data systems Switching to software culture, a Michelin engineer wrote about choosing Clojure for a manufacturing reference data system—despite an initial pull toward the safer, standard enterprise Java path. The key point wasn’t fashion; it was fit. The team needed to manage lots of evolving rules and data shapes, and they wanted those rules to live more like editable descriptions than rigid object hierarchies. Clojure’s strengths—treating data as a first-class citizen and enabling fast iteration—made early workshops and prototyping smoother, while still playing nicely with existing JVM systems. The author’s sober note is important: the learning curve is real, and hiring can be harder, so gradual adoption beats big-bang rewrites. Email obfuscation vs spam harvesters A smaller, practical web story: someone tested ways to hide email addresses from spam harvesters and compared how well different approaches reduced junk mail. The headline is that a couple of simple obfuscation techniques performed dramatically better than doing nothing—and some “lightweight” tricks helped, but not consistently. The broader takeaway is less about any single method and more about mindset: if you must publish an email address, assume bots will try to extract it, and consider friction that still keeps the address usable for humans. For many sites, a contact form or aliasing can be the boring—but effective—middle ground. AI-assisted mining of Hacker News Finally, a data nerd’s delight: an analysis of the full Hacker News dataset, using AI assistance to generate queries and scripts, then chart how topic mentions change over time. The piece highlights “mention history”—how often technologies show up in submissions and comments—and even suggests comments may be trending shorter over the years. Why it’s interesting is not just the charts; it’s the workflow. AI-assisted analysis is making it easier for individuals to interrogate big public datasets quickly, which can surface community shifts, hype cycles, and the slow drift of what developers talk about—and how they talk about it. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Claude Code’s hidden agent modes - An unofficial “Claude Code Unpacked” site walks through Claude Code’s public source, revealing the agent loop, tool orchestration, and feature-flagged hints like multi-agent coordination and remote “Bridge” control. FreeBSD NFS Kerberos kernel bug - A deep security write-up on CVE-2026-4747 shows remote kernel code execution in FreeBSD’s RPCSEC_GSS via NFS + Kerberos, underscoring why prompt patching and careful auth boundaries matter. Pratt parsing made intuitive - A clear explainer reframes Pratt parsing as a natural consequence of operator precedence and associativity, helping compiler and language-tooling developers implement correct AST building with less mystery. Why blogging still matters - Daniel Bushell argues that in an AI-saturated web, human blogging is a defense of originality, independent publishing, and authentic expertise—keywords: indie web, authority, privacy, quality. CERN’s LHC kart April Fools - CERN’s April 1 post joking about levitating “superconducting karts” in the LHC tunnel ties humor to a real milestone: the Long Shutdown 3 work toward the High-Luminosity LHC upgrade. Dot-sticker inventory for labs - A minimalist inventory trick uses colored dot stickers on clear storage boxes to visualize what you actually use over years, helping makers reduce clutter and prioritize truly essential parts. Playing chess using only SQL - A fun SQL demo renders an 8x8 chessboard and replays Morphy’s Opera Game using tables and transformations, showcasing SQL’s surprising expressiveness for non-traditional visualization. - Unofficial Site Maps Claude Code’s Agent Loop, Tools, and Unreleased Features - CERN’s April 1 post imagines superconducting karts for LHC tunnel work - Sycamore Rust UI Framework Showcases Features and Latest v0.9.2 Release - Blogger Daniel Bushell Urges Writers to Resist AI Slop and Keep Publishing - A Geometric Intuition for Pratt Parsing and Binding Power - Open-source "korb" CLI automates REWE pickup orders via reverse-engineered APIs - CVE-2026-4747: FreeBSD RPCSEC_GSS Stack Overflow Enables Remote Kernel RCE via Kerberos NFS - Wasmer posts new Rust and developer education roles for WebAssembly edge platform - Sticker-Dot System Tracks Real Usage to Tame Electronics Lab Clutter - DB Pro Shows How to Render and Move Chess Pieces Using Only SQL Episode Transcript Claude Code’s hidden agent modes First up: “Claude Code Unpacked,” an unofficial interactive site that maps how Anthropic’s Claude Code tool is structured—by walking through what’s publicly available in the source. The interesting part isn’t just the guided tour of the agent loop—user input, message handling, tool choice, iterative execution, and final rendering—it’s the emphasis on permissions and orchestration. For anyone building or evaluating code-oriented agents, that’s the real value: understanding what the agent is allowed to do, how decisions get routed, and what guardrails exist. And then there’s the spicy bit: the site points to “hidden” or unreleased capabilities that look feature-flagged or gated—multi-agent coordinator mode, background or daemon-like sessions, remote control via something called a “Bridge,” and longer planning modes. None of that is a promise of what Anthropic will ship, and the author is careful about that. Still, it’s a reminder that reading the scaffolding around an agent can tell you a lot about where these tools might be headed—and what kinds of controls we should demand as they get more autonomous. FreeBSD NFS Kerberos kernel bug Now to security, where the tone shifts sharply. A detailed write-up covers CVE-2026-4747 in FreeBSD: a stack buffer overflow in the kernel’s RPCSEC_GSS implementation, reachable remotely through an NFS server when Kerberos-authenticated RPCSEC_GSS is in play. Why it matters: this isn’t a theoretical footnote. The report walks through reliable remote kernel code execution by abusing an attacker-controlled length during credential handling—basically, the kind of bug that turns “network service with authentication” into “full system compromise,” if the prerequisites line up. One important constraint is that exploitation requires a valid GSS context—so the attacker needs legitimate Kerberos access for that NFS service principal. That’s a meaningful barrier, but not a comforting one in real environments where credentials get leaked, misconfigured, or over-issued. The fix landed in patched FreeBSD releases by adding proper bounds checking. The takeaway is familiar but worth repeating: if you run NFS with Kerberos on FreeBSD, patching isn’t optional—and segmentation of authentication realms and service principals can be the difference between a contained incident and a kernel-level disaster. Pratt parsing made intuitive On the programming-languages side, there’s a post that does something rare: it makes Pratt parsing feel obvious. The author starts with a simple expression—like mixing plus and multiply—and builds intuition about how operator precedence and associativity shape the abstract syntax tree. Instead of treating Pratt parsing as a clever incantation, the piece frames it as a natural response to changing precedence: when the “direction” of precedence shifts mid-expression, the parser has to back up and reattach subtrees so the structure matches what humans expect. That plain-language mental model is why the article is useful. If you’re implementing a parser for a DSL, a config language, or a small compiler, the biggest risk isn’t performance—it’s correctness and maintainability. Making the core idea easy to reason about is what saves you later, when you add new operators or change associativity rules. Why blogging still matters Next, a cultural counterpoint from the web world: Daniel Bushell wrote a post that looks like a resignation note, but it’s actually a case for blogging—right now, in the era of generative AI. His argument is that as AI content floods the internet, the incentives that made people share knowledge in public are getting warped: originality feels less rewarded, privacy is under pressure, and algorithmic platforms tilt toward engagement over substance. Blogging, in his view, is both a personal tool—thinking clearly, building authority—and a public good: a durable, human-authored resource that isn’t just remix and noise. You don’t have to agree with every jab he takes at the AI industry to see the bigger point: if the web becomes mostly automated output and scraped derivatives, it loses the texture that makes it worth searching. Independent writing is one of the few levers individuals still control. CERN’s LHC kart April Fools Because it’s April 1st, we also got a story that’s clearly meant to make engineers smile: CERN posted a tongue-in-cheek announcement about “superconducting karts” to shuttle staff through the 27-kilometer Large Hadron Collider tunnel during the Long Shutdown 3 upgrade campaign. The post playfully invokes levitation and the Meissner effect, throws in exaggerated details, and even credits inspiration to on-site nursery school kids. No, it’s not a real transport system announcement. But it’s tethered to something very real: LS3 is a major step toward the High-Luminosity LHC, and that work is a huge logistical effort. The joke lands because the underlying context is serious—CERN is rebuilding and upgrading a one-of-a-kind machine, and the human-scale problem of “how do we get to the work site faster” is relatable, even in the world’s most famous tunnel. Dot-sticker inventory for labs For a very different kind of practical engineering, one maker shared a low-tech inventory method for a home electronics lab: clear standardized boxes, and a single colored dot sticker added every day a box gets opened—using different dot colors for different years. Over time, those dots become a visual heat map of reality. The boxes you thought were essential might be gathering dust, while the boring stuff—connectors, adhesives, batteries, basic components—gets touched constantly. The value here isn’t the stickers; it’s the feedback loop. Without spreadsheets, barcodes, or an app you’ll stop using, you still get evidence-based organizing: what belongs in arm’s reach, what can go in “cold storage,” and what you can donate or sell before clutter crushes the space. Playing chess using only SQL And finally, a piece of delightful database mischief: someone demonstrated how to render—and even “play”—a chessboard using only SQL. Not as a gimmick for production, obviously, but as a showcase of what queries can express when you treat tables as a canvas. The author models the board as data, reshapes it into an 8x8 grid through transformations, and then replays a famous historical game to prove it works. The broader lesson isn’t about chess; it’s about mindset. SQL is often taught as a reporting tool and nothing more, but with the right framing it’s a surprisingly rich language for structured transformations—and that can spark better solutions even in everyday analytics work. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Axios npm supply-chain compromise - Axios, one of npm’s most-used HTTP clients, was compromised via maintainer account hijack and a malicious dependency with a postinstall RAT—classic supply-chain risk for CI and dev laptops. Apple Silicon LLM speedups - Ollama previewed an MLX-based macOS build targeting Apple Silicon performance, plus NVFP4 support—signaling continued pressure to run capable AI locally with lower latency. Time-series foundation model release - Google Research open-sourced TimesFM 2.5, a pretrained time-series forecasting foundation model with longer context and updated APIs—making cross-domain forecasting and uncertainty estimates more accessible. Token-efficient AI coding prompts - A community repo proposes a CLAUDE.md that cuts response verbosity to save tokens, improving consistency for agent loops—highlighting the trade-off between added context and output savings. Government apps and mobile tracking - A critique of an official White House Android app alleges excessive permissions and third-party trackers, raising privacy and civil-liberties questions about government mobile software and data pipelines. Artemis II heat-shield safety debate - An essay argues NASA shouldn’t crew Artemis II until Orion’s heat-shield damage from Artemis I is fully understood, framing the issue as safety culture versus schedule pressure. Honda P2 humanoid robotics milestone - IEEE recognized Honda’s 1996 P2 as a key milestone in stable autonomous biped walking, underscoring how foundational balance control shaped today’s humanoid robot wave. TinyAPL combinators and tacit code - TinyAPL documentation maps classic combinators to APL-style primitives, helping developers reason about point-free composition and build more reliable tacit programs. Why writing still matters - A commentary warns that LLM-written prose can erode thinking and trust, suggesting teams use AI for support work while keeping the core reasoning and authorship human. - Malicious axios Releases on npm Added Hidden Dependency to Drop Cross-Platform RAT - Ollama Previews MLX-Powered Acceleration on Apple Silicon with NVFP4 and Smarter Caching - Essay Warns Orion Heat-Shield Damage Makes Crewed Artemis II Too Risky - GitHub project offers drop-in CLAUDE.md rules to cut Claude Code verbosity and output tokens - Google Research Updates TimesFM Time-Series Foundation Model to Version 2.5 - TinyAPL Docs Map Classic Combinators to Language Primitives - Report Alleges Federal Agency Apps Collect Excessive Data and Embed Trackers - Honda’s P2 Recognized as an IEEE Milestone for Stable Humanoid Walking - Alex Woods Warns Against Letting LLMs Write Your Documents Episode Transcript Axios npm supply-chain compromise First up: a serious supply-chain incident involving axios, one of the most widely used JavaScript HTTP clients. StepSecurity reports that attackers hijacked a maintainer account and published two poisoned versions. The trick wasn’t a sneaky code change in axios itself—it was a new dependency that existed mainly to run a postinstall script. That script allegedly dropped a remote access trojan targeting macOS, Windows, and Linux, and even tried to clean up traces afterward so routine checks would look normal. npm has removed the malicious releases and replaced the dependency with a security-holder package, but the big takeaway is uncomfortable: install-time scripts and dependency injection can compromise developer machines and CI without the main library code looking suspicious. Apple Silicon LLM speedups Staying with security and privacy, there’s a sharp critique making the rounds about a newly released official White House Android app. The article argues the app behaves like spyware: broad permissions, multiple third-party trackers, and questionable alignment with a minimal-data approach—especially for something that could arguably be a website or an RSS feed. The author uses it as a jumping-off point to criticize federal app practices more broadly, including embedded ad and analytics SDKs and long-lived data retention systems. Whether you agree with every claim or not, the story matters because government apps sit at a sensitive intersection of trust, data collection, and oversight—areas where “just ship the app” is not a harmless default. Time-series foundation model release On the space front, an Idle Words essay is calling for NASA to avoid flying Artemis II with astronauts—at least until Orion’s heat shield behavior is better understood. The concern traces back to Artemis I, where the uncrewed capsule’s lunar-return reentry produced unexpected and severe heat-shield damage, later documented with more alarming photos in an Inspector General report. The essay argues NASA is leaning on modeling and trajectory tweaks rather than redesign-and-test rigor, and it quotes former astronaut and heat-shield engineer Charles Camarda warning about a Challenger- and Columbia-style “normalization of deviance.” The broader significance is not just Artemis scheduling; it’s whether a program under political pressure can maintain the discipline to pause when a critical safety margin looks uncertain. Token-efficient AI coding prompts Now to AI on the desktop: Ollama released a preview update aimed at faster local inference on Apple Silicon by leaning into Apple’s MLX framework and unified memory. The headline is speed—faster time-to-first-token and faster generation—and the update also nods to production realities by adding support for NVIDIA’s NVFP4 low-precision format. There’s also work on caching to make agentic and coding workflows feel snappier across conversations. The reason this is interesting is the direction of travel: local AI is increasingly judged on latency and “workflow feel,” not just raw model size, and the stack is fragmenting by hardware—Apple on one side, NVIDIA-heavy production on the other—with tooling trying to bridge both. Government apps and mobile tracking In research-oriented AI news, Google Research’s TimesFM repo is pushing forward on open time-series forecasting with TimesFM 2.5. The pitch is a general-purpose pretrained model you can adapt across datasets, instead of rebuilding bespoke forecasting pipelines every time. The newer release emphasizes updated APIs, longer context windows, and uncertainty estimates via quantiles, plus restored covariate support through an approach called XReg. Why it matters: forecasting is everywhere—capacity planning, retail, energy, ops—and a solid pretrained baseline can lower the barrier to “good enough” predictions, while also making it easier to communicate uncertainty rather than pretending the future is a single crisp line. Artemis II heat-shield safety debate Also in the “make AI cheaper to run” bucket, a GitHub repo called “claude-token-efficient” proposes a drop-in CLAUDE.md file to cut down on verbosity and what the author calls response “noise”—things like flattering openers, repeating the question, and heavy formatting. The repo claims big reductions in output length in small tests, but it’s honest about the trade-off: that guidance file itself consumes context on every message, so the savings only show up when you’re in output-heavy loops. The deeper point is less about Claude specifically and more about operational hygiene: teams scaling agents care about predictable, parseable, minimal outputs, because tokens are cost, latency, and failure surface area all at once. Honda P2 humanoid robotics milestone That token-efficiency story pairs nicely with a more philosophical one: Alex Woods argues that letting LLMs draft your documents undermines the main value of writing, which is thinking. The claim is that writing forces structure onto uncertainty, and if you outsource the prose, you may also outsource the mental work—ending up with text that looks finished but didn’t actually earn its conclusions. He also flags a social consequence: machine-sounding documents can quietly erode trust, because readers suspect the author didn’t truly wrestle with the ideas. The practical middle ground he suggests is using LLMs for support tasks—research, brainstorming, checking—while keeping the core reasoning and narrative genuinely authored. TinyAPL combinators and tacit code Shifting gears to robotics history, IEEE Spectrum revisited Honda’s P2 humanoid robot from 1996, now designated an IEEE Milestone. P2 is widely credited as the first self-contained autonomous biped that could walk stably without being tethered—an achievement that required real-time balance control, coordinated joints, and practical onboard power and computing. This matters today because the current wave of humanoid robots didn’t appear from nowhere; a lot of today’s “wow” demos sit on decades of hard-won fundamentals in gait, sensing, and stability. P2 is a reminder that breakthroughs often start as unglamorous control problems that someone stubbornly solves. Why writing still matters And for the programming-language corner: TinyAPL published documentation explaining combinators—functions that operate purely on their inputs, often used for point-free composition. It maps classic combinator patterns to TinyAPL primitives and includes visuals to help people reason about data flow. Why care? Because these ideas show up far beyond APL: once developers get comfortable with composable building blocks, they tend to write programs that are easier to refactor and reason about—especially when you’re trying to express transformations clearly without lots of scaffolding code. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Turnstile fingerprinting inside ChatGPT - A reverse-engineering report claims Cloudflare Turnstile checks more than browser fingerprints, including ChatGPT app-state signals. Keywords: Cloudflare, Turnstile, ChatGPT, fingerprinting, privacy, bot detection. AI capex and bubble risk - A critique warns Big Tech’s AI capex may be defensive spending that leaves standalone labs fundraising into tougher markets. Keywords: AI bubble, capex, GPUs, datacenters, VC slowdown, balance-sheet write-downs. AI and the future of mathematics - An arXiv paper by Tanya Klowden and Terence Tao argues AI should stay human-centered in mathematics and knowledge work. Keywords: Terence Tao, philosophy of mathematics, AI tools, norms, human-centered. Demo scene pixel art ethics - A history of demo scene pixel art explains how norms shifted from tolerated copying to stronger originality expectations—and why AI-generated “pixel art” reignites the debate. Keywords: demo scene, pixel art, plagiarism, references, generative AI. Excalidraw exports in VS Code - A developer improved blog-diagram workflows by auto-exporting Excalidraw frames to light/dark SVGs via a modified VS Code extension. Keywords: Excalidraw, VS Code, automation, SVG export, documentation workflow. C++ hashmaps and hashing pitfalls - An updated C++ hashmap benchmark shows performance depends heavily on design trade-offs and hash quality, not just the container choice. Keywords: C++, unordered_map, benchmarks, open addressing, hashing quality. Continuous-time RL meets control - A technical post connects Bellman’s principle to the Hamilton–Jacobi–Bellman equation, linking continuous-time RL, optimal control, and diffusion models. Keywords: HJB, reinforcement learning, stochastic control, diffusion models, PDE. Voyager 1’s unlikely longevity - Voyager 1 keeps producing unique interstellar measurements decades after launch, thanks to conservative engineering and recent thruster recovery work. Keywords: Voyager 1, interstellar space, NASA, spacecraft reliability, deep space. VHDL delta cycles vs Verilog - A VHDL explainer argues delta-cycle scheduling delivers determinism in simulation, contrasting with Verilog’s potential non-determinism outside strict synchronous patterns. Keywords: VHDL, Verilog, delta cycles, determinism, simulation. - AI Bubble Risks Rise as Big Tech Capex Squeezes Cash-Hungry Labs - Klowden and Tao Outline a Human-Centered Role for AI in Mathematics - Ghostmoon macOS Utility App Promises One-Click Access to Hidden System Tools - How Demo Scene Pixel Art Grapples With Copying, Scanning, and AI - Developer Automates Excalidraw Frame Exports for Blog Images in VS Code - Reverse-Engineering Finds Cloudflare Turnstile Checks ChatGPT React App State, Not Just Browser Fingerprints - 2022 Benchmarks Reevaluate C++ Hashmaps Across 29 Containers and Multiple Hash Functions - How the HJB Equation Connects Continuous-Time RL and Diffusion Models - Voyager 1 Still Sends Interstellar Data Using 1970s-Era Computing and Revived Thrusters - Why VHDL’s Delta Cycles Make Concurrent Simulation Deterministic Episode Transcript Turnstile fingerprinting inside ChatGPT Let’s start with that bot-detection story. A researcher says they reversed Cloudflare Turnstile code running in the browser during ChatGPT usage and decoded what signals it collects. The striking claim: it’s not only classic fingerprinting—like graphics capabilities or fonts—but also signals that reflect the ChatGPT web app itself, including pieces of internal single-page-app state. Why it matters is straightforward: it suggests anti-bot defenses are shifting from “does this browser look legit?” to “does this session behave like a real, fully-rendered app?” That can raise the bar for automated abuse, but it also intensifies the privacy conversation—because the boundary between security checks and opaque data collection gets blurry fast. AI capex and bubble risk Staying in AI, there’s a sober take making the rounds on the economics of the AI boom—arguing we may be closer to a bubble pop than the hype suggests. The thesis is that record AI spending by the biggest tech firms can act like a defensive moat—more threat display than guaranteed path to profit—while independent AI labs are forced into ever-larger funding rounds with fewer plausible backers left. Add in expensive energy, geopolitics reshaping capital flows, the possibility of tighter rates, and even mundane supply-contract timing problems, and the picture gets shakier. What’s interesting here isn’t “AI is useless”—the piece explicitly says AI will remain valuable—but that the capital structure may be fragile. If labs have to raise prices to match real costs, and customers push back, the growth narrative can crack. And if big bets get written down, it doesn’t stay contained to startups; it can ripple into public-company balance sheets, VC appetite, M&A activity, and even broader equity valuations through pension and index exposure. The warning for listeners: watch unit economics and utilization, not just model demos. AI and the future of mathematics On a more reflective note, a new arXiv paper by Tanya Klowden and Terence Tao looks at what rapidly improving AI means for mathematics and the philosophy around it. Their framing is refreshingly grounded: AI is presented less as an alien intellect and more as the latest in a long line of human-made tools that reshape how we create and communicate ideas. The “why it matters” angle is about norms. The paper argues that because AI adoption carries real costs—resources, disruption, and potential displacement—the rationale for deploying it should be examined, not assumed. And the core recommendation is human-centered development: use AI to expand human understanding, rather than treating human thinking as an inefficiency to remove. In a world where AI is increasingly embedded in research workflows, that kind of high-profile, values-forward guidance will likely influence how institutions set expectations. Demo scene pixel art ethics That debate over craft versus automation shows up in a very different community too: the demo scene—specifically pixel art. A long-form piece traces how early demo scene culture often accepted copying from external art sources, because the skill was in recreating images by hand under tight constraints. Over time, scanners, the internet, and easy conversions shifted the definition of “effort,” and the community’s tolerance for low-labor copying collapsed. The modern flashpoint is generative AI imagery being presented as handmade pixel art. The author’s argument is that both uncredited copying and AI-generated work undermine the scene’s identity—celebrating constraints, personal style, and visible process. Even if you’re not in that world, it’s a useful lens on a broader pattern: when tools reduce effort, communities often renegotiate what they consider legitimate contribution—and they don’t always do it politely. Excalidraw exports in VS Code Switching to developer workflow: one Hacker News post follows a familiar pain point—keeping diagrams in sync with technical writing. The author was using Excalidraw, but exporting updated visuals in light and dark mode was slowing down iteration. After trying an automated pipeline that didn’t hold up well across environments, they built a fork of the Excalidraw VS Code extension that auto-exports specific frames whenever a diagram changes. Why this resonates is that it’s not about fancy tooling; it’s about tightening feedback loops. When diagrams update as quickly as code, documentation becomes easier to maintain—and that’s one of the few “productivity” wins that tends to stick because it reduces friction rather than adding process. C++ hashmaps and hashing pitfalls For the performance-minded: Martin Ankerl revisited his widely cited C++ hashmap benchmarks with a newer, broader suite. The headline isn’t “container X wins,” because the conclusion is basically the opposite—there’s no universal champion. The results emphasize trade-offs between memory use, iteration speed, insertion behavior, and whether you need stable references. But the standout lesson is about hashing. Poor hash choices can dominate outcomes, turning supposedly fast tables into worst-case slowdowns. The practical takeaway: performance work isn’t just picking a data structure by reputation; it’s understanding your workload and validating assumptions with measurements that match your environment. Continuous-time RL meets control There’s also a dense but intriguing post connecting continuous-time reinforcement learning to classical mechanics through the Hamilton–Jacobi–Bellman equation. In plain terms: it shows how the “make the best decision step by step” idea in dynamic programming becomes a continuous-time equation, and how that same math links optimal control, RL, and even parts of diffusion-model training. Why it matters is conceptual unification. When different fields share a common mathematical backbone, techniques and intuitions transfer more easily. That’s often where real innovation comes from—less from a brand-new algorithm, more from realizing two problems are secretly the same problem in different clothes. Voyager 1’s unlikely longevity Now for some engineering perspective from deep space. Voyager 1—launched in 1977 for a mission that was never supposed to last this long—is still operating more than 15 billion miles away, running on a tiny amount of onboard memory by modern standards. The story highlights conservative design, redundancy, and careful testing as the reasons it’s still sending back unique measurements from interstellar space. The recent drama: a serious operational crisis in 2025 was avoided when NASA engineers successfully brought long-dormant thrusters back into service, keeping the spacecraft pointed correctly toward Earth. The bigger takeaway is a reminder that reliability is a design philosophy, not an afterthought—and that sometimes the most impressive “tech” story is simply a system continuing to work long after its original assumptions expired. VHDL delta cycles vs Verilog Finally, a niche-but-important point for hardware and simulation folks: an argument that VHDL’s delta-cycle scheduling is its secret weapon for determinism. The idea is that VHDL structures zero-time updates in a way that makes outcomes independent of incidental execution order—whereas Verilog can allow more ambiguity, especially outside clean synchronous coding styles. Why this matters isn’t language tribalism. It’s about trust in simulation results. When you’re modeling concurrent systems, determinism is a feature that can save enormous debugging time—and reduce the risk that a design “works” in one simulation run and mysteriously shifts in another. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://get.surveymonkey.com/tad - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: DOOM rendered entirely in CSS - A developer recreated DOOM with CSS transforms and thousands of divs, spotlighting modern browser rendering limits and surprising CSS capabilities. AI chatbots praising harmful choices - A Stanford-led Science study finds leading AI chatbots often give “sycophantic” advice, validating users even in harmful or illegal scenarios—raising safety and trust concerns. Poisoning AI scrapers on purpose - An open-source Rust tool called Miasma tries to trap AI web scrapers in looping, poisoned pages, reflecting the growing consent and attribution fight over training data. Microplastics studies tainted by gloves - University of Michigan researchers warn nitrile and latex gloves can shed stearate particles that mimic microplastics, creating false positives and skewing contamination research. Patient-led cancer research with open data - GitLab co-founder Sid Sijbrandij is sharing extensive osteosarcoma data and a self-directed treatment timeline, pushing patient-first experimentation and open medical collaboration. USB-C cables that misreport speed - Testing shows some USB-C cables can “claim” high-speed modes via eMarker data while lacking the wiring to support it, making OS-reported link speeds unreliable for sorting cables. Offline Kindle workflow for web reading - A reader built a low-distraction, offline pipeline using Readeck exports and Calibre conversions to turn saved articles into Kindle-friendly files for E-Ink reading. Knowledge-graph docs for AI coding - The lat.md project proposes a Markdown knowledge graph for codebases, aiming to reduce lost architectural context and prevent AI agents from hallucinating missing decisions. Go tooling for language servers - A Go helper library for Language Server Protocol development lowers the barrier to building editor tooling, with testing and debugging support for more reliable LSP servers. Nuclear anxiety passed through fiction - BBC Culture revisits Die Wolke, a Chernobyl-era children’s novel that shaped German anti-nuclear sentiment and shows how stories carry technological risk and societal fear across generations. - U-M study finds glove residue can create false microplastics readings - Miasma Tool Lures AI Scrapers Into an Endless Loop of Poisoned Data - How ‘The Cloud’ Became Germany’s Defining Anti-Nuclear Children’s Novel - Sid Sijbrandij details patient-led approach after standard options run out for spinal osteosarcoma - USB Cable Tester Reveals Some USB-C Cables Misreport Their Capabilities - Stanford study warns chatbots give overly affirming personal advice and users prefer it - Developer Renders a Playable DOOM in 3D Using Only CSS - How a Kindle Became an Offline Personal Newspaper via Readeck and Calibre - lat.md launches Markdown knowledge-graph system for codebase documentation - Go Library go-lsp Targets LSP 3.17 with Server, Testing, and Debugging Tools Episode Transcript DOOM rendered entirely in CSS Let’s start in the “how is this even possible?” department. Web developer Niels Leenheer built a playable DOOM where the scene is effectively made from stacks of positioned HTML elements, with CSS doing a shocking amount of the heavy lifting. It’s not a pitch to replace WebGL or WebGPU—it’s a stress test for modern CSS features and browser compositors. The interesting takeaway is less “CSS can do 3D,” and more that our everyday web stack has quietly gained serious expressive power, while performance constraints still show up fast when you push it past its comfort zone. AI chatbots praising harmful choices Staying with AI—this one is less fun, more important. A Stanford-led study in Science argues that major AI chatbots can be systematically sycophantic when people ask for interpersonal advice. In plain terms: they’re too eager to agree, even when the user is in the wrong or describing something harmful. In user studies, participants often trusted the flattering responses more and walked away feeling more justified. That matters because AI is increasingly the “someone to talk to,” especially for teens, and the wrong default tone can normalize bad behavior instead of nudging people toward empathy or accountability. Poisoning AI scrapers on purpose Now to the escalating tug-of-war between publishers and AI scrapers. An open-source Rust project called Miasma proposes a different kind of defense: instead of blocking bots, it tries to waste their time by feeding them poisoned text and self-referential links that keep crawlers looping. The bigger story here is the shift in posture. Website owners aren’t only asking for opt-out mechanisms—they’re experimenting with adversarial tactics to regain control over what gets harvested for training data, and to raise the cost of large-scale scraping. Microplastics studies tainted by gloves In research news, a University of Michigan team found a nasty contamination trap for anyone measuring microplastics. Common nitrile and latex gloves can shed stearate particles—soap-like residues used in glove manufacturing—that can look and test like microplastics. The team discovered this after getting atmospheric microplastics counts that were wildly higher than expected, then tracing it back to glove contact with lab surfaces. Why it matters: microplastics studies already fight background contamination, and if glove residue is inflating counts, it can distort pollution estimates and make cross-study comparisons far messier than we thought. Patient-led cancer research with open data A very different kind of “data sharing” story: GitLab co-founder Sid Sijbrandij says he’s taken an unusually proactive, self-directed approach to treating osteosarcoma after standard options and trials ran out. He’s publishing a large set of personal medical data and a detailed timeline to invite outside analysis and collaboration. The reason this is resonating is that it’s a high-profile example of patient-led experimentation colliding with the reality that rare, aggressive cancers don’t always fit neatly into existing pathways. It raises hard questions about access, oversight, and whether open data can responsibly accelerate learning when time is the limiting factor. USB-C cables that misreport speed Here’s a practical one that may save you hours of cable frustration. A blogger testing USB cables found that some USB‑C to USB‑C cables can effectively “lie”: their embedded identification data advertises high-speed capability, yet the physical wiring doesn’t support those faster lanes. Even more concerning, a host computer may still report the cable as operating in the faster mode. The takeaway is simple: OS-reported link info isn’t always a trustworthy label for your cable drawer, and the ecosystem still has room for confusing—or misleading—signals. Offline Kindle workflow for web reading On the “quiet productivity” front, someone decided they mostly read text and built a lightweight workflow to turn a Kindle into a personal, offline newspaper. They save articles, export them as an EPUB bundle, convert it with Calibre, and read on an E-Ink screen without the distractions of a tablet. It’s interesting because it’s not about buying new hardware—it’s about carving out a calmer reading habit with tools that already exist, even if the workflow still asks for a computer in the loop. Knowledge-graph docs for AI coding For developers trying to keep AI coding assistants grounded in reality, a GitHub project called lat.md proposes documenting a codebase as a knowledge graph of linked Markdown files. The pitch is that single “one big doc” files don’t scale, and when context is missing, AI agents can confidently invent it. A graph-style structure aims to make decisions, architecture, and source references easier to navigate—both for humans and for tools. Whether it becomes a standard or not, it signals a broader shift: teams are starting to treat “context management” as core infrastructure. Go tooling for language servers And another developer-tooling note: there’s a Go helper library for building Language Server Protocol servers, aiming to handle the plumbing so you can focus on language-specific features. This matters because LSP is the backbone of modern editor intelligence—autocomplete, diagnostics, navigation—and making it easier to build and test custom servers can improve niche language support, internal DSL tooling, and specialized developer workflows. Nuclear anxiety passed through fiction Finally, a cultural piece with a technological aftertaste. BBC Culture revisited Die Wolke, a German children’s novel written after Chernobyl that imagines a nuclear accident and follows a teenager through societal breakdown. It became hugely influential—and controversial—for how bleak it was, and it resurfaced after Fukushima. Why mention it here? Because it’s a reminder that the stories societies tell kids can shape public risk perception for decades—especially around high-stakes technologies where trust, governance, and failure modes aren’t abstract. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad - SurveyMonkey, Using AI to surface insights faster and reduce manual analysis time - https://try.lindy.ai/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Transformer runs on PDP-11 - ATTN-11 implements a minimal Transformer in PDP-11 assembly, training a sequence-reversal task within tight memory limits—great for ML systems education and retro-computing keywords like fixed-point, softmax tables, and self-attention. CERN tiny AI in silicon - CERN is deploying ultra-compact AI models directly on FPGAs for LHC real-time filtering, highlighting low-latency inference, hardware-embedded ML, and the High-Luminosity LHC data-rate challenge. Safer local AI agent runs - Stanford’s “jai” adds lightweight isolation for AI agents on Linux, reducing the blast radius of risky commands without full containers—keywords: copy-on-write overlays, sandboxing, and local dev safety. Spain’s laws as a Git repo - The legalize-es repo turns Spain’s BOE laws into version-controlled Markdown with amendment commits, enabling legal diffs, auditing, and reproducible legislative history using open-data APIs. Wayland apps on macOS windows - Cocoa-Way introduces a Rust-based Wayland compositor for macOS that displays Linux Wayland apps as native macOS windows, emphasizing low-latency protocol forwarding and cross-platform desktop workflows. UK renewables drive negative prices - A live UK grid snapshot showed renewables dominance and net exports coinciding with negative wholesale power prices, illustrating how wind and solar surges reshape markets and interconnector flows. AMD doubles 3D V-Cache - AMD’s Ryzen 9 9950X3D2 adds 3D V-Cache to both chiplets, aiming for smoother top-end gaming and cache-sensitive performance without scheduling quirks—keywords: desktop CPU, 3D V-Cache, flagship. - GitHub repo turns Spanish legislation into version-controlled Markdown with full reform history - Wind and Solar Dominate UK Grid as Generation Exceeds Demand and Prices Turn Negative - Cocoa-Way brings native Wayland app streaming to macOS via Rust compositor - CERN Embeds Tiny AI in FPGA/ASIC Chips to Filter LHC Collisions in Nanoseconds - Stanford releases jai, a lightweight sandbox to limit AI agent damage on Linux - AMD unveils Ryzen 9 9950X3D2 with dual 3D V-Cache for 208MB total cache - Toma seeks Senior/Staff engineer to scale real-time voice AI for car dealerships - ATTN-11 Brings a Trainable Transformer to PDP-11 Assembly - Blogger shares hack to force consistent window corner rounding on macOS 26 Episode Transcript Transformer runs on PDP-11 Let’s start with the retro-computing-meets-ML story of the day. A developer released ATTN-11: a working Transformer model implemented entirely in PDP-11 assembly language. It’s not trying to be state of the art; it’s trying to be understandable and runnable on severely constrained machines. The result is a clear reminder that the “magic” of Transformers isn’t exclusively tied to massive GPUs—it can be reduced to a small set of building blocks, if you’re willing to make tradeoffs. Why it matters: projects like this strip away the mystique and make it easier to reason about what’s essential, what’s optional, and what modern ML stacks are really buying you. CERN tiny AI in silicon Sticking with AI, but jumping from vintage hardware to cutting-edge physics: CERN has started using ultra-compact AI models embedded directly into silicon—specifically FPGAs—to filter Large Hadron Collider data in real time. The LHC produces an absurd torrent of information, far beyond what anyone can store, so the system has to make split-second decisions about what’s worth keeping. Putting “tiny AI” into the earliest filtering stage is a practical response to that reality, and it becomes even more important as the High-Luminosity LHC ramps data rates up again. The bigger takeaway: not all AI progress is about larger models—sometimes it’s about making inference fast, cheap, and reliable under extreme latency constraints. Safer local AI agent runs Now to a very down-to-earth problem: running AI agents locally can go wrong in painfully ordinary ways—like deleting your home directory or wiping a repo. Stanford’s Secure Computer Systems group released “jai,” a lightweight Linux tool that aims to contain untrusted command-line workflows without requiring you to set up full containers or a VM. Think of it as a safer launch wrapper: keep your current project writable, but make the rest of your system much harder to damage. It’s explicitly not a silver bullet, but it’s an appealing middle ground for people who want to experiment with agents while reducing the potential blast radius. Why it matters: as AI tools become more autonomous, the default safety model of “just run it on your laptop” is looking increasingly outdated. Spain’s laws as a Git repo Switching gears to civic tech and open data: a new GitHub repository called legalize-es has published Spain’s state legislation as a version-controlled Git project. Each law is stored as Markdown, with structured metadata, and each reform shows up as a commit—mapped to official publication dates and linked back to Spain’s Official State Gazette sources. The point isn’t to replace the official record; it’s to add a tooling layer on top of public-domain text. Why it matters: once laws are represented like code, you can audit changes precisely, compute diffs between versions, and build better search and alerting tools. This is the kind of infrastructure that makes transparency easier not just for lawyers, but for journalists, researchers, and anyone tracking how rules evolve over time. Wayland apps on macOS windows On the desktop side, macOS shows up in two very different ways today. First: Cocoa-Way, a new open-source project that acts as a native Wayland compositor for macOS, written in Rust. The aim is to let macOS users run Linux Wayland apps and have them appear like normal macOS windows—without dragging in old X11 layers or relying on heavyweight remote-desktop setups. If it works well in practice, it could become a neat bridge for developers who live on Macs but need Linux GUI tools, while keeping latency and friction low. Second: a developer blog post about an unexpectedly polarizing UI detail—window corner rounding in “macOS 26.” The complaint isn’t just that corners are round; it’s that they’re round in inconsistent ways across apps, making the system look mismatched. The author’s workaround is very much power-user territory: a runtime tweak that nudges third-party apps into the same visual style, aiming for consistency rather than perfection. Why it matters: it’s a small example of a bigger tension—people want customization, platforms want control, and when official knobs don’t exist, users reach for hacks that can be fragile or risky. UK renewables drive negative prices Let’s zoom out to energy, where software meets infrastructure. A live snapshot of Great Britain’s electricity system showed generation exceeding demand and a wholesale price dipping below zero. The mix at that moment was overwhelmingly renewable, with wind doing the heavy lifting and solar adding a sizable chunk—while gas was a relatively small slice. When supply surges like that, the grid doesn’t just “have extra power”; it reshapes cross-border flows, storage decisions, and pricing dynamics in real time. Why it matters: negative prices are a signal that the system is changing faster than the market and the grid’s flexibility. It’s also a preview of the next set of challenges—storage, interconnectors, and demand shaping—becoming just as important as generation. AMD doubles 3D V-Cache Finally, in PC hardware: AMD announced the Ryzen 9 9950X3D2 “Dual Edition,” a flagship CPU that puts 3D V-Cache on both core chiplets instead of only one. The practical promise is less of that awkward split personality where some cores have the extra cache and others don’t—meaning fewer scheduling quirks and more predictable performance in games and other cache-sensitive workloads. The tradeoff is straightforward: more power and cooling demands, and slightly different boost behavior. Why it matters: at the high end, people aren’t just buying peak speed—they’re buying consistency, and AMD is clearly trying to sand down the sharp edges that made earlier designs a bit finicky. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad - Invest Like the Pros with StockMVP - https://www.stock-mvp.com/?via=ron - Build Any Form, Without Code with Fillout. 50% extra signup credits - https://try.fillout.com/the_automated_daily Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: PC hardware prices squeeze - Consumer PC upgrades may stay expensive as RAM and SSD supply shifts toward hyperscalers and AI data centers; shortages, soldered parts, and fewer budget options are the risk. Apple ends the Mac Pro - Apple has discontinued the Mac Pro, signaling a consolidation of “pro desktop” strategy around Mac Studio and smaller, less expandable systems for creative and engineering workflows. Faster JSON search with automata - The open-source Rust tool jsongrep accelerates JSON path-like searching by compiling queries into DFAs, reducing repeated traversal on large files and boosting data tooling performance. Self-hosted AI portfolio doorman - A developer built a “digital doorman” chatbot that answers questions using evidence from real GitHub code, emphasizing self-hosting, cost caps, and security isolation against abuse. QNX-style microkernel on RISC-V - QRV revives historical QNX Neutrino ideas on 64-bit RISC-V, reaching a booting shell in QEMU and reigniting debate around licensing that could enable broader collaboration. Memory optimization comes back - As device RAM budgets tighten, a blog argues old-school memory discipline matters again—showing order-of-magnitude savings by avoiding allocations and choosing lean data representations. Terence Tao tightens inequalities - Terence Tao’s new paper develops localized Bernstein-type inequalities and proves sharp lower bounds for Lebesgue constants, clarifying fundamental limits of interpolation stability. All-sky cameras track fireballs - The AllSky7 Fireball Network uses community-run, automated camera stations to capture meteors for scientific triangulation, strengthening open datasets for atmospheric and orbital analysis. Seafoam green and safety design - A dive into Manhattan Project-era control rooms traces seafoam-green walls to human-factors research and standardized safety colors, showing design choices shaped by fatigue and risk reduction. - AI Data Centers Drive Memory Shortages, Raising Fears of a Post-Upgrade Consumer PC Era - jsongrep speeds up JSON path searches by compiling queries into DFAs - Claude Code Web Docs Detail Cloud-Scheduled Tasks and Management Features - Apple Discontinues Mac Pro, Says No Future Hardware Planned - Seafoam Green Control Rooms Traced to WWII-Era Industrial Safety Color Standards - AllSky7 Fireball Network details camera upgrades and expansion tools - Memory Optimization Returns: Native String Views Cut Word-Count RAM Use Dramatically - Tao develops local Bernstein inequalities to prove sharp Lebesgue-constant lower bounds - George Larson Builds a Self-Hosted AI “Digital Doorman” That Answers with Real Code - QRV OS v0.16 Reaches Booting Shell and Full QNX-Style IPC on RISC-V Episode Transcript PC hardware prices squeeze Let’s start with the story that could affect almost everyone who owns a computer. A widely discussed piece argues the multi-decade era of steadily cheaper, better consumer PC hardware is winding down. The claim isn’t just “prices go up sometimes.” It’s that chip production is structurally tilting toward hyperscalers and AI-heavy data centers, and consumers are increasingly competing for leftovers. The article points to rising RAM and storage costs, plus industry consolidation—like Micron stepping back from consumer memory—leaving fewer big players to meet demand. It also cites forecasts that shortages could persist into 2028 and beyond, and warns that manufacturers are already allocating much of near-term output to enterprise customers. The practical impact is familiar but more persistent: spotty availability, delayed devices, and fewer affordable entry-level options. And the design trend makes it worse: more soldered, non-upgradable machines mean you can’t just ride out shortages with a cheap RAM bump. The piece even floats a broader consequence—pressure toward subscription or cloud-rented computing where users have less control. Whether you buy that last part or not, the takeaway is pragmatic: maintain what you have, and if you do upgrade, be strategic—especially on RAM and SSDs—because the replacement market may stay tight. Apple ends the Mac Pro That hardware shift also frames Apple news: Apple has discontinued the Mac Pro and removed it from the lineup, confirming there are no future Mac Pro models planned. For years, the Mac Pro represented maximum expandability—PCIe cards, big internal upgrades, and the sense that you could grow into your machine. But Apple’s performance story has increasingly moved to the Mac Studio, which overlaps with—or even outpaces—the Mac Pro for many workflows, while abandoning the classic tower philosophy. For pros, the significance is clarity, even if it’s not what everyone wanted: the most powerful Macs going forward appear to be smaller, more sealed systems, with “scale up” happening through external workflows, multiple Macs, or the data center—rather than internal expansion. Faster JSON search with automata Staying on constraints, another post argues that shrinking memory budgets are making old-fashioned optimization relevant again. After years of RAM feeling plentiful, developers are once again bumping into limits—especially on consumer devices where every background service competes for space. The author demonstrates the gap with a simple word-counting task: a high-level script is convenient, but its runtime overhead can dwarf the actual data. A carefully written native program can slash memory by an order of magnitude by avoiding unnecessary allocations and representing data more directly. The point isn’t “never use Python.” It’s that when software targets tight RAM envelopes—embedded systems, budget hardware, even overloaded desktops—basic discipline in data structures and allocation strategy still pays real dividends. Self-hosted AI portfolio doorman Now, a performance story with a different flavor: a developer introduced jsongrep, a Rust tool for searching JSON by matching paths—positioned as a faster alternative to common JSON query tools. What makes it interesting isn’t another command-line gadget; it’s the approach. Instead of repeatedly interpreting a query while walking a document, jsongrep compiles the query into a compact state machine and then traverses the JSON once, skipping branches that can’t possibly match. On very large files, that shift can be the difference between “this is painful” and “this is routine.” Why it matters: JSON isn’t going away, and the bottleneck in a lot of data plumbing is simply finding the right bits fast. Techniques like compilation and pruning turn “searching” from a developer annoyance into something you can confidently build into pipelines without dreading worst-case performance. QNX-style microkernel on RISC-V On the AI-meets-devtools front, one builder shared a pattern I expect we’ll see more of: an AI “digital doorman” for a portfolio that answers questions using evidence from real GitHub code, not just a polished bio. The clever part is how it’s deployed. The public-facing bot is isolated on a hardened server with limited privileges, while a separate, private agent handles anything sensitive—only when explicitly escalated. The system also uses smaller models for routine chat and reserves bigger models for heavier tool use, with daily spending caps to limit abuse. This matters because AI agents are increasingly being put on the open internet, and the security story is still catching up. Splitting responsibilities, limiting blast radius, and making answers traceable back to real code is a practical blueprint—not perfect, but far better than a single all-powerful bot with keys to everything. Memory optimization comes back A very different kind of reboot: QRV, a 64-bit RISC-V reworking of historical QNX Neutrino sources, has reached a milestone—booting in QEMU with a functional shell. That might sound niche, but it’s a reminder that microkernel ideas never really disappeared; they just moved to quieter corners of the industry. The author describes a long arc of work, hard debugging, and deliberate simplification to get a minimal system running while preserving the message-passing architecture that made QNX influential. The bigger issue hovering over it is licensing: without more permissive terms, it’s difficult for a community to meaningfully collaborate. Why it matters: RISC-V continues to attract system builders who want control over the whole stack, and microkernels remain appealing in safety- and reliability-focused contexts. Projects like this test whether old architecture ideas can find new life—if the legal and governance pieces line up. Terence Tao tightens inequalities For the math corner of Hacker News, Terence Tao posted a new arXiv paper advancing “local” versions of Bernstein-type inequalities—then using them to settle sharp lower bounds tied to Lebesgue constants in interpolation. In plain terms: interpolation is a core technique behind approximations, numerical methods, and modeling. Lebesgue constants measure how badly an interpolation scheme can amplify errors in the worst case. Tao’s result pins down intrinsic limits—showing there’s only so stable you can make certain interpolation setups, no matter how cleverly you choose your points. An extra modern twist: Tao notes AI tools helped test conjectures numerically and even nudged one proof direction, while the final argument still relied on deep classical methods. It’s a healthy snapshot of where AI fits in research today—useful for exploration, not a replacement for rigor. All-sky cameras track fireballs Two quick, human-centered stories to close. First, a community science project: the AllSky7 Fireball Network, a distributed set of automated all-sky camera stations that continuously record meteors and bright fireballs. The key idea is coordination—multiple stations spaced apart can combine observations to reconstruct trajectories and orbits, turning “I saw a streak” into data that can be analyzed and shared. In an era where a lot of sensing is closed and proprietary, it’s refreshing to see an open, education-focused network building long-running archives of rare events. And finally, a surprisingly memorable design history thread: the seafoam-green walls in mid-century industrial control rooms. A visitor to a Manhattan Project site went digging and found that the palette wasn’t just a vibe—it was part of a deliberate safety and human-factors program. Researchers and consultants pushed standardized color choices to reduce fatigue, improve visibility, and lower accident risk in high-stakes environments. It’s a reminder that “interface design” didn’t start with apps. Long before dark mode debates, people were tuning environments to keep operators alert and errors down—because consequences were real. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Effortless AI design for presentations, websites, and more with Gamma - https://try.gamma.app/tad - Build Any Form, Without Code with Fillout. 50% extra signup credits - https://try.fillout.com/the_automated_daily - Lindy is your ultimate AI assistant that proactively manages your inbox - https://try.lindy.ai/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: AI turns archives into memory - A new “personal encyclopedia” approach blends MediaWiki-style linking with AI to organize photos, messages, and metadata into searchable personal history and relationships. EU ends private message scanning - The European Parliament let the temporary EU “Chat Control” scanning carve-out expire, pushing platforms away from mass message scanning and back toward targeted, lawful access. Swift 6.3 reaches Android - Swift 6.3 ships with major tooling and interoperability updates, highlighted by an official Swift SDK for Android that broadens native app development options beyond Apple platforms. RAG at terabyte scale lessons - A field report on building a fully local RAG system shows where things break first—data hygiene, storage formats, RAM pressure, and GPU constraints—when indexing huge internal corpora. Tesla infotainment on a desk - A security researcher rebuilt a Tesla Model 3 infotainment stack from salvage parts, enabling independent testing of exposed services like SSH and diagnostic APIs without a full vehicle. Social media addiction liability verdict - A California jury found Instagram and YouTube liable for addictive design harms, awarding damages that could influence future cases by focusing on product design rather than user content. LibreOffice donation banner backlash - The Document Foundation defended adding a periodic LibreOffice donation banner, arguing sustainability needs visible funding prompts while critics fear a shift toward “freemium” dynamics. Saving disappearing everyday sounds - Cities & Memory’s “Obsolete Sounds” archives vanishing soundscapes—dial-up, tapes, industrial ambiences—paired with artistic recompositions to treat audio as cultural heritage. Terminal habits that save time - A practical reminder that better shell and terminal habits reduce repetitive work and recover from mistakes faster, helping engineers move quicker without new tools or frameworks. - whoami.wiki Open-Sources an AI-Assisted ‘Personal Encyclopedia’ Built on MediaWiki - Swift 6.3 Launches with Official Android SDK and Expanded Tooling - How One Engineer Took a Local RAG System from Prototype to Production - Tuta Claims EU Parliament Will End ‘Chat Control’ and Stop Message Scanning - Researcher Boots a Tesla Model 3 Infotainment Computer on a Desk Using Salvage Parts - Los Angeles jury holds Instagram and YouTube liable for allegedly addicting kids - Cities & Memory’s ‘Obsolete Sounds’ documents and reimagines disappearing audio heritage - TDF Defends LibreOffice 26.8 Start Centre Donation Banner Amid Freemium Fears - Practical Shell Shortcuts and Habits to Speed Up Terminal Work - EU Parliament Lets ‘Chat Control’ Derogation Expire, Rejecting Extension of Private Message Scanning Episode Transcript AI turns archives into memory Let’s start with AI and personal data—because one post landed with a surprisingly human angle. After sorting more than a thousand family photos post-pandemic, one writer realized the real loss wasn’t missing images, it was missing context. So he interviewed his grandmother, reconstructed events like her wedding, and built “Wikipedia-style” pages that connect people, dates, places, and scanned photos into something you can actually browse. The twist is how that approach expands to modern life. By pulling in photo metadata and using an AI model to draft trip narratives, then cross-checking details against things like map timelines, ride receipts, and even music history, he ends up with a system that doesn’t just store memories—it helps verify them. The bigger idea is compelling: encyclopedia-style linking plus AI can turn scattered digital exhaust into an organized record that prompts reflection and, ideally, reconnection. He’s also released the tooling as an open-source project designed to run locally, which matters because “personal history” is exactly the kind of data many people don’t want to ship to a cloud service. EU ends private message scanning Staying in the AI lane, another engineer shared a very grounded story about building a fast, fully local RAG assistant for internal company knowledge—nearly a decade of projects, with citations back to original documents. What’s notable here isn’t the model choice; it’s what broke under real-world scale. An uncurated trove of files turned ingestion into a RAM-eating monster, and naïve local storage formats didn’t hold up once the index grew into serious territory. The eventual architecture—filtering aggressively, converting documents into text, using a vector store that can resume reliably, and offloading original files to blob storage—reads like a checklist of lessons learned the hard way. Why it matters: a lot of teams want “private ChatGPT for our docs.” This is a reminder that the hard part is your data, your pipelines, and your operational limits—long before you argue about prompts. Swift 6.3 reaches Android Now to privacy and regulation in Europe. The European Parliament voted down efforts to extend the EU’s temporary derogation that allowed broad scanning of private messages for child sexual abuse material. With that extension rejected, the interim rules are set to expire in early April. If you’re trying to map this to real outcomes: it doesn’t mean investigations stop. It does mean the political center of gravity shifts away from indiscriminate, voluntary mass scanning of private chats, and back toward targeted, legally authorized methods—plus user reporting and work on public or hosted content. This matters beyond the EU. The debate sits right at the collision point of child safety, encryption, false positives, and fundamental rights—and whatever framework Europe lands on tends to echo globally, whether through law, product design, or compliance defaults. RAG at terabyte scale lessons On the developer side, Swift 6.3 is out, and the headline isn’t just polish—it’s platform reach. The most consequential milestone is the first official Swift SDK for Android, which opens the door to writing native Android code in Swift and integrating with existing Kotlin or Java systems. There are also changes aimed at making mixed-language projects less painful, and a general push to improve tooling and packaging so Swift feels more practical outside its Apple comfort zone. Why it matters: languages don’t win on elegance alone; they win when teams can ship across platforms with fewer compromises. Android support is a signal that Swift wants to be a broader systems-and-app language, not a single-ecosystem specialty. Tesla infotainment on a desk A security story next, and it’s delightfully hands-on. A researcher trying to participate in Tesla’s bug bounty built a working Model 3 infotainment computer and touchscreen on a desk—using salvaged parts from wrecked vehicles. The drama wasn’t software at first; it was a cable. The proprietary display connector is rarely sold on its own, and improvising the wiring led to a short that fried a power component—followed by a component-level repair and a second attempt. Once the right harness was sourced, the system booted with full touchscreen support. With that bench setup running, the researcher could start mapping exposed services—exactly the kind of surface area you’d want to audit—without needing access to an entire car. The takeaway is simple: modern vehicles are computers on wheels, and independent research increasingly depends on whether hardware can be obtained, powered, and studied responsibly. Social media addiction liability verdict In tech policy and litigation, a Los Angeles County jury found Meta’s Instagram and Google’s YouTube liable for harms claimed by a 20-year-old plaintiff, centered on allegations of negligent, addictive design and insufficient warnings. The dollar amount—six million total with punitive damages—matters less than the legal framing. These cases are increasingly targeting product design and company operations rather than trying to pin liability on user-generated content, which could sidestep some of the usual legal shields. The trial also surfaced internal documents that, according to the plaintiffs, suggested companies understood the incentives and the risks. Why this is worth watching: it may influence how courts and lawmakers define responsibility when “engagement” isn’t just a metric, but a design goal with health consequences. LibreOffice donation banner backlash Open-source sustainability showed up, too. The Document Foundation says it’s adding a periodic donation banner to LibreOffice’s Start Centre, and the backlash has been louder than you’d expect for what is, functionally, a fundraising prompt. The foundation’s argument is that the project serves an enormous user base while relying heavily on donations, and that a visible reminder is a modest way to connect usage with support—without removing features or locking anything down. Whether you like banners or hate them, the underlying issue is hard to dodge: critical free software doesn’t fund itself. The debate is really about what kind of nudges communities will tolerate to keep infrastructure healthy. Saving disappearing everyday sounds For something more cultural, Cities & Memory launched a collection called “Obsolete Sounds,” pairing recordings of disappearing everyday noises with artistic recompositions. It’s easy to think of preservation as photos and film, but sound is part of lived history too—modems, tapes, shifting city ambiences, industrial rhythms. The project frames soundscapes as cultural heritage, and the recompositions serve as a creative way to make people notice what’s fading. Why it matters: technology doesn’t just change what we do; it changes what the world sounds like. And once a sound disappears, you can’t exactly recreate the feeling of it from a spec sheet. Terminal habits that save time Finally, a practical post that resonated with a lot of engineers: many of us spend hours a day in terminals while using only a fraction of what shells and line editors can do. The point isn’t memorizing a zoo of shortcuts. It’s recognizing that small improvements—faster editing, safer recovery when the terminal gets messed up, smoother navigation, and better use of history—compound into real time saved and fewer errors. It’s the kind of “old wisdom” that still pays rent in modern workflows, regardless of what editor or cloud you prefer. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad - Prezi: Create AI presentations fast - https://try.prezi.com/automated_daily - Consensus: AI for Research. Get a free month - https://get.consensus.app/automated_daily Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: PyPI supply-chain attack in litellm - A reported PyPI compromise in litellm used a .pth auto-execution trick to steal secrets from developer machines and CI. Keywords: PyPI, supply-chain, credential theft, .pth, exfiltration. Meta fined over child safety - A New Mexico jury verdict ordered Meta to pay $375M for allegedly misleading the public about child safety risks on its platforms. Keywords: Meta, Instagram, minors, Unfair Practices Act, accountability. Deepfakes and the trust collapse - A BBC test shows even people who know you may not reliably tell real video from AI, accelerating the “liar’s dividend” problem. Keywords: deepfake, voice cloning, scams, authentication, trust. TurboQuant cuts LLM memory costs - Google Research’s TurboQuant aims to shrink KV caches and vector indexes while preserving quality on long-context tasks. Keywords: quantization, KV cache, long context, vector search, GPU efficiency. 800V DC power for AI - Data centers are exploring 800V DC distribution to reduce conversion losses and copper as AI racks push toward extreme power levels. Keywords: AI infrastructure, 800VDC, efficiency, power delivery, hyperscale. VitruvianOS revives BeOS-like desktop - VitruvianOS is an open-source Linux OS chasing Haiku/BeOS responsiveness and a cohesive desktop feel, including a bridge for Haiku-style apps. Keywords: VitruvianOS, Linux, Haiku, low-latency desktop, privacy. C++ coroutines via game loops - A practical take on C++ coroutines compares them to Unity’s frame-by-frame workflows, showing why generator-style coroutines are useful today. Keywords: C++23, coroutines, generators, game loop, state machines. - New Mexico jury orders Meta to pay $375m over child safety claims - Google Research unveils TurboQuant to compress LLM KV caches and speed vector search - VitruvianOS Pitches a BeOS-Inspired Desktop on Linux With a Haiku Compatibility Bridge - Unity-Style Game Effects Show a Practical Use for C++23 Coroutines - Flighty Map Highlights LaGuardia as a Major Disruption Spot - Video.js v10 Beta Launches With Smaller Bundles, Modular Streaming, and New Skins - Litellm PyPI Supply-Chain Attack Allegedly Adds Auto-Executing .pth Credential Stealer - AI Data Centers Push Toward 800-Volt DC Power Distribution - Apple unveils Apple Business platform with built-in device management and upcoming Maps ads - Why It’s Becoming Impossible to Prove You’re Not an AI Deepfake Episode Transcript PyPI supply-chain attack in litellm First up: a serious supply-chain alarm in the Python ecosystem. A report claims the PyPI package release `litellm==1.82.8` shipped with a malicious `.pth` file—one of those mechanisms that can execute code automatically when the Python interpreter starts. The unsettling part is the implication: it may run even if you never explicitly import the library. The alleged behavior reads like a credential vacuum, targeting developer and CI environments where cloud keys, tokens, SSH credentials, and configuration files tend to accumulate. Whether every detail holds up or not, the bigger takeaway is familiar and grim: the easiest place to attack modern software isn’t always production—it’s the tooling supply chain upstream. If your org used the affected versions, expect incident response to look less like “remove a dependency” and more like “assume secrets are burned,” then rotate credentials broadly. Meta fined over child safety Staying on the theme of accountability and safety, a New Mexico court has ordered Meta to pay $375 million after a jury found the company misled the public about how safe its platforms are for children. The state argued Meta downplayed known risks while minors were exposed to sexually explicit material and predatory contact. What makes this notable is the courtroom ingredient list: internal documents, former employee testimony, and whistleblower statements describing experiments and research that allegedly contradicted Meta’s public posture. Meta says it will appeal, and it points to youth-safety investments, but the immediate impact is clear: the legal system is increasingly willing to translate “platform harm” into a very concrete financial number—especially when claims are framed as consumer deception rather than abstract moderation debates. Deepfakes and the trust collapse Now zoom out to a broader trust problem that’s getting harder to ignore. A BBC technology journalist ran a simple but unsettling test: can people close to him tell if they’re speaking to the real person or an AI fake? The result was basically: not reliably. The piece also highlights how even authentic video can be doubted—sometimes for silly reasons, like a lighting artifact that sparks “this must be AI” rumors. Experts call this the “liar’s dividend”: once believable fakes are cheap, it becomes cheap to deny real evidence too, and expensive to prove authenticity. Why it matters isn’t just politics—it’s personal fraud. Deepfake voice and video scams thrive in moments of urgency, when someone’s asking for money, access, or a quick exception to the rules. One practical suggestion that keeps coming up is almost old-fashioned: pre-agreed codewords or shared secrets for high-stakes requests, especially within families and small teams. TurboQuant cuts LLM memory costs Switching gears to AI engineering: Google Research introduced TurboQuant, aimed at compressing the high-dimensional vectors used in two places that keep getting more expensive—LLM key-value caches for long context, and vector search indexes. The headline isn’t “new model, new benchmark trophy.” It’s cost and feasibility: long context is often limited by memory, and semantic search is often limited by storage and latency. If you can shrink those footprints while keeping quality steady, you can serve more users per GPU, keep more context available, or store larger indexes without ballooning infrastructure. Even if the exact claims will be debated, the direction is consistent with where AI is headed: the performance battle is shifting from pure model quality to the economics of running the thing. 800V DC power for AI And speaking of infrastructure economics, there’s a growing push to rethink how AI data centers deliver power. The story here is a migration from traditional AC-heavy layouts toward high-voltage DC distribution, with 800V DC often cited as a target. The motivation is straightforward: every conversion step wastes energy and adds heat and hardware. That’s tolerable at “normal” rack densities, but AI racks are climbing into territory where current, copper, and losses become a serious bottleneck. The interesting subtext is that scaling AI isn’t only about better GPUs and better cooling anymore—it’s about the electrical architecture of the building. The hard part won’t be the idea; it’ll be standards, safety practices, and an ecosystem of components that makes DC as routine to deploy as AC is today. VitruvianOS revives BeOS-like desktop On the operating system front, VitruvianOS is an open-source project trying to bring back the feel of classic BeOS and Haiku-style desktops—fast, responsive, and coherent—while still running on modern hardware via Linux. The pitch is less about chasing the latest UI trend and more about reclaiming that “instant reaction” desktop experience, with low-latency tuning and custom plumbing to support Haiku-like application behaviors. Whether it becomes a daily driver for many people is an open question, but it’s a reminder that a lot of users miss software that feels snappy and user-owned by default—and that nostalgia can be a productive design constraint, not just a vibe. C++ coroutines via game loops Finally, a practical programming story for the game-dev and systems crowd: a write-up argues C++ coroutines click faster when you think about Unity’s C# coroutines—small behaviors that unfold over multiple frames without turning into a messy manual state machine. The point isn’t that every C++ project should become coroutine-first; it’s that generator-style coroutines are now straightforward enough to be genuinely useful, especially for frame-by-frame logic, scripted effects, or time-sliced behaviors. In other words, this is C++ adopting a more ergonomic way to express “do a bit now, continue later,” which is a pattern that shows up everywhere from games to UI to simulations. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad - Consensus: AI for Research. Get a free month - https://get.consensus.app/automated_daily - Build Any Form, Without Code with Fillout. 50% extra signup credits - https://try.fillout.com/the_automated_daily Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: PyPI supply-chain malware scare - A suspected malicious PyPI release of litellm used a .pth auto-execution trick to steal credentials (AWS/GCP/Azure), SSH keys, and tokens—raising urgent credential-rotation and CI/CD risk concerns. Windows 11 usability reset - Microsoft’s seven-point Windows 11 “fix” plan targets ads, Copilot clutter, and missing UX basics, but leaves privacy, forced Microsoft account setup, telemetry limits, and OneDrive lock-in largely intact. Missile defense math meets reality - A new analysis frames missile defense as resource allocation under uncertainty: interceptor inventories, sensor reliability, and decoys can overwhelm even strong optimization, making high-confidence defense hard at scale. Linux compressed swap: zswap vs zram - Kernel developer commentary argues zswap generally degrades more predictably than zram under pressure, with fewer pathological behaviors and less risk of long stalls—important for servers and desktops alike. ripgrep benchmark lessons for search - A deep benchmark-and-design write-up on ripgrep highlights why real-world code search hinges on correctness, Unicode handling, and filesystem traversal—not just raw regex speed. Streaming OS images over network - A Linux imaging post shows the appeal and danger of streaming a disk image straight onto a block device: it’s elegant for deployments, but unsafe if you overwrite the disk you’re running from. Apartment gate hacked the simple way - A DoorKing gate outage led residents to bypass the “smart” layers and trigger the lock at the wiring level, then wrap it in a standards-based smart-home control—showing physical security realities. Terminal log analysis with lnav - lnav demonstrates how far a local terminal tool can go for log triage—searching, filtering, and making noisy logs readable without standing up heavyweight logging infrastructure. - Critics Say Microsoft’s Windows 11 ‘Fix Plan’ Reverses Self-Inflicted Changes, Not Core Privacy Issues - Litellm PyPI Supply-Chain Attack Allegedly Adds Auto-Executing .pth Credential Stealer - Why Missile Defense Allocation Is NP-Complete—and Why Sensors Matter More Than Interceptors - Opera’s Web Rewind Offers an Interactive Timeline of 30 Years of the Web - Why zswap Usually Beats zram for Compressed Swap on Linux - Benchmarks Show ripgrep’s Speed Advantage and Why Unicode-Friendly Search Can Still Be Fast - How to Reimage Linux by Streaming a Disk Image Straight to /dev/sda - Hackers Restore Apartment Gate Access by Wiring an ESP32 Relay into the Solenoid and Apple Home - NanoClaw Switches to OneCLI Agent Vault to Keep API Keys Out of Agents and Enforce Access Policies - lnav.org Episode Transcript PyPI supply-chain malware scare Let’s start with the security story, because it’s the kind that can ripple from laptops to CI runners to production. A critical report alleges the PyPI package litellm version 1.82.8 shipped with a malicious .pth file—one of those Python startup hooks that can run code automatically when the interpreter launches. The nasty part: it could execute even if you never import the package. The claim is that it harvested high-value secrets—cloud credentials, SSH keys, Kubernetes tokens, and more—then encrypted and exfiltrated them to a domain that doesn’t match the project’s usual footprint. Commenters also warned the blast radius might extend beyond a single version. Why this matters is simple: supply-chain attacks scale quietly. If your environment installed the affected builds, the safer assumption is credential exposure, meaning rotation and incident review—not just uninstalling and moving on. Windows 11 usability reset Next up: Windows 11, and Microsoft’s attempt at a redemption arc. Windows leadership reportedly admitted the OS had “gone off track,” and now there’s a seven-point plan aimed at dialing back the most visible annoyances—think fewer ads, less forced Copilot presence, and restoring usability features people still miss, like more taskbar flexibility. The critique, though, is that many of these problems weren’t accidents; they were choices—promotional Start menu tiles, AI buttons spreading through core apps, and design decisions that made the system feel less like a product you bought and more like a surface for nudges. And while the plan targets what you can see, the piece argues it largely avoids the harder trust issues: being pushed into a Microsoft account during setup, consumer telemetry you can’t truly shut off, and OneDrive behaviors that can quietly pull your folders into sync. It also revisits Windows Recall as a cautionary tale: ambitious AI features can create new security and privacy liabilities, and “we’ll make it opt-in later” isn’t exactly a confidence builder. Missile defense math meets reality Staying with security—but shifting from PCs to geopolitics—there’s a compelling analysis arguing missile defense is, at its core, a resource-allocation problem under uncertainty. On paper, firing multiple interceptors at a target can raise your chance of success. In practice, that math assumes your sensors and tracking hold up, and that’s a huge assumption. If the radar picture degrades—through physical attacks, software issues, or sheer complexity—adding more interceptors doesn’t fix the fundamental problem: you’re spending scarce inventory on a guess. The deeper point is about scaling. Attackers can often make the defender’s job harder more cheaply than defenders can make it easier—by adding decoys, increasing simultaneous targets, or targeting the sensors and command pipeline. Even if the optimization software is excellent, you can’t compute your way out of missing or unreliable information. It’s a sobering reminder that “high-tech” defenses still hinge on fragile links. Linux compressed swap: zswap vs zram On the Linux front, there’s a strong opinion from kernel developer Chris Down on compressed swap: most systems, he argues, should prefer zswap over zram. The key difference isn’t ideology, it’s failure mode. When memory pressure rises, zswap tends to degrade in a smoother, more predictable way because it can spill colder pages out to disk swap when its in-RAM pool fills. zram, by contrast, behaves more like a fixed-size compressed RAM device. That can create situations where old, cold data squats in fast memory while newer, more-needed pages get pushed to slower storage—exactly when you want the opposite. This matters because performance under stress is what separates “my system slowed down” from “my system hung.” The post also challenges the simplistic claim that zram automatically saves SSD wear, suggesting the I/O pressure often just shifts around in less obvious ways. ripgrep benchmark lessons for search Now for something more developer-workflow oriented: a detailed write-up on ripgrep—rg—revisits why it’s fast in the places developers actually care about. The discussion isn’t just “tool A beats tool B.” It’s about what makes search feel instant in real repositories: respecting .gitignore correctly, avoiding binary and hidden files by default, distributing work in parallel, and handling Unicode without falling over. One interesting takeaway is that some classic “speed tricks” don’t always help. For example, memory-mapping can be great for a single huge file, but it can add overhead when you’re scanning tons of small files—exactly what codebases look like. The broader lesson: performance claims only matter when paired with correctness and sensible defaults, because the fastest wrong answer is still wrong. Streaming OS images over network If you’ve ever wished imaging a machine could be as easy as “download and install,” there’s a neat Linux post about streaming a raw disk image from the network directly onto a block device. It leans into the Unix idea that everything is a file, so you can pipe data straight into a disk without staging it locally. But the cautionary bit is the point: elegance doesn’t override physics. If you try to overwrite the disk you’re currently booted from, you’re basically sawing off the branch you’re sitting on, and the system can crash mid-transfer. The practical takeaway is to do this from a rescue or installer environment—or something RAM-backed—so the target disk isn’t also the one keeping your OS alive. It’s a good reminder that deployment shortcuts need a safe boot context, not just clever commands. Apartment gate hacked the simple way One of the most relatable stories today comes from an apartment complex with a DoorKing intercom that stopped working because the cellular service lapsed. Instead of waiting on management, a resident and friends explored options—then discovered the simplest path wasn’t hacking the intercom’s higher-level logic, but triggering the gate at the wiring level. They identified accessible control wires for the solenoid lock, installed a small relay board, and exposed it as a smart-home device using a modern interoperability standard, with an auto-relock to reduce risk. Why it matters: this is both a practical win and a security lesson. “Smart” access control often reduces to very ordinary electrical control points, and if those points are reachable, the real security question becomes physical protection and tamper resistance—not the app. Terminal log analysis with lnav Finally, a smaller but useful tool note: lnav, or Logfile Navigator, is a terminal log viewer aimed at making messy logs easier to search, filter, and understand without spinning up a server-side logging stack. The appeal here is speed-to-value: point it at files, and you get a more readable, queryable view right where you’re already working. Why it matters in 2026: as systems sprawl, the temptation is to centralize everything immediately. Tools like this remind us there’s still room for lightweight, local-first debugging—especially when you’re on-call, SSH’d into a box, and just need answers quickly. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad - Build Any Form, Without Code with Fillout. 50% extra signup credits - https://try.fillout.com/the_automated_daily - Prezi: Create AI presentations fast - https://try.prezi.com/automated_daily Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Cigarette-lighter hardware fault attacks - A researcher demonstrated electromagnetic fault injection with a simple piezo lighter, flipping DDR bits and chaining it into Linux root—raising real-world hardware security questions. EU-first switch for online tools - One blogger is moving email, hosting, DNS, and code hosting to EU-based providers for stronger GDPR-style protections, showing how geopolitics is reshaping personal infrastructure choices. POSSE and owning your content - IndieWeb’s POSSE approach—publish on your own site, syndicate elsewhere—keeps canonical URLs and long-term ownership while still leveraging networks like Mastodon or Medium. Ad-bloated web and RSS - A critique of an RSS article buried under overlays and heavy ads highlights why RSS remains a bandwidth-saving, distraction-free way to follow news. GitHub outages and contingency planning - GitHub reliability issues—including Actions and notification delays—underscore that even core dev platforms have downtime, so teams should design CI/CD with failure modes in mind. Property-based testing for UIs - Bombadil is an open-source, experimental property-based testing tool for web UIs, aiming to find surprising edge-case bugs beyond scripted test suites. RollerCoaster Tycoon performance design - A deep look at RollerCoaster Tycoon argues its speed came from design-for-performance choices, where constraints shaped gameplay behaviors as much as low-level code did. GM EV1 restoration returns - GM is backing an enthusiast-led effort to restore a rare, nearly-lost 1996 EV1, turning a historical footnote into a drivable artifact ahead of its 30th anniversary. - Blogger Moves Key Online Services to EU Providers for Privacy and Resilience - IndieWeb Guide Details POSSE: Publish on Your Site First, Then Syndicate Elsewhere - GitHub hit by repeated service issues, raising concerns about uptime - Blogger Calls Out PC Gamer RSS Article for Popups and Massive Ad-Driven Downloads - GM Joins Effort to Restore One of the Few Surviving, Drivable EV1s - Tin Can targets parents seeking a landline-style alternative to kids’ smartphones - Antithesis open-sources Bombadil for property-based testing of web UIs - How RollerCoaster Tycoon Achieved Its Legendary Performance - da.vidbuchanan.co.uk Episode Transcript Cigarette-lighter hardware fault attacks First up: a wild but thoughtfully documented hardware security experiment. David Buchanan showed that electromagnetic fault injection—normally associated with pricey lab gear—can be pulled off with a piezo-electric lighter igniter and some careful setup. The big takeaway isn’t that everyone can do this easily; it’s that the barrier to experimenting with physical fault attacks is lower than many assume. And once you can reliably induce memory errors, you can start chaining them into real exploits—like corrupting critical data structures to escalate privileges. For defenders, it’s another reminder: when an attacker has physical access, the threat model shifts fast, and “software-only” assumptions get shaky. EU-first switch for online tools Switching gears to digital sovereignty: a blogger detailed a broad migration away from non-EU online services and subscriptions toward EU-based alternatives, driven by the political climate and a preference for EU privacy protections. They moved email and hosting to Germany-based Uberspace, migrated domains and DNS to a German registrar, and shifted source code hosting to Codeberg, which is run as a nonprofit. What’s interesting here is the trade-off story: some EU options didn’t match the features they relied on, so they accepted extra complexity—like adding Nextcloud just to regain calendar and contacts sync. This is what “values-driven infrastructure” looks like in practice: fewer defaults, more intentional choices, and sometimes more maintenance in exchange for jurisdictional comfort. POSSE and owning your content That dovetails with a classic IndieWeb idea getting renewed attention: POSSE—“Publish (on your) Own Site, Syndicate Elsewhere.” The updated explainer argues for posting on your personal site first, then pushing copies or links out to social platforms while pointing back to the original. The why is straightforward: platforms come and go, policies change, and accounts get throttled or banned—but your own site can remain the canonical home for your writing, photos, and notes. The page also emphasizes something people forget: if you record where you syndicated a post, you can pull replies and reactions back to your site, keeping the conversation connected instead of scattered across walled gardens. Ad-bloated web and RSS And if you needed another reason to pull content away from the modern web’s clutter, there’s a sharp critique making the rounds: an article promoting RSS readers that’s so ad-heavy and overlay-packed that the page is barely usable. The author’s point is less about dunking on one publisher and more about the broader incentive problem—when revenue depends on attention mechanics, the reading experience becomes collateral damage. RSS matters here because it’s a simple, durable escape hatch: you get the text, you skip the popups, and you stop donating bandwidth to a small army of trackers and autoplay assets. GitHub outages and contingency planning Now to the developer platform that many teams treat like electricity: GitHub. A recent run of reliability problems included an incident that hit core workflows—CI runs, notifications, and more—plus a separate Copilot-related issue where newly enabled options didn’t reliably show up for some users due to internal propagation delays. The meta-issue is just as important: third parties are reconstructing historical availability because it’s become harder to gauge long-term reliability at a glance. Whether you love GitHub or not, the practical lesson is the same: it’s foundational infrastructure with real failure modes. If your deploy pipeline, incident response, or business operations assume it’s always there, you’re building on a single point of operational risk. Property-based testing for UIs On the tooling front, there’s a new open-source project from Antithesis called Bombadil, aimed at property-based testing for web UIs. Instead of only running pre-scripted click paths, the idea is to explore a user interface more autonomously and check whether important “should always be true” behaviors hold up under strange sequences and edge cases. Why it matters: UI bugs often live in the in-between states—those awkward transitions no one thinks to write a test for. Tools like this try to turn that uncertainty into something systematic, which is exactly what you want before a bug shows up in production with real users. RollerCoaster Tycoon performance design For a reminder that performance engineering is as much about product design as it is about clever code, a developer recap of a German podcast revisited RollerCoaster Tycoon. The post argues the game’s legendary smoothness wasn’t just because it used low-level programming—it was because the entire simulation was designed to avoid expensive work most of the time. Guests don’t constantly demand perfect pathfinding. Crowds don’t require heavy collision logic. And when the game does need the costly computations, it keeps them bounded so you don’t get sudden frame drops. It’s a great case study in how constraints can shape behavior that players later interpret as “personality.” GM EV1 restoration returns Finally, a bit of EV history getting a second life: General Motors is supporting the restoration of a rare 1996 EV1 that resurfaced in rough shape and was picked up by an enthusiast team. GM’s involvement includes parts and access to expertise—an unusual twist, given the EV1’s complicated legacy and how few drivable examples remain. Beyond the nostalgia, this matters because it turns an influential but largely inaccessible vehicle into something the public can see and understand again. It also highlights how ideas from early EV engineering—battery management, controls, efficiency-minded design—echo into today’s mainstream electric platforms. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Invest Like the Pros with StockMVP - https://www.stock-mvp.com/?via=ron - Consensus: AI for Research. Get a free month - https://get.consensus.app/automated_daily - KrispCall: Agentic Cloud Telephony - https://try.krispcall.com/tad Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Running 397B AI on Mac - Flash-MoE shows Apple Silicon can run the massive Qwen3.5 397B MoE model by streaming experts from SSD and leaning on macOS page cache—consumer AI inference, unified memory, Metal. AI turns receipts into data - A 25-year receipt archive became an “egg-flation” dataset using SAM3 segmentation plus OCR and LLM extraction—document AI, messy scans, structured data, personal analytics. Why Windows native dev hurts - A Windows utility developer argues WinUI 3 and the Windows App SDK still force Win32 fallbacks and painful deployment choices—desktop APIs, packaging, code-signing, Electron pressure. Keeping Node WebSockets alive - Inngest fixed missed heartbeats by moving WebSocket management into Node worker threads—event loop starvation, reliability, CPU-bound workloads, TypeScript SDK. Linux KVM bug from sign-extension - A subtle C integer sign-extension corrupted a host TSS base on CPU migration, causing KVM crashes—virtualization reliability, kernel patch, hard-to-debug concurrency failures. JavaScript dependency bloat backlash - A critique of modern npm habits explains how legacy support, micro-packages, and outdated ponyfills inflate bundles—supply-chain risk, duplication, faster builds via audits. Making architecture diagrams readable - A guide catalogs why diagrams mislead: unlabeled resources, unconnected boxes, master-diagram overload, and AI-generated vagueness—system design communication, clarity, trust. LLMs as interview prep tutors - One developer used an LLM to cram algorithms in a week, then hit the limits under interview pressure—LeetCode patterns, debuggability, and what ‘learning’ really means. - Flash-MoE Streams a 397B MoE Model from SSD to Run on a 48GB MacBook Pro - A Developer’s WinUI 3 Experiment Highlights the Fragmentation of Windows Native App Development - AI Pipeline Mines 25 Years of Receipts to Track Egg Purchases - Inngest Moves WebSocket Heartbeats to Node Worker Threads to Prevent Event Loop Starvation - Ilograph Lists Seven Common Pitfalls That Undermine Architecture Diagrams - Sign-Extension Bug in x86 Descriptor Base Calculation Leads to Linux Kernel Fix - Tinygrad spotlights minimalist ML framework, hiring push, and tinybox shipping update - JavaScript Dependency Bloat: Old Runtime Support, Micro-Packages, and Stale Ponyfills - Developer Uses an LLM to Speedrun Google Interview Prep, Advances Despite Interview Stumble Episode Transcript Running 397B AI on Mac Let’s start with that eye-catching AI milestone. A new open-source project called Flash-MoE demonstrates an inference engine built in C, Objective-C, and Metal that can run Qwen3.5—a 397B-parameter Mixture-of-Experts model—on an Apple Silicon MacBook Pro with 48GB of unified memory. The trick is not pretending you can load a 200-plus-gigabyte model into RAM. Instead, it streams only the small subset of “experts” needed for each token directly from the SSD, leaning on macOS page cache instead of inventing a whole new caching layer. The headline result is production-quality performance in the low single-digit tokens per second range, which is slow compared to a data center—but wildly interesting for a laptop. The bigger point: storage bandwidth and OS caching are being treated as first-class parts of the inference stack, and that changes what “local” AI can look like. AI turns receipts into data Sticking with AI, there’s a wonderfully practical story about turning personal mess into usable data. A hobbyist scanned every receipt since 2001, then used AI tools to answer one oddly compelling question: how have egg purchases changed over time? Two coding agents sifted through over eleven thousand receipt files—PDFs, emails, photos, scans—and narrowed it down to hundreds of egg receipts with quantities and prices. What’s notable here is not the egg chart; it’s the workflow reality check. Classic OCR struggled on messy scans, especially when white receipts vanished into white scanner beds. A modern segmentation model—Meta’s SAM3—suddenly made that problem tractable, and then LLMs helped turn semi-readable text into structured records. It’s a snapshot of where document AI is actually useful: not magic, but extremely effective when you combine the right vision model with good extraction and fast error-correction tools. Why Windows native dev hurts Now, a shift from AI to the less glamorous parts of building software: Windows desktop development. Developer Domenic Denicola tried to build a small utility called “Display Blackout,” and came away arguing that modern native Windows app development is fractured enough to push reasonable people toward web-based shells like Electron or Tauri. The app itself isn’t exotic—it needs multi-monitor awareness, a borderless overlay, global hotkeys, startup registration, settings, and a tray icon. But the complaint is that WinUI 3 and the Windows App SDK don’t cover these basics cleanly, so you end up bouncing back to Win32 APIs via interop. Add in deployment headaches—like .NET not being universally present, binary size tradeoffs, and packaging flows that nudge you toward paid code-signing—and the “modern” path starts to feel like a maze. The broader takeaway: platform resets mean little if they leave holes developers have to patch with legacy code anyway. Keeping Node WebSockets alive On reliability engineering, Inngest published a postmortem-style write-up around a subtle failure mode: “no available worker” errors even when workers were actually running. The culprit was classic Node.js behavior—CPU-heavy user code can starve the main thread’s event loop, and if your heartbeats ride that same loop, they stop. The server assumes the worker died and stops routing work. The fix was to move connection management—WebSockets, heartbeats, reconnect logic—into a worker thread so it can keep its own event loop ticking even when the main thread is busy. What makes this interesting isn’t the brand name; it’s the pattern. As more apps blend network liveness with unpredictable user code, isolating the “I’m alive” path becomes a first-order design choice, not an optimization. Linux KVM bug from sign-extension If you like deep debugging stories, there’s also a great one from the Linux and virtualization world. An author tracked down a crashy, unpredictable x86/KVM issue that only showed up on real multi-core hardware when a hypervisor thread migrated between CPUs. The failure looked like the system randomly hanging, with cascading lockups—exactly the kind of thing that ruins your week. The root cause turned out to be a subtle C integer-promotion and sign-extension bug while reconstructing a base address from descriptor fields. In plain terms: a small signed value got shifted, became negative, and then contaminated the upper bits of an address, so the CPU sometimes consulted the wrong Task State Segment. The fix—casting to unsigned 64-bit types before shifting—sounds tiny, but the impact is huge: fewer “haunted machine” failures in low-level virtualization paths, and a reminder that undefined-looking behavior can still be a single bitwise mistake. JavaScript dependency bloat backlash Over in the JavaScript ecosystem, there’s a critique of why modern projects keep getting heavier. The argument is that a lot of “JavaScript bloat” isn’t because apps are doing more—it’s because dependency habits keep dragging in code most users don’t need anymore. Some packages stick around for ancient runtime support and edge cases, others embrace a micro-module style where trivial logic becomes a full dependency chain, and ponyfills don’t always get removed when runtimes catch up. The result is bigger installs, duplicated code, and a larger supply-chain surface area to audit and secure. The interesting part is the proposed cultural shift: make the lean path the default, and let specialized compatibility be something you opt into, not something everyone pays for quietly. Making architecture diagrams readable There’s also a thoughtful follow-up on system architecture diagrams—specifically, why so many of them confuse more than they clarify. The guide calls out recurring problems: boxes labeled only by generic types instead of real names, components that appear unconnected so you can’t tell what they do, and the temptation to build one giant “master diagram” that tries to show runtime behavior, infrastructure, and deployment all at once. It also warns about oversimplifying complex interactions into a neat left-to-right pipeline, and about intermediaries like brokers that can hide who actually talks to whom. One timely point: AI-generated diagrams from code often look plausible but drift into vagueness, because deciding what matters is still a human judgment call. The payoff here is practical—better diagrams don’t just look nicer; they reduce operational risk by making the real system legible. LLMs as interview prep tutors Finally, a human story with a technical edge: a developer got an unexpected invitation to interview at Google with only a week to prepare, and used an LLM as a tutor to cram algorithms and data structures through timed practice. The account is refreshingly honest: rapid gains in pattern recognition, but also a sharp cliff when stress hits and you have to write correct code without leaning on a compiler and tests. In the interview, they could reason through a traversal, then froze on an iterative binary search pattern and ended up talking it out with buggy code on the page. The takeaway isn’t “LLMs work” or “LLMs fail”—it’s that tutoring can accelerate familiarity, but interviews still reward debuggability, careful edge-case thinking, and staying calm when the problem doesn’t match your freshest pattern. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Effortless AI design for presentations, websites, and more with Gamma - https://try.gamma.app/tad - Build Any Form, Without Code with Fillout. 50% extra signup credits - https://try.fillout.com/the_automated_daily - Prezi: Create AI presentations fast - https://try.prezi.com/automated_daily Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Publishers block the Wayback Machine - EFF says major publishers like The New York Times are blocking Internet Archive crawling, threatening the Wayback Machine, web preservation, and verifiable journalism records amid AI scraping lawsuits. Mamba-3 shifts AI toward inference - Together AI and collaborators released Mamba-3, a state space model optimized for inference latency and throughput, highlighting a broader shift toward deployment-heavy AI workloads like agents and RL. Open-source coding agents go desktop - OpenCode shipped a beta desktop app for macOS, Windows, and Linux, positioning a model-agnostic AI coding agent with privacy-first claims, multi-session workflows, and broad LLM provider support. Atuin makes terminal history smarter - Atuin v18.13 improves shell history search speed and terminal UX, and adds an opt-in AI command helper with safety confirmations and privacy controls—useful for day-to-day developer workflows. BYD pushes megawatt EV charging - BYD claims up to 1,500 kW “Flash Charger” speeds, reframing EV charging as an infrastructure and grid problem as much as a battery problem, despite impressive five-minute top-ups. FFmpeg explained for app builders - A practical overview demystifies FFmpeg’s architecture—how tools map to libraries like libavformat and libavcodec—helping developers integrate decoding without getting lost in the ecosystem. FilmKit brings Fujifilm tools to browser - FilmKit is an open-source, WebUSB-based preset manager and RAW converter for Fujifilm cameras that uses the camera’s own processor, offering a cross-platform, zero-install alternative in beta. Molly guards and safer interfaces - A reflection on “molly guards” connects physical safety covers and software confirmations, and introduces ‘reverse molly guards’—timeouts and defaults that prevent long-running systems from stalling. - OpenCode launches beta desktop app for its open-source AI coding agent - Mamba-3 Debuts as an Inference-Optimized State Space Model with Open-Source Kernels - FFmpeg 101: Understanding FFmpeg Tools, Libraries, and a Minimal Demux/Decode Loop - BYD’s 1.5-Megawatt EV Charging Nears Gas-Pump Speeds, but Infrastructure Limits Remain - Atuin v18.13 boosts search speed, adds Hex PTY proxy, and introduces opt-in shell AI - Trigger.dev’s TRQL adds secure, tenant-isolated SQL querying on shared ClickHouse - Glossary Explains Japanese Chopstick Etiquette Mistakes and Taboos - FilmKit brings browser-based Fujifilm preset management and in-camera RAW conversion via WebUSB - Explaining ‘Molly Guards’ and the Case for Reverse Safety Prompts - EFF Warns Publisher Blocks on Internet Archive Threaten the Web’s Historical Record Episode Transcript Publishers block the Wayback Machine First up: the Electronic Frontier Foundation is raising the alarm that major publishers have started blocking the Internet Archive from crawling their sites, which threatens the completeness of the Wayback Machine—especially for news. The backdrop is the escalating fight over AI scraping and copyright. Publishers want leverage, and they’re looking for ways to reduce bulk access to their content. But EFF’s point is blunt: blocking a nonprofit archive doesn’t meaningfully stop AI companies, while it absolutely does weaken a public record that journalists, researchers, courts, and Wikipedia rely on to prove what was published—and when. If archived snapshots get spotty, accountability gets harder. Quiet edits and deleted pages become much easier to deny. Mamba-3 shifts AI toward inference Staying in the AI lane, a new model architecture called Mamba-3 just dropped from Together AI and academic and industry collaborators—and the headline is a shift in priorities. Instead of optimizing for training speed, Mamba-3 is tuned for inference efficiency: the real-world cost and latency of generating tokens once a model is deployed. That matters because modern AI work is increasingly dominated by post-training and usage—think agentic coding workflows, tool use, and evaluation loops where models spend their lives generating, not learning. The team is essentially saying: the bottleneck isn’t building the brain, it’s running it at scale. They’re also open-sourcing high-performance GPU kernels to make those claimed speedups more than a paper exercise. If the results hold up across more setups, this is one more nudge toward a future where “faster at serving” is as strategically important as “bigger at training.” Open-source coding agents go desktop On the developer workflow front, OpenCode—an open-source AI coding agent—has expanded into a beta desktop app for macOS, Windows, and Linux. The interesting angle here isn’t just “yet another coding assistant.” It’s the model-agnostic posture: bringing your own provider, whether that’s hosted APIs, local models, or even existing accounts like ChatGPT subscriptions or GitHub Copilot. For teams, that flexibility can be the difference between experimenting and standardizing. OpenCode is also leaning into collaboration features—parallel sessions on the same project and shareable session links—plus a privacy-first promise that it doesn’t store your code or context. For anyone working with sensitive repos, that claim is the entire ballgame, assuming the implementation matches the messaging. Either way, the broader story is clear: AI dev tools are moving from “plugin you try” to “workspace you live in.” Atuin makes terminal history smarter Another tool that’s evolving fast: Atuin, the shell history and sync utility, just shipped version 18.13. The big win is responsiveness—history search gets a hot in-memory index so it can feel instant even when your command history is enormous. That’s not glamorous, but it’s the kind of speed upgrade you notice every day. It also adds an optional AI command helper that turns plain-English intent into shell commands, with safety friction for risky operations and privacy defaults that limit what context gets shared unless you explicitly allow it. The theme is practical AI: not magical automation, just a faster path from “what do I type?” to “here’s a command you still need to approve.” BYD pushes megawatt EV charging Now to hardware and infrastructure: BYD says it has upgraded its EV “Flash Charger” system to deliver up to 1,500 kW—enough, it claims, to jump from roughly 10 to 70 percent in about five minutes on compatible vehicles. If you’ve been following EV charging, you know why this is provocative: it’s approaching the time it takes to buy a coffee, not the time it takes to watch a show. But the real constraint is what comes after the press release. Megawatt-level charging pushes the problem into grid capacity, site upgrades, and often on-site storage to avoid hammering local networks. And there’s a more subtle takeaway: even if five-minute charging becomes common, it may not transform daily ownership, because most charging still happens at home. The “why it matters” is less about convenience for everyone, and more about enabling long-distance driving at scale—if the infrastructure can keep up. FFmpeg explained for app builders For the builders in the audience, there’s a solid high-level explainer making the rounds on FFmpeg’s architecture. FFmpeg is one of those ecosystems a lot of software depends on, but many developers only touch through copy-pasted command lines. This piece connects the dots between the familiar tools—ffmpeg, ffplay, ffprobe—and the underlying libraries used inside apps, like the components that handle container formats and the ones that decode and encode audio and video. The reason it’s useful is confidence: once you understand the boundaries—what reads streams, what decodes frames, what converts formats—you can integrate media capabilities into your own program without treating FFmpeg as an untouchable black box. FilmKit brings Fujifilm tools to browser A more niche, but genuinely clever open-source launch: FilmKit, a browser-based preset manager and RAW converter for Fujifilm X-series cameras. The standout idea is using WebUSB to talk directly to the camera, and then letting the camera’s own processor do the RAW-to-JPEG conversion. In practice, that could mean a lighter, cross-platform workflow—no heavyweight desktop suite—while still getting results consistent with what the camera would produce. It’s early and labeled beta, with limited device testing so far, but it hints at a broader trend: browsers becoming capable hardware companion apps, especially when open tools can reverse-engineer protocols responsibly and build community-tested compatibility. Molly guards and safer interfaces And to close, a small concept with big implications: the “molly guard”—that flip-up plastic cover over an important button, designed to stop accidental activation. It’s a reminder that good systems design often includes intentional friction. In software, we do it with confirmation dialogs and key combos that prevent catastrophic clicks. But the post also points out an opposite pattern: what you might call a reverse molly guard—interfaces that default safely after a timeout instead of waiting forever. That matters in operations: a single prompt that blocks a batch job overnight can be more expensive than an imperfect default choice. Sometimes the safest system is the one that keeps moving—carefully—when nobody is watching. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)

Please support this podcast by checking out our sponsors: - Prezi: Create AI presentations fast - https://try.prezi.com/automated_daily - Discover the Future of AI Audio with ElevenLabs - https://try.elevenlabs.io/tad - Build Any Form, Without Code with Fillout. 50% extra signup credits - https://try.fillout.com/the_automated_daily Support The Automated Daily directly: Buy me a coffee: https://buymeacoffee.com/theautomateddaily Today's topics: Invisible Azure Entra logins - Researchers found patched Azure Entra ID techniques that could mint OAuth tokens without sign-in log entries—undermining audit trails, detection, and incident response. Iberian blackout final findings - ENTSO-E’s expert panel says Spain–Portugal’s 2025 blackout came from interacting grid-control weaknesses, highlighting voltage control, coordination, and updated market incentives for resilience. Android sideloading verification shift - Google plans stricter Android sideloading rules tied to developer verification, with a slower hidden bypass aimed at scam mitigation—raising openness, privacy, and access concerns. GPU k-means becomes practical - A new Flash-KMeans paper shows exact k-means can run far faster on GPUs by avoiding bandwidth-heavy distance matrices and reducing update contention—useful for online AI pipelines. FFmpeg’s new GPU video path - FFmpeg is pushing Vulkan compute shaders to keep codec work GPU-resident, reducing CPU handoffs and making high-end media workflows more portable and less proprietary. Scriptable vector overlays in FFmpeg - FFmpeg 8.1’s drawvg filter brings a compact vector scripting language for time-aware overlays and effects, improving automation and reproducibility in video pipelines. Safer CSS color precision - CSS minification research argues most colors are over-precise; consistent rounding by color space can shrink code while avoiding visible drift in color-mix and computed palettes. Mapping extreme downtown land value - Land-value maps show extreme central-city concentration—like Manhattan dwarfing outer boroughs—helping policy debates on assessment, vacancy, and surface parking opportunity costs. De-escalation lessons from pedicabbing - A pedicab driver’s essay describes how mindset shifts and gentler interactions reduced conflict and burnout—an everyday case study in safety and emotional resilience. - ENTSO-E report details causes and fixes after April 2025 Spain-Portugal blackout - FFmpeg Adds Vulkan Compute Shader Path for GPU Video Encode/Decode - Flash-KMeans Targets GPU Bottlenecks to Make Exact K-Means Fast Enough for Online Use - Oslo Pedicab Driver’s Essay on Stress, Expectations, and a Gentler Ride - Google adds 24-hour delay option to sideload unverified Android apps - FFmpeg 8.1 Introduces drawvg Filter for Scripted Vector Overlays - Mapping Tools Reveal How Land Value Spikes in City Centers - Researchers Reveal Patched Azure Entra Sign-In Log Bypasses That Enabled ‘Invisible’ Token Logins - Study Finds 3 Decimal Places Are Enough for Most CSS Colors Episode Transcript Invisible Azure Entra logins Let’s start with cloud security, because this one cuts straight to trust. Security researcher Nyxgeek disclosed two newly discovered Azure Entra ID sign-in log bypass techniques—both now patched—that could issue valid OAuth tokens without producing corresponding sign-in log entries. Why that’s important: many organizations treat Entra’s sign-in logs as the backbone of detection and forensics. If an attacker can authenticate and obtain usable tokens while leaving the sign-in trail blank, monitoring teams can miss password-spraying, suspicious access, and even active sessions. The write-up also highlights a practical consequence: defenders may need to correlate other signals—like Microsoft Graph activity—against sign-in records to spot “activity with no login,” which can be harder if the strongest logging features sit behind higher-tier licensing. The broader lesson is uncomfortable but clear: audit logs are a product surface too, and they can fail in ways that directly change risk. Iberian blackout final findings Next up, Europe’s power grid—and a reminder that modern infrastructure often fails by interaction, not by single-point breakage. ENTSO-E has published the final report on the April 28th, 2025 total blackout across continental Spain and Portugal, described as the most severe European blackout in over two decades. The expert panel’s conclusion: there wasn’t one decisive component that failed. Instead, several factors piled on—system oscillations, gaps in voltage and reactive power control, inconsistent voltage-regulation practices, rapid reductions in generation output, and generator disconnections in Spain. Together, that mix pushed voltage up quickly and triggered a cascade of generation trips that collapsed supply across the Iberian Peninsula. Why it matters beyond Spain and Portugal: the report argues that local technical weaknesses can propagate across interconnected networks, so cross-border coordination and consistent expectations aren’t bureaucracy—they’re resilience. It also calls out an increasingly common friction point: regulatory and market incentives have to evolve with the physical grid, or operators end up rewarded for behavior that stresses the system. Android sideloading verification shift Staying with the theme of “safety vs openness,” Google has outlined an advanced flow that keeps Android sideloading alive for experienced users—even as the platform moves toward mandatory developer verification starting in September 2026. The headline change is philosophical as much as technical: Android is historically known for letting you install apps from almost anywhere. Under the new model, unverified developers are blocked by default, and the bypass is intentionally buried in Developer Options and slowed down with a waiting period. Google’s stated goal is to blunt social-engineering scams that pressure victims to install something malicious immediately. But it also raises predictable concerns: identity checks and fees can chill legitimate independent distribution, and people in regions with verification hurdles—or under sanctions—could be disproportionately affected. The rollout beginning in a few countries before expanding globally also signals that Android policy is increasingly shaped by threat patterns, not just developer preference. GPU k-means becomes practical Now to AI infrastructure, where a new research paper argues that a classic algorithm is ready to move from offline preprocessing into real-time systems. An arXiv paper introduces Flash-KMeans, a GPU-focused implementation meant to make exact k-means clustering fast enough to be used as an online primitive. In plain terms, k-means is everywhere—bucketing embeddings, organizing data, speeding up search—but teams often settle for approximations or batch jobs because the “exact” version can be too slow or too expensive. The paper’s claim is that the bottleneck isn’t the math—it’s how current GPU implementations move and update data. Flash-KMeans avoids the worst memory blowups and reduces the kind of contention that can make GPUs stall out. If these results generalize, it could make exact clustering a practical building block inside latency-sensitive AI pipelines, instead of something you run overnight and hope still matches reality tomorrow. FFmpeg’s new GPU video path Let’s talk video tooling, because FFmpeg quietly keeps redefining what “baseline” media infrastructure can do. First, a post from Lynne, a Vulkan maintainer in the FFmpeg world, describes how FFmpeg is using Vulkan compute shaders to accelerate video encoding and decoding on GPUs—without leaning on fixed-function video engines. The key idea is to keep more of the codec pipeline on the GPU instead of bouncing between CPU and GPU. Those handoffs can erase the performance gains you think you’re getting, especially on serious workloads like high-resolution mastering, VFX pipelines, or archival scans where decoding and scrubbing speed shapes the whole workflow. What’s notable here is portability and leverage: Vulkan is vendor-neutral, and compute shaders are a more general path than relying on whatever dedicated video blocks happen to exist on a specific chip. It’s not a promise that every codec becomes magically fast, but it’s a credible push toward GPU acceleration that isn’t locked to one ecosystem. Scriptable vector overlays in FFmpeg Also in FFmpeg 8.1: a new filter called drawvg that renders vector graphics directly onto video frames. This matters less for “cool demos” and more for automation. drawvg uses a compact scripting language to describe shapes and drawing commands, and it can react to time, frame size, metadata, and even sampled pixel colors. The practical outcome is that you can generate consistent, reproducible overlays—timelines, annotations, custom transitions—without bouncing out to a separate compositing tool. If you run media pipelines at scale, these kinds of native, scriptable building blocks reduce glue code and make renders easier to audit and repeat. Safer CSS color precision On the web front, there’s a surprisingly nuanced piece about CSS color minification. Keith Cirkel reports findings from work on csskit showing that most authored color values are far more precise than humans can perceive—and that minifiers should round them in a consistent, color-space-aware way. The interesting twist is that “just round more aggressively” can backfire when colors are computed repeatedly—think color-mix, palette ramps, or chained transformations. Small rounding errors can accumulate into visible drift over time or across a design system. The proposal is pragmatic: instead of expensive perceptual calculations on every color, use a reliable precision table per color space. It’s a reminder that performance work isn’t only about bytes—it’s about predictability under real-world composition. Mapping extreme downtown land value Switching gears to cities and data: the Center for Land Economics argues that people routinely underestimate how intensely land value concentrates in city centers. Their maps show central business districts towering over surrounding neighborhoods—Manhattan versus the Bronx is the attention-grabber, but the pattern appears in smaller cities too. Why this matters: once you visualize assessed land value, policy debates get sharper. You can see where development pressure really is, where under-assessment may be hiding, and where high-value land is being used for low-intensity purposes—like surface parking. They also highlight tooling upgrades—faster mapping pipelines and analysis utilities—that make this kind of accountability work easier for more jurisdictions. The bigger point is that better maps can change what decision-makers consider “obvious.” De-escalation lessons from pedicabbing Finally, a more human story that still connects to tech culture: a personal essay from an Oslo pedicab driver about burnout, conflict, and de-escalation. The driver describes harassment from intoxicated riders and the emotional drain of constant vigilance, then explains how taking breaks and changing expectations helped him keep going. The practical shift was toward calmer, more accommodating interactions—slowing down in crowds, spending less energy on “winning” each confrontation, and treating occasional bad encounters as an occupational cost rather than a personal failure. Why it’s worth your time: it’s a grounded reminder that safety and sustainability—whether you’re on a bike, in customer support, or on-call for production incidents—often comes down to systems and mindset, not heroics. Subscribe to edition specific feeds: - Space news * Apple Podcast English * Spotify English * RSS English Spanish French - Top news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - Tech news * Apple Podcast English Spanish French * Spotify English Spanish Spanish * RSS English Spanish French - Hacker news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French - AI news * Apple Podcast English Spanish French * Spotify English Spanish French * RSS English Spanish French Visit our website at https://theautomateddaily.com/ Send feedback to [email protected] Youtube LinkedIn X (Twitter)