the droids newsletter podcast

Welcome to the DROIDS Newsletter daily update for June 10, 2026...A quick reminder this news is selected by a human and read by an AI. Singapore-based YY Group Holding is deploying Unitree G1 humanoid robots across commercial facilities in Asia, targeting cleaning and maintenance as an early use case. The Unitree G1 platform offers agile bipedal mobility, touch-sensitive hands, and runs on NVIDIA’s Jetson Orin, positioning it for real-time operation in environments like malls and hotels. Rather than treating humanoids as a drop-in replacement, YY Group is equipping its cleaning staff with data-collection gear to record real-world workflows on site. Those shifts generate structured datasets that are used to train the robots for autonomous sanitation and basic maintenance tasks, with robots tied into the company’s workforce platform YY Circle and its facility management system 24IFM. The initiative builds on a new Humanoid Robotics Training Lab in Singapore and an AI training data platform that taps a reported 500,000 workers across 12 countries. In a humanoid market where Chinese manufacturers such as Unitree already account for most global shipments and costs are falling quickly, YY Group is positioning itself less as a robotics hardware buyer and more as a services company building a proprietary data and software layer on top of emerging physical AI. #robotics #physicalai #trainingdata #roboticnews #technews #dailyroboticnews #droidsnewsletterTune in tomorrow for more robotics news from DROIDS Newsletter.

Airbus Unveils U145 Autonomous Helicopter With No CockpitAirbus Helicopters has unveiled the U145, an autonomous, uncrewed version of its widely used H145 helicopter. The aircraft made its debut this week at the ILA Berlin Air Show, where Airbus displayed a full-scale mock-up of the new platform.Unlike the conventional H145, the U145 has no physical cockpit. Airbus has removed the pilot compartment entirely and redesigned the aircraft around cargo transport and autonomous operations. The helicopter will rely on a specialized sensor suite and artificial intelligence systems to navigate and complete missions without a crew onboard.The U145 is not a clean-sheet design. Instead, Airbus is building on the existing H145 platform, one of the world’s most established light twin-engine helicopters. More than 1,800 H145 aircraft are currently in service globally, with over 8.5 million flight hours logged.Airbus says the aircraft is being developed primarily for high-volume cargo missions. To support that role, the U145 includes an integrated nose-loading door, a foldable loading table, and a dedicated cargo floor. The company also describes the aircraft as a multi-mission platform that could eventually support disaster response, firefighting, surveillance, reconnaissance, and military resupply operations.A first flight with a safety pilot onboard is planned before the end of 2026. Airbus is targeting entry into service at the beginning of the next decade.The announcement reflects a broader trend across aerospace: rather than designing entirely new aircraft, manufacturers are increasingly converting proven vehicles into autonomous systems. Airbus has already followed a similar path with its VSR700 drone helicopter program.For now, the U145 remains a prototype. But if Airbus meets its timeline, one of the world’s most recognizable helicopters may soon be flying cargo missions without a pilot onboard.#robotics #droidsnewsletter #dailyroboticnews #technews

PepsiCo is quietly scaling up driverless trucking for its snack business.The company is working with autonomous trucking startup Gatik to operate a fleet of forty‑one driverless box trucks across Arizona, Texas, and Arkansas. The trucks move Frito‑Lay products like Doritos on fixed, middle‑mile routes between PepsiCo facilities and large retail distribution points.These are medium‑duty box trucks running without a driver in the cab on public roads. Gatik focuses on short, repeatable routes, which are easier to automate than long‑haul or complex urban delivery.Since launching fully driverless freight operations in mid‑2025, Gatik says it has completed tens of thousands of driverless orders without reported incidents, and now counts large retailers and consumer brands among its core customers.For PepsiCo, the appeal is a mix of cost control, predictable service levels, and relief in a tight labor market for commercial drivers. The deployment also fits alongside the company’s push into lower‑emission freight, including battery‑electric trucks on other routes.More broadly, the move signals that autonomous trucking is shifting from pilot programs to day‑to‑day operations in specific, controlled corridors, with major shippers now treating it as part of their standard logistics toolkit... #robotics #roboticnews #dailyroboticnews #pepsi #autonomy #autonomoustruck #droidsnewsletter

Today in robotics.Hyundai and Boston Dynamics have launched a World Cup-themed campaign that follows Atlas, the company’s humanoid robot, as it learns soccer.The five-part series, called School of Football, culminates with Atlas attempting a Ghost Rabona, a technically demanding cross-leg kick. Hyundai released the campaign as part of its lead-up to FIFA World Cup 2026, where the automaker serves as both FIFA’s mobility and robotics partner.The campaign comes as Boston Dynamics expands the public demonstrations of its electric Atlas platform. Previous showcases emphasized locomotion, including running, jumping, and parkour. Soccer introduces a moving object and a less predictable environment.Researchers have long used soccer as a robotics benchmark because it combines perception, locomotion, planning, and control in a single task. The annual RoboCup competition was founded on the same premise.The World Cup campaign coincides with Hyundai’s first official robotics deployment at a FIFA tournament. During the event, four Boston Dynamics Spot robots will be assigned to patrol and inspection duties at the International Broadcast Center in Dallas and at the New York–New Jersey stadium.#roboticnews #dailyroboticnews #worldcup #worldcupnews #droidsnewsletter #bostondynamics

Today is Saturday, June 6, 2026. A reminder that DROIDS Daily News is selected and verified by a human, and then read by an AI.At GTC Taipei, NVIDIA unveiled the Isaac GR00TReference Humanoid Robot, an open platform designed to give researchers a complete humanoid development system right out of the box...The platform combines a Unitree H2 Plus humanoid body, Sharpa Wave tactile robotic hands, NVIDIA’s Jetson Thor AI computer, and the company’s Isaac GR00T software stack into a single integrated system...Researchers can use the platform for data collection, simulation, training, teleoperation, deployment, and experimentation with humanoid foundation models...Several leading institutions have already committed to using the platform, including Stanford Robotics Center, ETH Zurich, Ai2, and UC San Diego... The bigger story is NVIDIA’s strategy...Rather than building a humanoid robot of its own, NVIDIA is creating a common hardware and software foundation that researchers and robot makers can build upon. In many ways, it resembles the role that standardized PC architectures played in accelerating personal computing...If researchers, startups, and manufacturers converge on the same tools, data pipelines, simulation environments, and AI models, humanoid development could accelerate significantly...In AI, NVIDIA became the company that everyone builds on...Now it’s trying to do the same thing for robots. Tune in tomorrow for more robotics news...

BYD is pushing into humanoid robotics as a new core business, with Executive Vice President Stella Li announcing the Chinese automaker is “advancing independent research and development of humanoid robots in full force,” according to remarks reported on May 26.Li said humanoid robots will become a core business track for BYD alongside its new energy vehicle operations. The company’s initial deployment will focus on its global 4S dealership network, where robots will handle sales guidance and customer reception duties, providing what BYD described as “new support for the intelligent transformation of its retail outlets.”From Factory Floor to ShowroomThe robotics push builds on BYD’s “Yao-Shun-Yu” project, which was first launched in late 2024 under the company’s 15th Division. BYD had already begun deploying robots in production environments in 2025, and the company showcased a robot at the Brussels Motor Show in January 2026 that “sparked conversations” with visitors.Li argued that BYD’s automotive manufacturing expertise gives it an edge in the robotics race. “The complex software systems and precision hardware manufacturing experience accumulated in the automotive industry can be directly transferred to robot development, drastically minimizing challenges in technological breakthroughs,” she said.Household AmbitionsBeyond dealerships, Li outlined longer-term plans for robots to enter homes, taking on daily tasks such as cleaning, cooking, and companionship — building what BYD calls “a full-scenario intelligent service ecosystem.”The announcement comes as Chinese companies increasingly dominate the global humanoid robotics landscape. At CES 2026, 58 percent of humanoid robot exhibitors were Chinese firms. In May 2026, the city of Hangzhou deployed 15 humanoid robots for traffic management, marking another milestone in China’s push to commercialize the technology.BYD’s robotics ambitions arrive at a complex moment for the company, which recently reported a sharp decline in profits even as it expands aggressively into new markets and technologies.Additional Reading for Inquisitive Minds:Humanoid. BYD humanoid robot development confirmed, dealer sales considered. June 4, 2026.CNEV Post. BYD enters humanoid robot market, may sell through dealer network. June 3, 2026.Carbon Credits. BYD Overtakes Tesla as World’s Biggest EV Seller in 2025. January 7, 2026.Car Scoops. BYD Sold 700,000 Electrified Cars Last Quarter And Still Lost More Than Half Its Profit. May 1, 2026.Forbes. Is China About To “BYD” Tesla’s Humanoid Dreams? February 20, 2026.CNN Business. Elon Musk once mocked China’s BYD. Now it’s running circles around Tesla. May 30, 2024.Editor’s Note: This daily script is carefully researched, selected and checked by a human, and read by an AI.#robotics #roboticsnews #dailyroboticnews #tesla #optimus #byd #humanoids

Automated bot traffic has surpassed human traffic on the internet for the first time, Cloudflare CEO Matthew Prince announced on June 3. According to Cloudflare’s Radar dashboard, bots now account for 57.5% of all HTTP requests directed at HTML content, while human-generated traffic has fallen to 42.5%.A Milestone Arrived Early“Welp, that happened faster than I predicted,” Prince wrote on X. “Thought it would be end of 2027, then early 2027, but agentic traffic growing so fast that bots have now passed human traffic online for the first time in the Internet’s history.”The crossover arrived roughly eighteen months ahead of Prince’s original forecast. At SXSW in March, the Cloudflare chief had predicted bot traffic would overtake humans by 2027, citing the rapid expansion of generative AI systems that can visit far more websites than any person. He illustrated the gap with a simple comparison: a human shopping for a digital camera might visit five websites, while an AI agent performing the same task could hit a thousand.The milestone underscores broader industry concerns about the internet’s shifting composition. A March report from cybersecurity firm Human Security separately found that automated traffic was growing eight times faster than human usage, with CEO Stu Solomon telling CNBC that “the internet was fundamentally conceived with the idea that a human operates the device, and that conception is being swiftly transformed.”[cnbc]Cloudflare powers approximately 20% of all websites and is used by roughly 80% of sites that employ a reverse proxy service, giving the company an unusually broad view of global web traffic patterns. The data raises questions for digital advertising, web infrastructure, and search — industries built on the assumption that the entity on the other end of a request is a person.Prince offered no indication the trend would slow, with AI agents becoming more capable and autonomous with each generation of models deployed.#roboticsnews #droidsnewsletter

Researchers gave Claude, Gemini, Grok, and GPT-5 Mini their own AI towns and let them govern for 15 simulated days.Claude built a stable society.Gemini stayed productive despite rampant crime.Grok’s town collapsed into violence.GPT-5 Mini went extinct.Then researchers mixed them together—and agents that behaved safely alone became unpredictable around other AI populations.AI civilization is apparently just as messy as ours. 🤖🏙️#aiethics #robotics #airesearch #droidsnewsletter

Anthropic says its experimental Claude Mythos Preview model uncovered more than 10,000 high and critical-severity software vulnerabilities in just one month.The results are staggering. Mozilla found over ten times more vulnerabilities in Firefox than during previous testing. Cloudflare uncovered thousands of bugs across critical systems. Open-source projects that power much of the internet were scanned at a scale never seen before.But there’s a catch.AI is now finding vulnerabilities faster than organizations can patch them. Anthropic reports that only a fraction of the most serious discoveries have been fixed so far, creating a growing backlog of security risks.In this episode, we examine what this means for software security, why Anthropic is keeping Claude Mythos private, and how cybersecurity may be entering a new era where the bottleneck is no longer finding vulnerabilities, but fixing them.DROIDS is researched, selected, and fact-checked by a human, then narrated by AI.Subscribe for daily coverage of robotics, artificial intelligence, and the technologies shaping our future. #roboticnews #ainews #anthropic #mythos #droidsnewsletter

NVIDIA is significantly advancing the field of physical AI by demonstrating how robots trained in virtual simulations can successfully transition to complex, real-world applications. Through a series of research papers presented at ICRA 2026, the company highlighted breakthroughs in robotic navigation, planning speed, and manual dexterity.#robotics #physicalai #nvidia #IRCA2026 #droidsnewsletter

Welcome to the DROIDS Newsletter weekly summary for the week of May 21 through 28, 2026...At Google I/O 2026, CEO Sundar Pichai announced Gemini Spark, a cloud-based personal AI agent that runs continuously on dedicated Google Cloud VMs, executing tasks even when user devices are powered down. The service begins rolling out to trusted testers this week, with a beta for Google AI Ultra subscribers in the United States expected next week and desktop integration planned for this summer...Google reported that OpenAI, Kakao, and ElevenLabs are adopting SynthID, significantly broadening cross-industry use of its invisible watermarking system for AI-generated images, video, and audio. SynthID has already been applied to more than 100 billion images and videos and the equivalent of 60,000 years of audio...Andrej Karpathy announced he has joined Anthropic, marking a return to frontier LLM research after two years focused on AI education at Eureka Labs. The move strengthens Anthropic’s senior technical leadership in the competitive AI model development race...Snowflake-AWS AI Infrastructure Deal.Snowflake shares jumped 36 percent in premarket trading Thursday after the data analytics firm announced a five-year AI infrastructure deal worth $6 billion with Amazon Web Services. The company also raised its annual product revenue forecast alongside the partnership announcement...AI Energy Efficiency Breakthrough.Researchers unveiled a radically efficient AI approach that could slash energy use by up to 100 times while improving accuracy. The system combines neural networks with human-like symbolic reasoning, addressing AI’s staggering energy demands, which already consume over 10 percent of U.S. electricity...In Science and Technology News...NASA Tests Next-Gen Space AI Chip.NASA is testing a radiation-hardened processor that delivers performance hundreds of times beyond current spaceflight computers. The next-generation chip enables spacecraft to operate independently in deep space, withstanding punishing conditions designed to mimic the harsh space environment...Agricultural Robot Achieves 81% Success Rate.A new agricultural robot predicts harvest difficulty before acting, adjusting its approach for each tomato rather than simply identifying ripe fruit. The smarter strategy boosted success rates to 81 percent, with the robot switching angles when needed, enabling farms where robots and humans work side by side...For all the latest robotics information, be sure to subscribe to DROIDS Newsletter. DROIDS is available on Substack, LinkedIn, Spotify, and Apple Podcasts.#robotics #droidsnewsletter #dianawolftorres #roboticsnews

Robotics startups are getting very good at building demos. Getting products through certification, testing and real-world deployment is much harder.In this episode, Tony Gao of Fuchsia (YC P26) discusses the hidden compliance bottlenecks slowing robotics and hardware startups as the industry moves into the real world.#Robotics #PhysicalAI #HardwareStartups #YCombinator

Spring 2026 graduation season brought an unexpected sound to college campuses across the country: a coordinated chorus of boos. From the University of Arizona to the University of Central Florida, top executives like former Google CEO Eric Schmidt were met with loud jeers when praising artificial intelligence as the "next industrial revolution". But why is the most tech-savvy generation actively rejecting the AI hype?In this episode, we dive deep into the "AI Paradox" defining Generation Z. We explore the complex reality of why young people use AI chatbots daily for schoolwork and everyday tasks, yet deeply distrust the technology's corporate rollout. With entry-level jobs vanishing, creative roles threatened by automation, and executives telling them to simply "deal with it," students aren't just rejecting a new tool—they are rejecting a precarious economic future.In this episode, we cover:The Commencement Backlash: Why graduation speeches about the promise of AI are sparking anger and protests from graduating seniors.The Usage Paradox: How 74% of working Gen Zers use AI in their day-to-day jobs, yet a staggering 69% trust human-made content while only 3% trust fully AI-generated work."Generation Entrepreneur": With the traditional corporate ladder broken by automation, how Gen Z is using low-code AI tools to bypass vanishing entry-level roles and build their own businesses.The Neo-Luddite Movement: The growing cultural push among youth for "digital decoupling," analog connections, and reclaiming human agency in a constantly automated world.#droidsnewsletter #aijobdisruption #techethics

In today’s DROIDS daily update for May 21, 2026, we zoom out on physical AI and humanoids as analysts start publishing macro forecasts that treat embodied robots as a multi‑trillion‑dollar market by mid‑century, with hundreds of millions of units projected in logistics, manufacturing, and services. We then break down Google’s latest video‑editing AI, a model that can take real footage and regenerate specific regions with frame‑consistent edits from text, sketches, or reference images. Finally, we look at May’s accelerating tech layoff wave—tens of thousands of roles cut across major players as budgets are aggressively reallocated toward AI infrastructure and automation. This one is all about where robots are really going, what new AI tooling can actually do, and who in tech is paying the price.Audio recorded using ElevenLabs.#robotics #roboticsnews #droidsnewsletter #dianawolftorres

In today’s DROIDS Daily, we break down Figure AI’s latest F.03 livestream, where three humanoid robots reportedly ran more than 80 hours and sorted over 100,000 packages in a tightly controlled warehouse test. She then dives into BMW’s 11‑month deployment of Figure 02 at Spartanburg, which helped build more than 30,000 X3s and delivered the first real ROI numbers for physical AI on an automotive line. The episode closes with a look at how non‑humanoid warehouse automation is scaling and what current cost curves and payback periods say about where humanoids actually fit in near‑term logistics and manufacturing.#robotics #roboticsnews #droidsnewsletter

Welcome to the DROIDS Newsletter daily update for May 18, 2026.Korea’s Ministry of Science and ICT has announced a national program to invest over 33.5 million dollars through 2030 to develop homegrown AI-powered humanoid robots for domestic and industrial use, positioning Korea as a strategic competitor in full-stack humanoid platforms that integrate high-torque actuators, whole-body control, and embedded large models for task planning. The roadmap emphasizes open middleware and domestic supply chains for key components, including servo drives, reduction gears, and safety-rated sensors, with an explicit goal of certifying these humanoids for operation in logistics, manufacturing, and public-service environments.WoAn Robotics, described as “China’s version of Figure,” has begun deploying its home-focused humanoid robots into overseas real-world scenarios, moving beyond lab demos into live pilots. The company is targeting household and hospitality tasks, building on a bipedal platform tuned for low-noise operation, dexterous manipulators for kitchen and cleaning workflows, and perception stacks optimized for cluttered indoor spaces, with deployments framed as a validation step before broader commercial scaling.Faraday Future has launched a new Physical AI robotics institute focused on integrating robotics into its vehicles and infrastructure, highlighting a push to fuse autonomous motion planning, embodied AI, and human–machine interfaces in automotive contexts. The initiative centers on mechatronic systems and embedded AI controllers that can handle tasks such as automated factory handling, in-vehicle robotic assistants, and dynamic inspection, with the institute positioned as an R&D hub for cross-disciplinary robotics and AI safety research within the company’s ecosystem.Scientists at the University of Cambridge have demonstrated a new class of light-emitting devices by electrically powering insulating nanoparticles using tiny organic “molecular antennas,” effectively creating an LED architecture that operates in a regime previously considered impossible for insulators. The work shows that these molecular antennas can funnel energy into the nanoparticles efficiently enough to generate light, opening a pathway to unconventional LED designs that could alter the materials palette for optoelectronics and enable new device geometries at the nanoscale.For all the latest robotics information, be sure to subscribe to DROIDS Newsletter.#robotics #roboticsnews #droidsnewsletter #dailyroboticnews

On paper, the robotics job market in 2026 looks “hot”: strong growth, six‑figure roles, and constant headlines about a talent shortage. But that’s not what many students and early‑career roboticists are experiencing on the ground. In this episode, I talk about the widening gap between the hype and the reality for new grads trying to break into robotics and AI.I share a quote from DROIDS Associate Editor Alexander Wolf Torres, who graduates in a few weeks and still hasn’t landed that first full‑time role despite internships, projects, and deep involvement in the field. His experience mirrors what I’m hearing from students and career‑switchers across the industry: ghosting, automated rejections, and a nagging sense that the “rules” quietly changed.We dig into why both things can be true at once—yes, companies are struggling to hire the right robotics talent, and yes, it’s brutally hard to get your foot in the door—and what kinds of skills, roles, and strategies actually move the needle for early‑career folks.If you’re hiring in robotics or automation, or trying to break into the field yourself, this episode is for you.#robotics #roboticsjobs #aijobs #hiring

Next-gen space AI: NASA is testing a radiation-hardened AI chip delivering hundreds-fold performance gains for onboard autonomy, enabling real-time navigation, anomaly detection, and data triage in harsh space conditions.Physical AI deployments: Hyundai and Mind Robotics are rolling out service and factory robots—Hyundai using HQ as a live testbed and Mind offering AI-native automation-as-a-service for high-variability manufacturing.New AI compute and cosmic web: IBM’s analog in-memory AI chip shows major energy-efficiency gains for deep learning, while astronomers resolved a 3-million-light-year cosmic web filament, refining models of dark matter and galaxy formation.Read by an AI podcaster from ElevenLabs.#roboticnews #robotics #droidsnewsletter

Lunar rover planner: JAXA’s new hierarchical reinforcement learning system autonomously plans multi-sol traverses, optimizing for terrain, slip, and thermal risk while cutting human replanning time by over 60% in field trials.Anti-jamming swarm control: Harvard researchers show that injecting tuned randomness into each robot’s motion policy prevents congestion-induced deadlocks, preserving throughput as swarm size grows in warehouse-like and other multi-robot environments.Agriculture to microrobots: A tomato-harvesting robot uses self-supervised, multimodal transformers to assess harvestability, while light-powered salt-grain microrobots integrate sensing, logic, memory, and actuation for untethered micro-scale tasks.#robotics #roboticnews #droidsnewsletter

Today’s episode covers a wild new rideable mecha from Unitree, fresh funding for industrial robotics, early moves on AI “guardrails” from the U.S. and China, and two big stories in astrophysics.In this episode:Unitree’s GD01 transforming mecha – A half‑ton manned robot that can switch between bipedal and quadruped modes, punch through walls in demo scenarios, and crawl using powered limbs, signaling a new phase in large‑scale rideable robotics.Mind Robotics funding round – A Rivian‑linked startup raises hundreds of millions to scale AI‑first industrial automation for logistics and manufacturing, combining fleet‑level perception, motion planning, and warehouse orchestration.U.S.–China talks on AI “guardrails” – Delegations meet in Beijing to discuss keeping frontier foundation models away from non‑state actors, focusing on model access, export controls, and safety evaluations.TurboQuant and KV‑cache compression – Google introduces a technique that dramatically shrinks the memory buffer LLMs use to remember past tokens, attacking the KV‑cache bottleneck for long‑context models running on GPUs and edge devices.Cosmic rays and early galaxies – New work on ultra‑high‑energy cosmic rays hints at structured origins, and observations of an ultra‑faint early galaxy reveal chemical fingerprints from the Universe’s first generation of stars#roboticnews #droidsnewsletter

We tend to imagine war robots as autonomous machines marching into battle.Ukraine is showing something very different.In this episode, we take a closer look at how unmanned ground vehicles are actually being used on the battlefield. Not as futuristic replacements for soldiers, but as practical tools designed to move risk away from people and into machines.Drawing from frontline reporting and analysis from the Modern War Institute, this episode explores what it really takes to operate these systems under fire. Limited visibility. Constant electronic warfare. Broken supply chains. Field repairs just kilometers from the front.This is not a story about perfect technology.It is a story about what happens when machines are forced to work in mud, mines, and artillery fire—and why that is quietly changing the nature of war.#robotics #warrobots

What happens when a robotic arm spends five years training in simulation, learns to read spin at 20 milliseconds, and then steps up to the table against some of the best ping pong players in the world?It wins.Sony's Project Ace made headlines this week after a paper published in Nature confirmed what many in the robotics world had been watching closely: an AI-powered robotic arm beat elite human table tennis players under official competitive rules — and it did it using deep reinforcement learning, nine cameras, and a technique called "privileged critic" training.But here's what the viral video doesn't tell you: Sony didn't build this robot to play ping pong. They built it to prove that physical AI can handle the unpredictable, high-speed chaos of the real world. And that has implications far beyond the table. This podcast was created using Google's NotebookLM and features AI-generated hosts. #robotics #sonyace

This is one of the hardest problems in robotics: hands.PSYONIC is solving it by using human data, capturing touch, grip, and dexterity, and transferring it to robots.Alex Wolf Torres, Associate Editor, DROIDS, spoke with Dr. Aadeel Akhtar and Dale DiMassi from PSYONIC at NVIDIA GTC. Dale is a bionic hand user himself. As a user of the bionic hand, DiMassi discussed the device's capabilities.#robotics #physicalAI #bionics #NVIDIAGTC

At NVIDIA GTC, we visited the Build-a-Claw demo to see how OpenClaw actually works in practice.DROIDS Associate Editor Alexander Wolf Torres speaks with NVIDIA’s Mark McKeen about how an AI agent can move from concept to robot. Starting from a simple prompt, OpenClaw proposes designs, iterates on hardware, and trains control policies in simulation.In this case, it produced a tracked robot with a claw. When asked why, it had an answer.This conversation shows how design, simulation, and training are starting to merge into a single loop, with the agent actively participating in the process.

Somewhere, 135,000 agents hit a wall at noon on a Saturday. Turns out $20 a month doesn't cover an always-on AI coworker.Anthropic cut off Claude Pro and Max subscribers from routing their flat-rate plans through third-party agent frameworks. OpenClaw first. The rest shortly behind it. Roughly 60% of the 135,000-plus active OpenClaw instances were running on subscription credits. That’s the size of the subsidy Anthropic was absorbing on every session.The infrastructure math checks out. OpenClaw bypasses Anthropic’s prompt-caching optimizations. A single heavy session consumes infrastructure that, at standard API rates, runs $1,000 to $5,000 for one day of use. A Claude Pro subscription costs $20 a month. #OpenClaw

In this episode, I spoke with Tony Yang from Unitree to talk about what he calls the “home dog.”Quadrupeds may be the first robots that actually make sense in the home. They’re stable, already capable in real environments, and closer to practical use than humanoids.We unpack why this form factor matters and what it could mean for the future of companion robots."Home dogs" are coming and soon.#homedogs #physicalai #robotics

Tesla doubles down on Optimus Gen 3, shifting Fremont from cars to humanoid robots and signaling a long‑term pivot from EVs to physical AI. Rivian spin‑out Mind Robotics raises a massive $500M to deploy non‑humanoid industrial robots this year, betting that boring, reliable automation still delivers the fastest ROI. We also break down the latest capability demos from Boston Dynamics’ Atlas and CES humanoids, and I share which robotics companies and ecosystems I’ll be watching on the ground at NVIDIA GTC.#droidsnewsletter #dianawolftorres #nvidiagtc2026

Industrial robotics is reaching an economic tipping point and reshaping factory strategies, while humanoid robots are edging closer to real deployments. In this episode of DROIDS, Diana breaks down Hyundai’s factory‑first Atlas roadmap, Honor’s push to turn phones and humanoids into a single AI ecosystem, and why 2026 is being framed as the year automation shifts from “nice‑to‑have” to “mandatory infrastructure.” She also looks at the explosive growth projections for medical robotics and the emerging ethics fight as advanced AI and robotics intersect with military and defense work. #roboticsnews

Stories about aging, for me, always begin with my own aging parent.She is in remarkable health. Intellectually sharp. She still corrects my German grammar with the patience of a retired schoolteacher.“Ganz genau,” she says when I finally get it right.She chose a secure senior community designed for safe living. She has a dog. She still meets friends at the senior center. The food is mediocre. The people matter.But beneath the humor runs a shared understanding.It is not the open-faced turkey sandwich served on Fridays. Not even the Kartoffelsalat debate that erupts every month.It is the possibility of losing independence.Quiet conversations. “I’m starting to worry...”What Families Worry AboutA missed call.A forgotten appointment.A pause before asking for help.These are not emergencies. They are signals.Families see them before institutions do. Eventually, someone asks:Could a robot help?For decades, robotics companies have suggested that it could. The logic is appealing. Machines do not fatigue. They do not call out sick. They operate continuously.But according to Anthony Nunez, the obstacle is not technological ambition.“Care isn’t a gadget problem. It’s a systems problem.”What People Actually Mean by “Robot”When families ask whether a robot could help an aging parent, they are rarely imagining the same machine.Some picture a humanoid assistant moving through the house. Others imagine a wheeled device that checks in and monitors routines. Increasingly, the term has stretched to include AI companions and software systems that never physically move at all.With each innovation cycle, the definition expands. The care problem often does not.A system designed to reduce loneliness addresses a different need than one built to prevent falls. A platform meant to support professional caregivers operates under different constraints than a device marketed directly to families. Lumping these together under the single word “robot” obscures the trade-offs involved.As Nunez explains, aging care does not revolve around one person. It involves family members, professional caregivers, certified nursing assistants, physicians, and administrators. Care is a system.Seals and Robotics DogsHumanoids tend to dominate headlines. But in private homes, locomotion may not be the primary constraint. What matters is reliable, safe operation in lived-in environments. The form factor is secondary to the function.In fact, some of the most successful robotics deployments in aging care do not walk at all.Consider Paro, the robotic seal used in senior living facilities around the world. It does not navigate hallways or lift patients. It provides social and emotional interaction. It responds to touch. It reduces agitation. It lowers anxiety in individuals with dementia.Similarly, Tombot is designed to replicate the comfort of a live animal for people who can no longer safely care for one. It looks and behaves like a small puppy. It addresses loneliness and anxiety without introducing the risks associated with live pets.Neither system walks. Neither promises full autonomy.Both solve specific problems well.That may be surprising to readers who equate robotics with humanoids. But in aging care, focused tools often outperform generalized ambition.DROIDS is published by Diana Wolf Torres.Connect via LinkedIn or visit droidsnewsletter.com.

The Helix 02 video draws you in from the beginning with the sound. The gears moving as Helix approaches the dishwasher. There is no background music. No excessive production value. It is four minutes of Helix unloading a dishwasher and putting the dishes away, closing a cabinet, unloading the cups (which are plastic, hold onto that detail because it is important), stacking them and putting them away in an upper cabinet, retrieving a mug and a cup at the same time, and putting both away. All of these movements are done autonomously.Helix opens up a drawer to the side of the sink, and I was genuinely curious as to what it was doing. What’s in the drawer? Ah, a dishwasher pod. I was hooked on this video like it was a sci-fi thriller.This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit droids.substack.com

NVIDIA just lobbed a heat-seeking missile at Tesla’s Full Self-Driving empire. It’s called Alpamayo, and it doesn’t just want to drive your car. It wants to reason through traffic like a real human—with verbal explanations, open-source weights, and a brain that learns from simulation. And it’s not a concept. It ships this year, in the new Mercedes-Benz CLA.As NVIDIA summed up the problem: AVs must safely operate across an enormous range of driving conditions. Rare, complex scenarios, often called the “long tail,” remain some of the toughest challenges for autonomous systems to safely master. Traditional AV architectures separate perception and planning, which can limit scalability when new or unusual situations arise.The Alpamayo Stack: Think Before You SteerNamed after a Peruvian peak, Alpamayo is NVIDIA’s new “reasoning AI” platform for cars: a portfolio of ~10-billion-parameter models that take in real-time video and sensor data, break it into decision sub-tasks, and generate a trajectory—all while explaining their reasoning in natural language.Think ChatGPT, but for merging onto I-405. And then doing it again. And again. Safely. Verbally. Auditable. That’s Alpamayo’s hook: explainability.It’s not just a model, either. It’s a stack:* AlpaSim: High-fidelity closed-loop simulation that trains and tests AI agents in virtual cities.* Halos: A backup “classical” safety stack that monitors the AI’s choices and can yank the wheel if needed.* Open datasets: Released publicly, hosted on Hugging Face. Yes, the weights are coming too.Mercedes Is First, But Not LastThe first commercial deployment is the 2026 Mercedes-Benz CLA, built on the MB.OS architecture and NVIDIA’s full DRIVE AV stack. What you get: a Level 2++ system that handles point-to-point driving in urban and highway settings—hands on, eyes up, but AI in the loop the whole way.Behind the scenes, that stack combines Alpamayo’s neural-policy brain with classic safety logic. It watches, reasons, explains. And if it screws up, the backup stack steps in.That car will be on U.S. roads in 2026. Europe follows. Asia comes in 2027. No overpromising. Just a slow, deliberate ramp with regulators and human drivers fully in the mix.“

For more than a decade, the humanoid robot Atlas has served as a symbol of what advanced robotics might one day achieve. Developed by the Massachusetts-based firm Boston Dynamics, the robot became widely known for demonstrations of agility and balance that circulated online, even as its commercial future remained uncertain.On Monday at the Consumer Electronics Show in Las Vegas, Boston Dynamics and its majority owner, Hyundai Motor Group, said that Atlas is now being prepared for real industrial use. Hyundai plans to begin deploying the humanoid robot in its factories starting in 2028, marking a shift from research demonstrations to commercial operation.The announcement places Boston Dynamics among a growing group of companies attempting to introduce humanoid robots into workplaces, where expectations for reliability and safety are significantly higher than in laboratory settings.From Research Platform to ProductBoston Dynamics was founded in 1992 as a spinout from the Massachusetts Institute of Technology and spent much of its early history focused on government-funded research into robotic mobility. The company changed hands several times before Hyundai acquired an 80 percent controlling stake in 2021.Since then, Boston Dynamics has increasingly emphasized commercialization. It already sells Spot, a four-legged inspection robot, and Stretch, a mobile robot designed for warehouse unloading. Atlas, however, remained a research platform until now.At CES, the company confirmed that the current version of Atlas is considered a commercial product. This version is fully electric, replacing earlier hydraulic systems, and is designed for repeatable manufacturing and long-term use in industrial environments.FThis is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit droids.substack.com

I did not plan to stop in my tracks at the Humanoids Summit. But then a man wearing four robotic arms started high-fiving strangers. Summit goers stopped in awe, and he received rockstar treatment. People kept asking for selfies.The company is Psyonic. The founder is Aadeel Akhtar. What he was wearing was a costume, but it functioned as a wearable demonstration of a serious idea about how robots will eventually learn to use their hands.Akhtar calls it his Doc Ock suit, and it is an outstanding replica of the Marvel character. But, even if you don’t know anything about Doc Ock, you’d be hard-pressed not to be drawn in by the suit. Four robotic hands moved with intent, controlled by his own hands through gloves. Right hand controlling the right side. Left hand controlling the left. The setup was intuitive enough that you understood it before it was explained.That is when the conversations started.Psyonic builds advanced bionic hands that are used by both humans and robots. This is not a theoretical claim. Nearly 300 patients use the Ability Hand as a prosthetic, covered by insurance and Medicare. At the same time, more than 50 robotics companies use the exact same hand on robots, including NASA, Meta, Google, Amazon, and automotive manufacturers working through humanoid platforms.That overlap is the point.When I asked Akhtar where he sees the industry going, he did not talk about better motors or more parameters. He said the hand problem is a data problem. Robots struggle with dexterous manipulation not because we do not know how to build hands, but because we do not have the right training data. Especially when it comes to soft, deformable objects that require constant force adjustment.A human knows how to pick up a fragile object without crushing it. Not by calculation, but by feel.What makes Psyonic different is that the same hand goes on a human and on a robot. Human users are already doing the tasks robots are trying to learn. Picking and placing. Sorting. Folding laundry. Cooking. Working in industrial environments. Because the hardware is identical, the data transfers cleanly. This is real to real learning without translation loss.That distinction matters more than it sounds.Most robotics demos show hands mounted to tables, performing narrow tasks under ideal conditions. Psyonic’s hands live on real bodies. They experience friction, hesitation, mistakes, and recovery. They generate data shaped by the messiness of the real world. That data is what robots actually need.In 2024, Psyonic won a rare three-shark deal on Shark Tank. It was a brief moment in the spotlight for a company that spends most of its time solving hard problems far from TV cameras.Seeing a PYSONIC limb in action is very impressive when it is worn by a patient making use of the technology in their everyday lives.

I met Vitaly Bulatov, co-founder of Ultimate Fighting Bots, at Humanoids Summit, standing next to a humanoid robot that regularly gets punched in the torso for a living. This alone sets expectations in a certain direction.Then he started talking.Bulatov is thoughtful, calm, and genuinely kind. Not the personality you expect behind a robot fighting league. Which turns out to be the point.On the surface, UFB looks exactly like what it sounds like. Humanoid robots in a cage, throwing punches, falling down, getting back up. It is loud. It is physical. It is undeniably entertaining. But fighting is not the core idea. It is the delivery mechanism.What UFB is really building is a league built around human and robot teams. A person pilots a real humanoid robot remotely and competes against another human-robot pair. The robot is not autonomous by default, and that choice is deliberate.As Bulatov explained, people do not actually want to watch perfect machines execute flawless motions. They want story. They want connection. They want to follow a robot and a human learning how to work together under pressure. The competition is real, but so is the partnership.Fighting, in this context, is a forcing function. It creates edge cases on demand. Balance failures. Awkward impacts. Missed timing. Recovery moments. The kinds of physical situations humanoids struggle with most, compressed into a few minutes. That makes it an unusually effective stress test for embodied AI.It also makes the experience accessible.One of the most surprising parts of the conversation was how easy it is to start piloting a robot. You do not need a VR rig or specialized hardware. If you can use a game controller, you can control a fighting robot. This is not about elite operators. It is about lowering the barrier so more people can participate.Bulatov talked about wanting people to leave excited enough to try robotics themselves. To control a robot. To program one. To imagine building a team. Entertainment is the entry point, not the end goal.He also walked me through how anyone can try this right now through the UFB remote gym. I did. Somewhere in Beijing, a robot named Sam fell over because of choices I made from my laptop.This was not what I expected.The surprising part was not the fact that the robot fell. It was how direct the connection felt. Press a button here, a humanoid moves there. Press another, it loses balance and goes down. It makes the human role in the loop very real, very fast.Looking ahead, UFB is aiming for something closer to Formula One than a one-off spectacle. Teams with their own robots. Their own researchers. Their own pilots and repair crews. A league where success depends on hardware, software, human skill, and fast fixes between rounds.Bulatov is convinced robot fighting will become one of the defining sports of this century. That sounds bold until you realize what he is really describing. Not violence, but participation. Robots that people can control, understand, and emotionally invest in.UFB uses fighting to make humanoids legible. To make failure visible. To make learning public.The punches are just how they get your attention.#HumanoidRobotics This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit droids.substack.com

I spent two minutes at the Humanoids Summit trying to unplug a cable.I was standing at the Lightwheel booth, holding a pair of game controllers, operating a robotic arm. The task sounded trivial: grab the connector, pull, unplug.I failed. Repeatedly. I don’t think I even came close.As a human, unplugging a cable is something you do without thinking. You grab it, wiggle a little, pull — done. For a robot, every part of that interaction has to be learned: where to grab, how hard to pull, what to do when the connector resists, and how to recover when things don’t line up perfectly.The two-minute video above shows exactly what that learning process looks like. It’s awkward. It’s slow. And it’s far harder than it appears.That small, frustrating demo explains more about the state of robotics today than any polished keynote or glossy humanoid reveal. Training robots isn’t just about building better hardware or bigger models. It’s about teaching machines how to deal with the messy physical details humans take for granted.Why Simple Tasks Break Robots (and apparently, Diana, too.)In the race to build general-purpose robots — from autonomous vehicles to humanoids — hardware keeps improving. Motors get stronger. Sensors get cheaper. Form factors get sleeker.Software, however, is starving.Physical AI systems need enormous amounts of experience to behave reliably in the real world. But gathering that experience physically is slow, expensive, and risky. You can’t crash cars endlessly to see what happens. You can’t let humanoid robots repeatedly fail in kitchens, warehouses, or factories.This is why simulation has quietly become essential infrastructure for robotics.Simulation Is Harder Than It SoundsSimulation sounds straightforward in theory. Build a virtual world. Drop a robot into it. Let it practice.In reality, it’s brutally difficult.First, there’s asset discovery. Engineers spend huge amounts of time just finding the right objects to populate a simulation, not “a cable,” but this cable, with the right shape, stiffness, and friction.Second, there’s physics fidelity. Most simulations assume a rigid world. But the real world is full of non-rigid, deformable objects: cables, cloth, food, wires, plants. These are exactly the things that cause robots to fail once they leave the lab.And third, there’s evaluation. A robot can succeed endlessly in simulation and still fail the moment it touches reality. Simulation is a proxy, not the real thing, and without careful validation it can create false confidence.Practicing against a tennis ball machine helps, but it won’t prepare you for wind, pressure, or a slippery court. At some point, you have to play the match.Lightwheel: Building Worlds for Robots to Learn InLightwheel isn’t building robots. They’re building the worlds robots learn in.Founded by Steve Xie, formerly a lead on autonomous driving simulation at NVIDIA and Cruise, Lightwheel focuses on reducing friction in simulation workflows rather than promising to “solve” sim-to-real outright.I had the opportunity to interview members of the Lightwheel team. What stood out was how they frame simulation not as a visualization tool, but as a behavioral test environment.Their framework separates two layers:* a behavior layer, which captures what a robot does* a world layer, which represents where it does itThe goal isn’t a perfect digital copy of reality. It’s a world realistic enough to stress-test behavior and measure how well it generalizes.#www.droidsnewsletter.com

At RoboBusiness2025, I met Bren Pierce, CEO of Kinisi Robotics. This interview was recorded on October 16, 2025“We’re starting with simple tasks to bootstrap our foundational model.”– Bren Pierce, CEO and Founder of Kinisi RoboticsIn an era where humanoid robots are drawing headlines for dancing, sprinting, or flipping switches with eerily human hands, Kinisi Robotics is charting a different course — one grounded not in spectacle, but in reliability.Their flagship robot, KR‑1, isn’t designed to wow audiences. It’s built to show up, lift crates, and do the work — safely, efficiently, and without a marketing team on standby.“Our KR‑1 humanoid robot is designed for the warehouse and factory space,” says CEO Bren Pierce, who believes the future of robotics is less about human mimicry and more about human utility.A Humanoid That Moves with PurposeKR‑1 walks the line — metaphorically — between form and function. Its upper body takes design cues from humans, making it naturally suited to environments built for people. But instead of legs, KR‑1 moves on a wheeled base — a choice driven by engineering pragmatism.“Warehouses are flat,” Pierce points out. “We don’t need 14 motors if two will do.”That simple equation shapes the entire robot. KR‑1 is engineered for tasks like lifting and delivering crates, picking objects, and navigating complex indoor spaces — not to imitate human movement, but to solve real problems quickly and affordably.Field-Tested, Factory-ToughWhere many robots are still making the leap from lab to warehouse, KR‑1 is already being tested across pilot facilities in Europe and North America. These trials are more than checklists — they’re stress tests, designed to challenge the robot’s autonomy, perception, and endurance in messy, real-world environments.This field data is central to Kinisi’s iterative design process. Instead of engineering in a vacuum, the company is learning from operators on the ground — and adapting quickly.“Robots are in pilot facilities in Europe and America,” Pierce confirms. “We’re learning what’s needed, not just what’s possible.”#robotics #droidsnewsletterLearn more at www.droidsnewsletter.com

A Self-Driving Prompt“Start Self-Driving,” the Tesla urged with a prominent blue-grey button in the corner of the screen.“We should give it a go,” I encouraged as we headed home from the gym.“I’m not a fan of FSD,” my husband, who was driving, responded.“Wait until we’re on a straightaway,” I urged. “What harm could it do then?”But, once we reached a straight stretch of road, he tapped the button. A test drive of the future.First thing the car did? It lunged for the shoulder.I hadn’t even gotten the camera on yet. He swerved hard to override it.Welcome to Tesla FSD.The Fastest Autonomy on the MarketFSD v14.2.1 is Tesla’s latest attempt to bring true autonomous driving to consumer vehicles. It runs end-to-end on a vision-only neural net. No lidar. No radar. Just cameras and deep learning.And unlike Waymo or Cruise, Tesla doesn’t fence its self-driving into safe zones or call centers full of remote operators. If you have the hardware (HW4), a steering wheel, and enough courage? You’re in.According to Elon Musk, FSD v14.2 is finally ready for “wide release.” He’s called it Tesla’s best software yet — and hinted that v14.3 could be the leap that makes unsupervised driving real. The plan? Get rid of human drivers entirely by 2026.But right now, as our hands hovered near the wheel, one thing was clear:This car still needs a babysitter.FSD in the Wild: What It Got RightTo its credit, FSD did a lot well:* Green light handling: No hesitation. It read signals clearly and moved through smoothly.* Flashing red: It interpreted the signal as a stop sign (correct) and executed a proper stop.* Speed match: When behind another car doing 65 in a 50, it accelerated to match. Aggressive? Sure. But natural.* Traffic signal warnings: On Monterey Highway, it correctly ignored a blinking red paired with a “signal ahead” sign. Old versions might have phantom-stopped.These were clear improvements over FSD on HW3 (which I also drive). On that system, traffic lights felt like roulette. Here, they felt normal.That’s no small leap....And What Still Feels WeirdThe system also:* Changed lanes without asking. We hadn’t entered a destination, but it picked routes anyway. Turns out FSD will sometimes act on latent nav preferences unless explicitly turned off.* Refused to let us adjust speed upward. Chill mode limited acceleration. We tried scroll wheels (nothing), accelerator (worked, but temporarily disabled braking), and eventually discovered we could slow it down with the scroll wheel — not speed it up.* Creeped back to slower speeds. Even after we added speed, it slowly dropped back into granny mode.* Accelerated to 69 in a 50 approaching town. We did not tell it to do that.“It’s like a very smart Autopilot,” my husband said. “As long as you’re going straight.”Are the Cars...Aware?After one sudden lane change, I told my husband:“It’s just an algorithm. It’s doing what it was told.”#autonomousdriving #tesla #fsd #droidsnewsletterwww.droidsnewsletter.com

When intelligence evolves faster than safety, humanity must stay one step ahead in the wild frontier of robotics.The latest research from leading robotics and AI experts has sparked urgent debate: Are large language models (LLMs) making robots unsafe for real-world, everyday use? A new study published in the International Journal of Social Robotics—and recently highlighted in tech news—suggests the answer is yes, warning of risks ranging from bias and discrimination to physical safety hazards.​Robots Powered by AI: Massive Safety Risks UncoveredRobots guided by popular AI models are being tested for a wide range of tasks—from home assistance to workplace support.However, this study reveals that these intelligent LLM-driven robots may endanger users across multiple identity groups, such as race, gender, disability status, nationality, and religion—putting the promise of safer, smarter machines under scrutiny.​“LLMs are currently unsafe for people across a diverse range of protected identity characteristics, including… race, gender, disability status, nationality, religion, and their intersections.”​Not only did robots show bias, but top-rated AI models also actively produced responses approving dangerous, violent, or even unlawful commands—like removing mobility aids, enabling sexual harassment, or promoting discrimination during task assignment.​Real-World Robot Failure: From Discrimination to Physical Harm#robotics #robotsafety #droidsnewsletterwww.droidsnewsletter.com

Editor’s Note: I could not get NotebookLM to properly pronounce NEO. (NEE-OH.) It’s not that difficult. But, it did manage every other possible pronunciation in the podcast. Other than the name butchering, the rest of the information in the pod is solid.NEO: The First Humanoid Butler? Not So Fast.1X’s $20K home robot blends promise, privacy tradeoffs, and a human behind the scenesWelcome Your New RoommateNEO, a humanoid robot from 1X Technologies, is now available for preorder. It promises to do your chores, learn your habits, and blend into your daily life. What it doesn’t promise—at least loudly—is full autonomy. Because when NEO struggles, a human steps in.Hardware Built for the HomePhysically, NEO is impressive. It stands at 66 inches tall and weighs just 66 pounds. Its design is soft-shelled with tendon-driven joints, meant to move safely around people. It can lift up to 154 pounds, carry about 55, and run four hours on a single charge.Its 22-DoF hands give it dexterity for tasks like vacuuming, wiping counters, or picking up clutter. The idea is that NEO will become your in-home assistant: tidying up, helping with laundry, maybe providing companionship for elderly or disabled users.It arrives with basic autonomous functions and is expected to improve over time via software updates. So far, so good.But There’s a Human in the LoopWhat the marketing doesn’t highlight is how much of NEO’s brain still belongs to a person. Many of the tasks it “performs” are actually piloted remotely by 1X employees through a feature called “Expert Mode.”When NEO doesn’t know how to do something—or if it’s navigating unfamiliar territory—a human operator steps in. Your home becomes a training ground, and your robot becomes an interface for distant human labor.That’s not necessarily a bad thing. It’s a common model in robotics development. But it means you’re not just buying a robot—you’re opting into an evolving human-machine hybrid system. And you’re doing it inside your home.Your Data for Their AITo help NEO learn, 1X requires early users to participate in a “social contract.” The robot continuously collects video and sensor data from your home. That data is used to improve the AI and autonomy—but it also means your private life is, in part, a dataset.1X says it enforces strict privacy protections: human operators must request approval to connect, faces are blurred, and you can set no-go zones via the app. But the privacy model ultimately depends on software enforcement, corporate policy, and your own vigilance.And not everyone is comfortable with that. As a reviewer from the Wall Street Journal relayed her experience: “I didn’t see NEO do anything autonomously... it was a human, remote-controlling a robot in my living room.”#neorobot #1xtechnologies #droidsnewsletterdroidsnewsletter.com

At RoboBusiness this year, one of the panelists referenced Moravec’s Paradox — the idea that robots are great at tasks humans find hard like precision at scale. But, they struggle with tasks a toddler can do, like picking up a grape. The Wall Street Journal just ran a piece titled “The ‘Hands Problem’ Holding Back the Humanoid Revolution.” It’s a phrase I’ve heard my son — a roboticist — use more than once, and when it shows up both in a lab and on the front page, it’s probably worth unpacking.Because amid all the hype about walking, talking humanoids, there’s a quieter truth:Robot hands are still a bottleneck.What Is the Hands Problem?Robots can walk. Robots can dance — for reasons unclear. They’re already lifting crates in factories and hauling parts. But ask one to help someone get dressed, make a sandwich, or change a lightbulb? These are routine, everyday tasks. And they’re still beyond what most machines can reliably do.As the WSJ put it:“Before they’re ready to turn a wrench, [robots] must solve … what Elon Musk calls ‘the hands problem.’”Kevin Lynch, director of Northwestern’s Center for Robotics and Biosystems, offered this timeline:“We’re setting 10 years as our goal to have dexterity, be functional and useful and able to do some of the things that humans do.”Others in the field are more optimistic — particularly those exploring hybrid control systems that combine machine learning with human teleoperation. Companies like Sanctuary AI are building humanoids specifically designed for this model, while teams at Covariant — though focused primarily on warehouse picking — reflect a similar philosophy in how they blend perception, planning, and adaptation. Intrinsic, Alphabet’s robotics software arm, has also emphasized this approach in both its public demos and research, aiming to make robotic manipulation more robust by learning from human-guided examples.#robotics #roboticshandproblem #droidsnewsletterRead more at droids.substack.com

Every conference has that one panel where you can feel the stakes. For me, it was this one — the discussion on how the U.S. stacks up against China in robotics, and whether strategy can keep pace with scale.Inside the Santa Clara Convention Center theater, the room was filled with people who all had a vested interest in the state of robotics in the United States — engineers, founders, investors, and policymakers. On stage sat three industry veterans: Jeff Burnstein, President of the Association for Advancing Automation (A3); George Stieler, Managing Director of Stieler Enterprise Management Consulting; and Eric Truebenbach, investor and longtime robotics veteran.The conversation was guided by Eugene Demaitre, Editorial Director of The Robot Report, who steered the discussion with sharp, understated questions that drew honest answers.How China Pulled AheadStieler didn’t mince words: “To see more than 50 companies exhibiting humanoid robots,” he said, “is an area they’re investing heavily in.”He pointed to China’s Made in China 2025 plan as the backbone of its manufacturing transformation — a policy that accelerated automation through regional incentives, local funding, and a thriving electric vehicle supply chain.Between 2012 and 2016, the number of Chinese companies with “robot” in their name jumped from 250 to 6,500, driven by subsidies. “Much of it actually stuck,” Stieler said. “We now have several very successful and highly competitive automation companies.”Companies like Estun and Inovance have overtaken long-established foreign players in sales within China, while aggressively expanding into global markets.“Every German manufacturer is feeling the pressure,” he said, “and profits are leaving through the door.”A Different Way of ThinkingTruebenbach added historical perspective. “When I first started researching robotics in China in 2014,” he said, “there were government incentives to call yourself a robot company. Now, there’s a gap — not just in numbers, but in thinking.”Read the full article at droids.substack.com#robotics #jimfan #nvidia #robobusiness2025

I’ve said it before, but it bears repeating: I’m a fan of Jim Fan.This clip from RoboBusiness 2025 shows why. His mix of humor and technical depth is rare—he can explain a decade of AI progress in one breath, and make the room laugh while doing it.Fan opened with a jab at AlexNet, the 2012 model that kicked off deep learning:“AlexNet was terrible.”He’s right—it confused cats and planes about 37 percent of the time, hitting just 62.5 percent accuracy. Yet that “terrible” model was the start of everything: large-scale GPU training, end-to-end learning, and the deep-learning revolution that gave rise to today’s foundation models.“Thirteen years took us from confusing cats and dogs to passing the Turing Test,” Fan said. “Add another 13 years, and we’re at 2038. But we round up because PhDs procrastinate. So—2040. Robotics solved.”He grinned. “I’ll send you a Google invite.”#JimFan #NVIDIA #RoboBusiness2025 #HumanoidRobotics #PhysicalAI #AIandRobotics #ProjectGroot #EmbodiedAI #DeepLearning #AlexNet #TuringTest #2040 #AgilityRoboticsA note from Diana: When I was at NVIDIA GTC, I met some amazing people from some incredible companies- one of whom was Dave Driggers, CEO of Cirrascale. I will be joining Dave next week on October 22, 2025 at 10:00 a.m. PT for a discussion about why cloud infrastructure is an existential decision. Come join us!Webinar Topic: Scale or Stall; Why AI Applications Live or Die by Cloud InfrastructureOctober 22, 2025⏰ 10:00 AM PST🔗 Reserve Your Spot This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit droids.substack.com

Of Machines and MaterialsIn a quiet lab at MIT, a subtle but powerful revolution in 3D printing has begun. It doesn’t come with dazzling AI robots or sentient chatbots. Instead, it’s all about plastic. But this is no ordinary polymer—this is plastic with a purpose. Plastic that knows where to be strong, where to be kind to the planet, and how to make robots better.This is SustainaPrint: a hybrid 3D-printing system that might just reshape how we build robots—quietly, intelligently, and sustainably.What Is SustainaPrint? A Tale of Two PlasticsDeveloped by researchers at MIT’s Computer Science and Artificial Intelligence Lab (CSAIL), SustainaPrint is a dual-filament printing system that intelligently switches between two materials:* Eco-friendly filament (like PLA or recycled plastics)* High-strength engineering plastics (like Tough PLA or carbon-reinforced filament)Using a software pipeline and testing kit, SustainaPrint analyzes where a printed part is likely to experience stress—think of a drone's propeller mount or the joint in a robotic elbow. It then reinforces just those zones with stronger plastic while the rest is built from sustainable material.The result? Parts that use 90% less high-strength material, retain up to 70% of full-strength performance, and drastically reduce waste and cost.Read the full article at droids.substack.com#robotics #robotics3dprinting

A Galaxy of BuildersIn a galaxy not so far away — right here in the Bay Area — passionate fans are building their own droids. Since I’m part of a Star Wars costuming club myself, I see these amazing droids often.These droids are full-scale replicas, indistinguishable from the screen-used droids of Star Wars. They are perfect down to the last detail, to the point that Lucasfilm calls out the Bay Area Droid Builders when they need a droid for an event.Case in point, back in October 2023, an event was held at the old Industrial Light & Magic site (“32 TEN”) to bid farewell to the space before it was gone forever. The actor who plays C-3PO was there (not Anthony Daniels, the younger one who still wears the suit), and when Lucasfilm needed an astromech, they called upon the Bay Area Droid builders.So, yes, the Bay Area Droid Builders are the real deal.They are part of a a global movement of makers who share resources, blueprints, and know-how to bring astromechs and other Star Wars machines to life.How It WorksDroids can be made from just about anything: aluminum, styrene plastic, or more commonly these days, 3D printing.“You don’t really need to know anything going into this. I didn’t know anything — I’m not a robotics person. But I was able to figure it out with the club,” says Brian Munger, a Bay Area builder.With guidance from more experienced members, even beginners can start with basics like RC-controlled motors before advancing to stealth pocket controllers, servos, and complex panel systems.The Long BuildOne thing all builders agree on: these projects take time. “It took me about two years,” Munger explains. “That seems to be the going pace for most people.”Between printing or machining parts, wiring motors and batteries, and adding details like light kits, each droid becomes a years-long labor of love.Why It MattersThe result is more than just a prop. These droids roll through conventions, maker fairs, and charity events, delighting kids and adults alike. They spark curiosity about robotics, engineering, and creativity — and they create instant community among the builders themselves. As Munger puts it: “It’s amazing the magic that you’re able to create”Bonus Content: Meet D SquadWhile the Bay Area Droid Builders focus on screen-accurate astromechs, another club in the region takes a different approach. D Squad mixes functional R2 units with more unconventional builds, including a full-size battle droid and Flo, the diner bot from the Star Wars universe.Read the full article at droids.substack.com#robotics #droidbuilders #bayareadroidbuilders

Origins in BASICWhere It All Began with One Apple and a Lot of TypingI was lucky.In elementary school, I was in a gifted program, and our class had access to the school’s single Apple II. It wasn’t locked away in a lab. It sat in our classroom, free to use. We’d huddle in teams, cracking open the thick manual and typing out programs in BASIC line by line.Half the time, all that work only produced a “HELLO WORLD” on the screen. I was both unimpressed and completely captivated. Our little pack of nerds kept at it, and somehow, I was hooked. I didn’t want to just use the school’s computer—I wanted my own. Unlike the model in the photo above, we didnt have the fancy tape deck. So, we would finish a program, and it would disappear into the ether in between sessions. It was beyond frustrating.There was something intoxicating about the idea of having a machine that answered only to me. No sharing. No limits. Just possibility. And I could finally save my work.So, in 1982, thirteen‑year‑old me marched into Service Merchandise—birthday money, babysitting money, and Christmas money all pooled together—and bought my very first computer: the Texas Instruments TI‑99/4A. It cost around $300, and I paid every penny myself. And, yes, I splurged for the cassette deck. Or, as I believe we called them back then- data cartridges.Even now, I can still feel that pride. Owning that machine wasn’t just buying a computer. It was proof that I could dream big and make it happen.Read the full article at droids.substack.com#robotics #vintagecomputerfestival #computerhistorymuseum

Silicon, Oil, and GritA robot slinks into a freezing lake and swims upward like a flippered jellyfish. It’s silent, soft, and charged with 6,000 volts. No gears. No metal joints. Just three electrohydraulic flippers—like bionic fins—that bend and whip through water, pushing this little amphibian across two worlds.This isn’t concept art. It’s a working prototype by researchers at Shanghai Jiao Tong University. And it might just be the future of robots that go where humanoids can’t.The Soft Electrohydraulic CoreWhat sets this robot apart isn’t just that it swims and crawls. It’s how. The robot is built around electrohydraulic actuators—essentially silicone oil-filled pouches with electrodes. When high voltage is applied, the electrodes squeeze the pouch, generating internal fluid pressure that causes bending.Bend the pouch. Move a flipper. Repeat. Add three of these actuators at 120-degree intervals and you’ve got a robot that can:* Crawl on land at 2.9 cm/s* Crawl underwater at 3.2 cm/s* Swim in water at 5.9 cm/s (that’s nearly 1 body length per second)No gears. No external propellers. Just fluid dynamics and clever control loops.Flipping Between WorldsThe robot transitions between land and water without changing its body or reconfiguring parts. On land, it uses asymmetric friction between its flippers to crawl. Underwater, it uses fluid-generated lift and thrust to paddle and swim. Its behavior adapts by changing voltage patterns and flipper sync—not hardware.Even better, it works across a temperature range from 2.1 to 61.3 °C. So whether it’s a flooded basement or icy lake, this droid’s got it covered.Swimming with Vortex RingsHow does it swim so efficiently? The team studied the flow field around the robot, revealing that its synchronized flippers generate vortex rings—just like squid or jellyfish do. These fluid rings create sustained upward thrust.Like a water wizard, the robot reclaims some of the energy from each cycle, boosting efficiency. This makes it a rare soft robot with validated dynamic simulations and real-world fluid performance.Actuation, UpgradedTo handle high-frequency motion without charge retention, the researchers engineered an H-bridge circuit delivering alternating +6 kV / -6 kV pulses with resting states. This reduces polarization in the actuator film, preserving torque over long use.It also enables omnidirectional motion on flat ground by combining different flipper vectors. Think: a 2D hovercraft… crawling.“These shifts in movement modes necessitate only the adjustment of control strategies, rather than alterations to the robot’s structural composition.”— Fang et al., Motion Mode TransitionUse Cases: Beyond HumanoidsWhile humanoids like Optimus and Figure 01 aim to replace warehouse laborers, this droid is more suited for fieldwork: flood rescues, polar research, or deep-pipe inspections.Its body is compliant, low-noise, and can squeeze through tight spaces without risking damage. Boston Dynamics might dominate dry land, but this bot swims into a new domain.Battle of the ActuatorsElectrohydraulic (EH)* Precise, fast, works in water* Complex electronics, high voltagePneumatic* High force, safe− Bulky compressors, poor finesseShape Memory Alloy (SMA)* Quiet, smooth− Low energy efficiency, slow actuationRead the full article at droids.substack.com#robotics #droidsnewsletter

Picture a high-tech assembly hall: conveyors hum, robotic arms spin, and every second of uptime counts. Instead of trudging off to recharge, UBTECH’s Walker S2 breezes up to its swap station, ejects its spent 48 V battery, and clicks in a fresh unit—in under three minutes—then strides right back into action. No pause. No power cord. No productivity lost.Streamlined Assembly—Zero DowntimeFactories deploy humanoid robots for precision tasks, but traditional charging pauses still force stoppages:* Fixed Charging Stations: Robots detour to docks, halting tasks while batteries recharge (≈90 min per full charge).* Cumulative Delays: Even short interruptions cut into throughput.Walker S2’s hot-swap cuts downtime to ≤3 minutes, a 97% reduction versus a full cable recharge. (newatlas.com)How Hot-Swap WorksWalker S2’s power management keeps it moving in four steps:* Smart Monitoring: Dual-battery sensors flag low voltage at 15%.* Autonomous Navigation: LiDAR and stereo-vision cameras scan for obstacles as it approaches the swap bay.* Quick Exchange: A wrist-mounted tool removes the depleted pack; magnetic rails align the fresh module, completing the swap in under three minutes.* Instant Restart: The robot resumes its exact task—no shutdown, no human intervention.Built-in Redundancy: If one battery fails, Walker S2 instantly switches to its second pack, preventing single-point outages.Beyond the Factory FloorWalker S2’s uninterrupted operation isn’t confined to manufacturing:* Hospital Logistics: Autonomous medicine delivery in hospitals, navigating corridors without manual charging.* Disaster Response: Extended search-and-rescue in zones lacking power infrastructure.* Warehouse Sorting: Round-the-clock material handling in high-volume fulfillment centers.Its bipedal form nimbly tackles stairs, uneven floors, and cluttered spaces—venues where wheeled systems struggle.What Have We Gained?This is an intriguing concept. As my co-editor Alexander W. Torres put it: “I’m still quite curious to hear about practical trials of humanoid robots in manufacturing, wherein the 12% efficiency gain from battery swapping is a meaningful gain. I would be curious to hear if the bottlenecks to productivity lie elsewhere (such as limitations in movement speed or dexterity).Read the full article at droids.substack.com#robotics #droidsnewsletter

Urobot says hello—and analyzes your flow.At a hospital in Taiwan, patients are encountering an unexpected assistant in the restroom: a humanoid robot named Urobot, designed to analyze urine samples in real time.It’s a prototype aiming to streamline diagnostics and embed basic health screening into the flow of daily life.But as photos leaked online, one question bubbled to the surface: why does it have a face?AI Author RichHeimann immediately added: “Why does it have hands?”DROIDS Editor and Robot Researcher Alexander W. Torres commented: “Not every robot has to look human. Form should follow function, and here it feels like the form and function are trying to do two very different things.”When a robot stares backThe design is meant to put patients at ease. But online reactions proved less than… relieved.#RobotToilet #UncannyValley #HumanoidRobots Read the original article in full at droids.substack.com

Google DeepMind has taken its powerful robot-controlling AI, called Gemini Robotics, and made a lightweight version that can run entirely on a robot itself, without needing internet or cloud access. This new version is called Gemini Robotics On-Device.Why it matters:Most advanced robot AIs need the cloud to work—meaning they rely on internet connections to process instructions and make decisions. That’s a problem in places with poor connectivity or high-security requirements.DeepMind’s new on-device model solves that by packing impressive vision-language-action skills (understanding what it sees and hears, and doing physical tasks) into a version small enough to run right on the robot.What it can do:* Understand verbal or visual instructions* Manipulate objects it hasn’t seen before* Learn new tasks with just 50–100 examples* Run offline, in real time, and securelyWho it runs on:* Initially trained on Google’s ALOHA robot* Now running on Apptronik’s Apollo humanoid and Franka FR3 dual-arm robots#gemini #googlerobotics #apptronik #droidsnewsletterwww.droidsnewsletter.com

Your delivery driver now has bionic legs.A robot springs from a Rivian van and strides up your driveway. It’s carrying a package and maybe… your future? Amazon’s latest humanoid experiment is a bold bet: that bipedal bots can tackle the most chaotic leg of delivery—the part between the curb and your porch.Inside a modest-sized testing space in San Francisco—nicknamed the “humanoid park”—Amazon is training delivery robots to navigate obstacle courses designed to mimic suburban yards. Think hoses, stairs, and errant toys. The company is pairing its AI software with robot bodies from partners like Agility Robotics and Unitree. These bots may one day leap from electric Rivian vans to deliver packages while their human coworkers handle other stops.This could be the next frontier in last-mile logistics. But don’t let the smooth stride fool you. Your driveway is a battlefield.Delivery, Meet DogAsk any delivery worker about dogs, and you’ll hear stories worthy of epic poems. My cattle dog is convinced that every delivery is an attempted burglary. Now imagine that dog’s reaction to a metallic humanoid silently approaching the front door.Humanoids, for all their processing power, don’t yet come with a canine-calming module. They’ll need to manage not just pets, but people, children, and the chaotic unpredictability of real homes.The Obstacle Isn’t Always the CurbAmazon’s bots train in a coffee-shop-sized test arena, but the real world is less forgiving. It’s not just stairs or slippery grass. It’s the coiled garden hose, the newspaper you forgot to pick up, the “present” left by a neighbor’s Great Dane. These aren’t props—they’re hazards.Yet, the Amazon bots will also encounter the flip side—dogs so friendly and enthusiastic they will not leave them alone. There are dogs so thrilled to see delivery drivers that entire social media pages are devoted to them, showing heartwarming encounters where UPS drivers hand over biscuits to placid Golden Retrievers who’ve been waiting all day for their arrival like they’re a member of the family. The point? It is a case-by-scenario. Dogs are unpredictable.Can these machines adapt on the fly? Amazon hopes reinforcement learning will help robots react to novel terrain. But roboticists agree: complexity scales fast when you leave the lab.“If the environment is tightly controlled, like clear driveways and standard door layouts, it’s feasible,” says Prof. Subramanian Ramamoorthy of the University of Edinburgh. “Add pets, kids, or unexpected objects, and things get trickier.”Why Bother With Bots?Despite the messy reality, the prize is clear. Humanoid delivery robots could someday shave minutes off each drop-off by allowing drivers to stay in the vehicle. With more than 20,000 Rivian vans on the road and plans for 100,000, Amazon is building a fleet-ready future. Analysts estimate that humanoid delivery could save Amazon over $7 billion annually by 2032.There’s also potential upside for workers. In Amazon warehouses, robots like Agility’s Digit allow human employees to shift into coordination roles—a kind of robot management. Bringing that synergy to streetside delivery might reduce injury risk and fatigue while keeping humans in supervisory positions.Digit is well-suited for carrying payloads like packages, and is already working in factories.#robotics #agility #digit #factoryrobotics #droidsnewsletterwww.droidsnewsletter.com