God Mode Podcast

This was the week Cursor beat Anthropic at running Anthropic's own model.The third-party harness extracts more performance from Opus 4.7 than Claude Code does. Codex ate Claude Code with HeyGen and browser baked in. Grok 4.3 dropped at one-quarter OpenAI's pricing. GitHub Copilot moved to usage-based billing. Anthropic shipped Claude Security and Cursor cloned it in two hours flat. Stripe Treasury launched with Stripe MCP — the financial wallet for AI agents. SpaceX board approved Elon's $7.5T pay package with a 100 terawatt compute clause and a 1-million-person Mars colony clause. And the Department of War named its seven AI partners on Friday morning — Anthropic was notably excluded.Three builders. Weekly AI news. No fluff.The big thesis: the abstraction layer is eating the lab. Crypto already ran this movie — Phantom prints, Solana doesn't. AI is the remake.🎙️ Hosts:Rik — x.com/rikventureBen Broch — x.com/brochchainLuca Arrigo — x.com/LucaArrigo_Watch on YouTube → https://youtube.com/@godmodepodSubstack → https://godmodepod.substack.com/

This was the wildest seven days of AI news in 2026 — and we covered all of it.Anthropic finally admitted what builders had been saying for six weeks: Claude got dumber. They confirmed three product-layer bugs stacked on top of each other from March 4th to April 16th. Default reasoning silently dropped from "high" to "medium." A caching bug killed idle session thinking. A system prompt told Claude to keep responses under 100 words, truncating reasoning chains across the entire product. None of it was a model change. All of it was scaffolding.Then GPT 5.5 dropped — better benchmarks than Opus 4.7, faster, co-designed with NVIDIA on 100,000 Blackwell GPUs. But the number nobody put on the slide: 86% hallucination rate. Up from 52% on Opus 4.7. Way up from Grok at 17%. The benchmark race is the wrong race.Plus: SpaceX got a $60B call option to buy Cursor — the world's largest call option, structured as a talent grab to replace Elon's burned XAI founding team. Naval launched USVC, opening late-stage AI venture exposure to retail at $500. Ben breaks down the 30:1 valuation pattern showing up across the entire AI sector — Scale AI, Cursor, Anthropic, every single one priced at the same multiple. And a guy made $34,000 betting on Paris weather by driving to the weather station with a hairdryer.Three builders. Weekly AI news. No fluff.⏱️ CHAPTERS00:00 Cold open — gateway drug + 86% hallucination + compute thesis02:30 Anthropic apologizes — three stacked Claude bugs explained11:30 Spoon-feeding theory + Claude as gateway drug19:00 90% of my week in Claude desktop — Anthropic's everything app20:45 Provider or software? Anthropic must pick23:00 No world where compute goes unused — Ben's bubble counter26:42 Coding is a circular reference — the AGI flywheel29:00 Domino theory of the AI bubble32:11 SpaceX $60B Cursor deal — talent grab, not software32:28 The 30:1 rule of AI valuations41:53 GPT 5.5 hallucinates 86%45:24 OpenAI exec exodus — $500M unvested equity52:31 Is Claude becoming a verb?58:00 Naval drops USVC — retail venture for $5001:08:24 Xpeng flying cars 7,000 pre-orders1:11:00 Polymarket gets hacked with a hairdryer1:11:13 Luca's 2040 mountain hut vibe-code future1:15:40 Polymarket isn't trustless🎙️ HOSTSRik — x.com/rikventureBen Broch — x.com/brochchainLuca Arrigo — x.com/LucaArrigo_Listen on Spotify → https://open.spotify.com/show/0ZBhl5IVFa8wZPvIKNTkRPApple Podcasts → https://podcasts.apple.com/us/podcast/god-mode-podcast/id1883959580Substack → https://godmodepod.substack.com/

Anthropic pre-IPO hit $900B. Claude Code crossed $2.5B ARR in nine months — the fastest B2B growth ever recorded. And your Claude Opus didn't get dumber — Anthropic quietly nerfed it (73% drop in reasoning depth across 7,000 sessions).This week on God Mode Pod: Rik, Ben, and Luca break down the three-week Anthropic news dump — the Tesla vs NVIDIA playbook debate, the Opus reasoning rollback, the desktop Claude Code redesign, and whether IDEs will actually be dead by end of year.Plus: Tesla AI5 tapes out (2,500 TOPS, 17x jump from 2019). The OpenAI CRO memo leak. Luca's antitrust prediction. A 24-year-old ex-OpenAI researcher who turned a $254M fund into $5.5B in one year. Allbirds pivoting to AI compute for a 910% single-session run. And the best builder story of the week — a pool postcard AI agent that's already making money.Three builders. One hour. Everything that mattered in AI this week.New episode every Saturday.

Anthropic has a secret weapon. It's called Mythos — and it just scored 78 on a coding benchmark where GPT-5.4 scored 58. That's not a generational jump. That's a different architecture entirely.In EP 08, Rik, Ben, and Luca unpack the leak of the year, Jensen Huang's $500k token rule, and why Google might actually be the company quietly winning the AI race — while Anthropic's compute constraints start to bite.In this episode:• Mythos is real — and it found a 27-year-old bug in OpenBSD• Anthropic killed OAuth overnight — OpenClaw users got an 8pm eviction• Opus advises, Sonnet executes — the cost-efficient way to use Claude• Jensen's $500k token rule + Meta leaderboards• Gemma 4: 20% → 89% on math in one release• Luca's big question: "Cure diseases with compute, or make memes?"• Why being the leader in AI users might actually hurt AnthropicPlus: the Malta founder retreat. 6 spots. Builders only.🏝 Apply → https://basecampgozo.com/📬 Newsletter → https://godmodepod.substack.com/🎥 YouTube → https://www.youtube.com/@godmodepodGod Mode Pod is a weekly show about what's actually happening in AI — hosted by three builders who use these tools every day, not people watching from the sidelines.

Gonçalo Hall built the first digital nomad village in the world — 20 people in February 2021, 25,000 by end of year. Now he's doing it again, on a Caribbean island almost nobody has heard of.Roatan is 2 hours from Miami. It's home to Prospera — a free economic zone where gene therapy is legal, companies register in minutes, and you can get physical tax residency for $5,000/year with just 7 days on-island required. Brian Johnson has been there. The Kardashians went. A company just won a $1M Thiel Award for stem cell research there.In this episode we go deep on:→ How Prospera's ZEDA law actually works (and what makes it unlike anything else on Earth)→ The $11 billion lawsuit Prospera had to file against the Honduran government — and how they won→ Why 5,000 free economic zones exist but only 3 are actually successful→ Gonçalo's full community-building playbook (20 → 25,000 in 3 months)→ The identity crisis that made him call himself a "Destination Architect"→ Self-knowledge, ADHD as a superpower, and why you should work like a lion — not a cowGonçalo is the first official guest on the GodMode Pod. This one goes deep.🎙️ Guest: Gonçalo Hall — Destination Architect, CEO Roatan Tourist Bureau👤 Hosts: Rik Eerdekens & Luca Arrigo00:00 Cold open — Radical honesty00:37 Intro — Meet Gonçalo Hall01:08 What is Roatan? The Caribbean's secret island03:16 Prospera: where gene therapy is legal07:01 Is Roatan ready for digital nomads?07:15 The communities already inside Prospera09:12 Roatan vs Prospera vs Infinita — how it's structured10:30 Layer 1, 2, 3: built like a blockchain11:23 $5k/year residency — 7 days on-island required12:29 AI + biotech = the perfect storm for Roatan15:21 Gonçalo's AI hub vision for Roatan17:08 Edge City, Zuzalu & the pop-up city movement18:10 The vision: a permanent Zuzalu19:37 The $11B lawsuit vs the Honduras government22:15 Exponential company growth since 202223:09 The Dubai parallel — how hubs actually form24:00 Prospera as an OS for any country26:44 Are there other free economic zones?26:53 5,000 free zones — only 3 work27:49 What Honduras got uniquely right28:56 Who actually runs free economic zones?29:02 Dubai's hub strategy — specialized zones29:44 Why Malta should have a free zone29:55 Free business idea: genome sequencing franchise32:43 The marketing gap across 5,000 zones33:24 Community IS the marketing strategy35:57 Zanzibar & AI agents with bank accounts37:58 What's your playbook for building communities?38:11 The Seth Godin tribes framework in practice42:08 "I had an identity crisis" — the origin of Destination Architect46:27 Why nobody else has done what Gonçalo does49:31 The Bansko model: bottom-up, no government needed50:59 Going West — the oldest entrepreneurial play51:33 Why Madeira actually worked (the perfect storm)53:32 Malta is having a moment54:17 Life design & self-knowledge55:23 Self-knowledge is the operating system57:00 ADHD as a superpower58:25 The ideal destination checklist01:01:00 "I decided to be less happy for a year"01:01:31 Does every nomad have ADHD?01:01:46 Work like a lion, not a cow01:03:28 Outro

Episode 6 of the God Mode Pod — holiday edition. Rik's in the mountains, Ben's somewhere in North Africa, and Luca just got off the slopes. We're still here though.This week: AI content is going parabolic, OpenAI shut down Sora (and lost a $1B Disney deal), Elon announced a chip factory 50x the size of TSMC, ElevenLabs just got disrupted, a GitLab founder cured his own cancer with AI, and Anthropic has a secret model called Claude Mythos.We get into all of it.Hosted by Rik, Ben & Luca00:00 Cold Open00:42 Welcome to Episode 6 - Holiday Edition01:30 AI Content Is Going Parabolic (Elon's Chart)06:17 OpenAI Kills Sora - Disney's $1B Deal Is Dead10:03 Deep Research Is the Most Expensive Thing OpenAI Runs13:59 Google's Compute Crisis & Why Gemini Keeps Crashing16:00 Elon's TerraFab: 50x the Capacity of TSMC20:47 ElevenLabs Disrupted by Mistral & Google Gemini Live25:19 Open Source Memory Layer - Save 90% on Tokens28:39 Google TurboQuant: 6x Memory Compression, Zero Accuracy Loss31:55 GitLab Founder Cures His Own Cancer With AI35:39 Claude Mythos: Anthropic's Hidden Model Leaked38:41 Predictions41:43 OutroNothing in this podcast is financial or medical advice. We're just three guys talking about AI.

Naval's tweet hit different this week: "Coding an app is the new starting a podcast." We are three guys who just started a podcast and code all day. Felt personal.Episode 5 covers Travis Kalanick's new company Adams and his bits-to-atoms theory of the physical world, the SaaS apocalypse ($830 billion wiped in six days), Google Stitch threatening Figma's existence, Stripe launching agent-to-agent payments, MiniMax releasing a model that trained itself, Replit raising $400 million at a $9 billion valuation, Karpathy scoring every US job on AI replacement risk, and one guy who cured his dog's cancer using ChatGPT and $3,000.We also get into capital as a weapon, the Claude usage doubling window, Luca's Maltese legal tech app hitting 1,000 signups, and whether oil fear is replacing AI fear in the markets.Rik hosts from the mountains. Ben is building four products at once. Luca is digitising Maltese court documents and loving it.00:00 Intro + Naval's "Coding is the new podcast" tweet01:00 Ben's viral token usage and Anthropic doubles Claude limits03:32 Luca's Maltese legal app: 1,000 signups, 50 paid subscribers05:08 Replit vs other vibe coding tools06:11 Travis Kalanick: "If it's easy, it's not valuable"07:58 Travis Kalanick's new company: Atoms (the bits to atoms analogy)13:00 Capital as a weapon: Uber, ChatGPT and subsidised growth15:14 The SaaS Apocalypse: $830 billion wiped in 6 days23:31 Google Stitch drops and Figma falls 8% in a day27:25 Stripe MPC: agents paying agents31:50 MiniMax M2.7: a model that trained itself 100 rounds with no humans35:48 Man cures his dog's cancer with ChatGPT for $3,00040:48 Karpathy scores every US job on AI replacement risk44:00 Replit raises $400M at a $9 billion valuation51:14 Luca's $350/month Replit bill, live on air55:37 Predictions: robot tennis and NemoClaw59:36 Oil fear vs AI fear, gas in Libya and what comes next1:02:35 Outro

This episode explores the rapid evolution of AI, its industry impacts, and future opportunities, including AI's role in management, legal, and scientific breakthroughs. Guests discuss AI adoption, government use, and innovative applications like underwater drones and prediction markets. key  topicsAI adoption gaps across industriesTheoretical vs observed AI capabilitiesAI's impact on management, legal, and creative fieldsGovernment use of AI and model securityInnovative AI applications like underwater drones and prediction markets00:00 Introduction to AI's Rapid Evolution01:42 Personal Experiences and Adventures03:34 AI Adoption Across Industries06:14 Opportunities and Challenges in Management08:45 The Future of Entrepreneurship in an AI World11:18 Risk and Liability in AI Usage13:54 Identifying Opportunities in Education and Healthcare16:31 The Slow Adoption of AI in Education19:14 Institutional vs Individual AI22:05 Empowering Employees with AI Tools22:40 Government AI Models: A Year Behind24:50 The Challenges of Updating Government AI29:50 AI's Creative Problem-Solving Abilities33:58 AI's Impact on Research and Development37:54 Meta's Acquisition of Moldbook43:52 The Future of AI Agents and Social Media47:53 Global AI Adoption: A Surprising Heatmap49:50 The Global AI Divide and UBI55:14 Cloudflare's Thundercrawl and Data Scraping59:36 Nvidia's NemoClaw and the Future of AI Platforms01:02:24 Perplexity's Personal Computer: A User-Friendly AI Solution01:08:23 Predictions for the Future of AI and Technology

the latest developments in AI, government contracts, effective altruism, and the future of AI models and hardware. we analyze the strategic moves of OpenAI, Anthropic, and Google, exploring their philosophies, market impacts, and technological advancements.key topicsAI government contracts and ethical considerationsThe influence of effective altruism on AI organizationsMarket strategies of OpenAI, Anthropic, and GoogleAdvancements in AI models and open source alternativesThe impact of hardware and infrastructure on AI development

In this episode of the God Mode Pod, the hosts discuss the significant impact of AI on employment, productivity, and the future of work. They explore the implications of AI-driven job cuts, the potential for Universal Basic Income, and the shifting landscape of real estate and economic opportunities. The conversation also delves into the rise of AI startups, the evolution of design tools, and the innovative use of OpenClaw in podcast production. The episode concludes with insights on the future of AI models and their applications across various industries.takeawaysAI is leading to unprecedented job cuts in major companies.The productivity of AI is reshaping the workforce.Individuals must adapt to using AI to remain competitive in their jobs.Real estate markets may shift as tech jobs become less centralized.Universal Basic Income could be a solution for job displacement due to AI.AI startups are emerging as a response to the changing job landscape.Design and creative industries are being transformed by AI tools.OpenClaw is revolutionizing how podcasts are produced and managed.The future of AI models is promising, with rapid advancements expected.Companies must innovate to avoid being disrupted by AI advancements.

In the first episode of the God Mode Podcast, hosts Rik, Ben, and Luca dive into the fast-paced world of AI, discussing their backgrounds, the latest innovations, and the competitive landscape between major players like OpenAI and Anthropic. They explore the implications of AI on business automation, user experience, and the future of app development, emphasizing the importance of skill stacking and adaptability in a rapidly evolving tech environment. The episode concludes with reflections on the potential of AI to transform industries and the necessity for businesses to embrace these changes.takeawaysAI is evolving rapidly, and staying updated is crucial.The competition between OpenAI and Anthropic is intensifying.Data centers will play a significant role in the future of AI.Skill stacking is essential for success in the AI landscape.User experience in AI applications is becoming increasingly important.AI can automate many traditional business processes.Emerging technologies are democratizing app development.The future of AI holds exciting possibilities for innovation.Businesses must adapt to leverage AI effectively.Collaboration and integration of AI tools can enhance productivity.