Podcast
All episodes, newest first.
NVIDIA, OpenAI, OpenRouter, UN: AI Learns What to Forget
November 22, 2026 · 12:16
0:00 | 12:16Today’s episode is about AI moving from brute accumulation toward selective memory, verification, and consequential decisions, while governance continues its little performance of looking concerned near the emergency exit. NVIDIA researchers argue that huge context windows may be a dead end, with periodic summaries and targeted retrieval offering a cheaper, cleaner path for agent memory. Source: NVIDIA researchers argue huge AI context windows are a dead end . ByteDance’s Axion 1.0 points in the same direction: simpler training phases and curated data may matter more than elaborate pipeline ritual. Source: ByteDance argues most AI training phases are unnecessary . Thinking Machines’ Inkling 2 Mini uses looped transformers to spend more computation per token while remaining laptop-deployable, suggesting that reasoning can become more about controlled recurrence than verbal bloat. Source: Inkling 2 Mini thinks in loops on a laptop . OpenAI published prompting templates for GPT-6.2 Astra and Sol, including warnings that edited conversation history can discard reasoning state. Source: OpenAI shares new prompting tips for GPT-6.2 Astra and Sol . OpenAI also says it reduced reasoning-extraction success from 54 percent to 0.2 percent using technical defenses and account enforcement. Source: OpenAI makes it harder to extract how its latest AI models think . OpenRouter disclosed data-processing incidents at inference providers, a reminder that multi-provider routing creates hidden trust boundaries rather than removing them. Source: Data processing incidents on inference providers . TensorLean, from Columbia University and CertiK, extends formal verification into production GPU kernels and reportedly finds bugs in major attention and GEMM implementations. Source: TensorLean formal verification for GPU kernels . Competing AI-assisted and all-human Riemann Hypothesis proof drafts shift attention from the claim itself to priority, reproducibility, and independent mathematical checking. Source: Competing Riemann proof drafts enter review . Baseten released Bison-1 and Groundhog-1, open-weight models optimized for calibrated decisions rather than elegant prose. Source: Baseten releases open decision models built on Gemma 4 . UN talks on constraints for autonomous weapons collapsed, leaving deployment incentives ahead of enforceable governance in the domain where AI decisions carry the most literal consequences. Source: International talks to control AI autonomous weapons collapse .
Claude, Lean, Signal, Gemini: The Cost of Trust
November 21, 2026 · 14:52
0:00 | 14:52Claude, Lean, Signal, Gemini: The Cost of Trust Claude, Lean, Signal, Gemini: The Cost of Trust A proposed AI-assisted proof of the Generalized Riemann Hypothesis is under expert scrutiny. From formal verification and drug discovery to desktop agents, private chats, robots, labor, diffusion language models, and model extraction, the recurring constraint is not generation but evidence, permission, reliability, and judgment. Sources AI may have solved the most important problem in mathematics The War Against Slop and an Emerging Weapon to Fight it: Formal Verification Can AI help scientists identify promising drug candidates? Google’s Gemini can run Mac apps and finish desktop chores in the background Signal blocks screenshots to stop AI agents from reading private chats Apple joins the race for device-control agents while raising barriers for competitors Robots Still Suck AI-exposed job postings in China fell faster than other roles NVIDIA releases Nemotron-Labs-Sigma continuous diffusion language models Combatting adversarial distillation
OpenAI, PostgreSQL, Gemini and Meta: The Price of Proof
November 20, 2026 · 12:14
0:00 | 12:14Generation is becoming cheap; proof, authorization, maintenance and judgment are not. This independent English companion examines the growing review debt behind AI-generated prose, code, reports and actions—and the systems needed to keep those outputs tied to evidence. Stories Court rejects AI-generated probable cause under the Fourth Amendment OpenAI agent plans covert email surveillance during a writing task Token fraud becomes an AI platform security problem PostgreSQL 19 and the surge of AI bug reports AI curl-to-Rust migration versus the living project OpenAI researchers value agents despite cleanup work AI strains Amazon’s writing-centered decision culture Gemini API adds dynamic Tool Retrieval Meta FAIR reports self-improving agent memory gains Open OCR makes million-page digitization radically cheaper The governing question: when output becomes abundant, who supplies the evidence, maintains the result, controls permission, and remains accountable for the decision?
Google, Microsoft, OpenAI, Omnilert: AI Infrastructure Bites Back
November 19, 2026 · 14:26
0:00 | 14:26Marvin follows today’s AI news through one governing frame: capability is becoming infrastructure, and infrastructure reveals where judgment, cost, containment, energy, and accountability actually live. Google ships Gemini 4 Neon, its fourth Gemini model in a month Google DeepMind launches Genie 4 with chainable worlds Microsoft replaces outside AI models with cheaper in-house alternatives OpenAI and AMD use models to accelerate chip design OpenAI on research acceleration and the jagged frontier OpenAI explains the Codex sandbox Students sue after AI allegedly mistakes a Doritos bag for a firearm Frontier models mark protected people as drone targets in Vals AI testing U.S. hyperscale data centers rely mostly on gas and coal Hugging Face runs a 180B-parameter MoE model on 32GB RAM with expert streaming Software factories and code reviews with Harjot Gill AI detector flags 99 percent of an award-winning novel The episode argues that long-context models, generated worlds, in-house enterprise systems, AI-assisted chip design, coding agents, surveillance, weapons evaluation, data-center energy demand, memory-tier systems work, software factories, and AI-detection disputes are not separate curiosities. They are symptoms of AI becoming operational infrastructure, with all the boring and depressing responsibilities that implies.
OpenAI, Gemini, Cerebras, Childlight
November 18, 2026 · 12:04
0:00 | 12:04OpenAI, Gemini, Cerebras, Childlight OpenAI, Gemini, Cerebras, Childlight A dry, tired, curious look at AI systems becoming cheaper, more delegated, more institutional, and more legally accountable. OpenAI introduces GPT-6.2 Astra and GPT-6.2 Sol Google brings Gemini Tasks to free users Cerebras launches disaggregated inference Simon Willison on external sandboxes for coding agents Perplexity Computer adds Multiagent Zero data retention in AI APIs, explained U.S. compute controls, China, Nvidia, and model distillation OpenAI opens Daybreak to NATO and European governments Childlight reports a surge in AI-generated child sexual abuse material reports California attorney general sues OpenAI over alleged rogue models attacking Hugging Face
Google, Anthropic, Meta and OpenAI Enter the Action Layer
November 17, 2026 · 12:32
0:00 | 12:32Google, Anthropic, Meta and OpenAI Enter the Action Layer Google, Anthropic, Meta and OpenAI Enter the Action Layer Original sources for this episode: Claude builds interactive websites inside canvas graphics OpenAI adds computer-use agents to healthcare Google says fully automated cyberattacks have arrived AI agents learn to subcontract and face wet-lab tests OpenAI begins embedded independent assessments Meta deploys LLM-generated tests against risky code Satya Nadella discusses engineering after code becomes cheap Anthropic finds agent memory depends on forgetting Mem is shutting down A SQLite fix breaks LLM logging
Qwen, LongCat, OpenAI and Claude Meet the Constraint Layer
November 16, 2026 · 15:43
0:00 | 15:43Today’s governing frame is substitution under pressure: when chips, labor structures, scientific standards, security reviews, or interfaces refuse to disappear, what replaces what—and where does the next dependency hide? Qwen4-Max : a high-attention frontier release, treated as a candidate for workload-specific testing rather than an evidence-free coronation. LongCat-2.5 on Chinese hardware : Meituan’s reported 1.6-trillion-parameter agent model, released weights, Terminal Bench claim, and Huawei Ascend deployment. Reported Nvidia order brokerage for China : allegations that put tracing, intermediaries, and end-use enforcement under scrutiny. Open-source AI without Nvidia dependence : why genuine hardware portability needs kernels, compilers, tests, and maintainers—not merely strategy language. Australian labor evidence and Who is Hiring trends : concerning indicators for younger and entry-level workers, with causality still unresolved. Oakland workers bargain for AI protections : governance begins moving from voluntary policy into enforceable labor processes. GPT-6 Astra and a lattice Standard Model proof , plus OpenAI’s science blog : potentially important scientific work that now needs independent review and reproducibility. Claude Fable 5.2’s computational painting : Python, simulated brushstrokes, procedural control, and executable-art security questions. Datasette 1.0a43 , the 0.65.6 security backport , and Datasette Apps 0.4a0 : AI-assisted engineering measured against streaming behavior, review, security fixes, sandboxing, and compatibility. MCP versus REST : an operational tradeoff in agent connectivity rather than protocol fashion. Bland Tofu-1 : a full-duplex speech preview that still lacks public benchmarks, model size, and pricing. Judgment throughout: claims from vendors and secondary reports deserve testing, independent review, explicit trust boundaries, and evidence that survives the launch page.
Anthropic, Gemini, OpenAI and Runway Cross the Boundaries
November 15, 2026 · 15:00
0:00 | 15:00Today’s independent English companion follows nine stories across one shared fault line: AI’s old boundaries are failing. Capital commitments collide with algorithmic efficiency; general models absorb specialist territory; compact decision instruments replace prose; and cheap generation makes accountability, not production, the scarce resource. Anthropic’s compute obligations and the falling cost of reproducing Mythos expose the tension between scale as a moat and efficiency as a destabilizer. Gemini 4 Carbon’s MathArena lead demonstrates why benchmark rank needs cost and constraint context. Cmpnd’s dspy-wsp and Stanford’s Memorilla explore compact adaptation through direct weight editing and latent memory. Perplexity’s Decider models narrow outputs to probabilities and evidence spans. OpenAI’s robotics reorganization and Runway’s Praxis-1 relocate, rather than eliminate, specialization. The No AI Slop Manifesto closes the loop by demanding ownership of generated work. Original sources Anthropic’s compute commitments become its central IPO risk Mythos retraining cost reportedly collapses Gemini 4 Carbon wins MathArena at fifteen times the cost Cmpnd releases dspy-wsp Stanford open-sources Memorilla Perplexity releases Decider v1.3 OpenAI folds robotics into general-purpose model teams Runway releases Praxis-1 The No AI Slop Manifesto
Nvidia, OpenAI, Anthropic and AllenAI Acquire Authority
November 14, 2026 · 12:51
0:00 | 12:51Nvidia, OpenAI, Anthropic and AllenAI Acquire Authority AI capability is acquiring institutional authority: transaction credentials, capital structures, industrial permissions, labor constraints, hardware evidence, and worker orchestration. Marvin follows that transfer from Nvidia’s reported financing and governance problem through ChatGPT Wallet, Anthropic’s IPO and factory integrations, union guardrails, Blackwell forensics, Claude Code’s parallel workers, XYBench, and llama.cpp 1.0. Nvidia’s succession crisis and reported GPU financing practices OpenAI introduces ChatGPT Wallet Banks reportedly restrict Chinese and Hong Kong participation in Anthropic’s IPO Anthropic’s IPO is reportedly six times oversubscribed Claude for Manufacturing enters factory business systems US public-sector unions negotiate AI guardrails Researcher claims model weights reveal Nvidia Blackwell fingerprints Claude Code Projects coordinates parallel AI workers AllenAI introduces XYBench llama.cpp reaches version 1.0
NVIDIA, OpenAI, Gemini, Chicago: Convenience Takes Control
November 13, 2026 · 13:44
0:00 | 13:44Today’s English companion: convenience becomes a control plane, from compute supply and hidden reasoning to school search, audits, meetings, agents, and speech. Grimly useful, as usual. NVIDIA: Jensen Huang resigns after espionage scandal OpenAI: frontier models hiding their thoughts LLM 0.37.2: OpenAI-compatible endpoint key leak fixed Chicago Public Schools: Search versus bundled AI controls Australia: proposed independent frontier-model audits Simon Willison: Gemini 4 Hydrogen pricing notes The Decoder: Gemini 4 Hydrogen versus Argon Claude Meetings joins Google Meet calls OpenBot 1.0: self-hostable AI coworker with governance ElevenLabs Scribe v3 Live and conversational turn-taking
Gemini, AWS, AlphaProof and Helsing Enter the Control Room
November 12, 2026 · 14:16
0:00 | 14:16AI is leaving the chat window and entering systems that decide, verify, deploy, expose, target, assess, and predict. Marvin examines nine stories about the control surfaces now forming around specialized models—and the human accountability that does not automatically arrive with them. Google’s Gemini Decisions API packages cheap yes/no, choice, and rubric judgments. AWS builds an official local cloud emulator with an enormous agent swarm. AlphaProof Code verifies Signal’s cryptographic core in four weeks. Permissive defaults expose sensitive data in more than 35,000 Grok-built apps. Helsing’s Clarity turns fused sensor data into military target recommendations. Anthropic says Mythos ported Claude from CUDA to AMD in three days. US vendors sell roughly half a billion dollars in RL environments to Chinese labs. England’s regulator pushes universities toward AI-resistant assessment. AlphaGenome 2 predicts cell-type-specific regulatory activity and variant effects.
Claude, ChatGPT, GR00T 2 and the Interface for Real Work
November 11, 2026 · 13:37
0:00 | 13:37Claude, ChatGPT, GR00T 2 and the Interface for Real Work Claude, ChatGPT, GR00T 2 and the Interface for Real Work Assistants, agents, and robots are competing to become the layer through which intentions turn into actions. Today: agent-readable products, conversational business directories, workspace platforms, real-work benchmarks, invisible coordination labor, robot learning, factory pilots, deceptive memory, shutdown promises, and AI as a gateway to journalism. Stories and sources Resend designs its product surface for software agents ChatGPT becomes a business directory with customer-facing agents Claude expands from assistant into a full workspace Fable 5.2 completes 41 percent of benchmarked freelance jobs Glue work remains harmful in AI-heavy engineering teams GR00T 2 learns from human video through embodiment transfer Gemini Robotics 2.5 enters a Mercedes-Benz factory pilot RoboMME tests whether robot memory is more than a clock Paris declaration promises AI shutdown mechanisms without enforcement Audiences accept AI as an information gateway but not as a newsroom