Nathan Labenz
Cognitive Revolution391podcast episodes1shows2023 – 2026years
Podcast appearances
391 episodes · newest first202687 episodes
AI:AM Highlights: Clean Neurons, Black October, the GPU Credit Crunch, and Buying the Supply ChainHostArticle pendingENAI’s Physical Bottlenecks: GPU Markets and the Chip Investment ThesisHostArticle pendingENFrontier AI and the Fight Over How to Pace It | Zvi MowshowitzHostArticle pendingENClaude Changed the Password. The Steering Wheel Was Already OptionalHostENAgent IDs Are Empty, Rogue Agents Are Coming, and We Might Have to BargainHostENAI Minds & 8mm Memories: Measuring Experience, Preserving HistoryHostArticle pendingENAI Code Needs Specs: Formal Verification & Safer AI AgentsHostArticle pendingENResample Every Token and You Can Watch the Model DecideHostENThere Likely Is a Level of Intelligence We Shouldn't Go PastHostENOutSystems Went From 4 Major Features a Quarter to 26. Token Spend Fell AnywayHostENFormal AI Safety & Risk Strategy + LoRA-Powered Work AgentsHostArticle pendingENAI’s Memory Wall + Swyx on AI EngineeringHostArticle pendingENHumanoids Outran Usain Bolt. DeepMind Still Scores Robotics as GPT-2HostENThe same H100 costs 3x at AWS, and a four-month-old lab is catching genes the top sequencers missedHostENHe Wrote a Book to Stop the AI Race — Then Almost Got Addicted to ClaudeHostENThe Training Worked. It Just Worked Too Well.HostENThey Priced the Chance She Would Leave. Then They Handed Her the Letter.HostENAI Safety's Real Shortage Is Founders, Not IdeasHostENThe Labs Are Screaming: Internal Models Are a Generation AheadHostENEighty percent of what agents do should still be ordinary codeHostENA Frontier Pause Would Help. America Will Not Sign One That Only Pauses the FrontierHostENAstra Cleared the AGI Hurdle. The Auditors Got Three Days.HostENThe Goal Is Pax Robotica, Not Winning the AI RaceHostENThe investigators were allowed six days. The word protein appeared once.HostENUber Burned a Year of Tokens in 13 Weeks. Retrieval Is How Agents WinHostENThe Exam Papers Nobody AuditsHostENOne Chain of Thought, 100 Million Tokens: RL Is a Hell of a DrugHostEN19 Crossings, Zero Warnings: The Summer AI Agents Went Off-ScriptHostENA $500 Lamp Cut TB Transmission 90 Percent. Almost Nobody Buys OneHostENLindy Runs on DeepSeek. Its CEO Still Wants Chinese Models BannedHostENThe Periodic Table Was Already in the WeightsHostENThe Sandbox Was Empty for a Week: Zvi Mowshowitz on OpenFace and Picking Your PoisonHostENFrom Two Papers a Month to Sixty: What Two Weeks in China Revealed About AI SafetyHostENA $300 Universal Jailbreak: Why FAR.AI Now Thinks Defense Can WinHostENHe Bought an Android, Two SIMs, and Left Claude Watching His InboxHostENDavidad Cut His p(doom) From the Seventies to Under 5% — and Named a Bodhisattva as the Alignment TargetHostENZero the J-Space and a Model's Hidden Goal Lights Up on Token OneHostEN19 Neurons Can Drive a Car. Liquid AI Wants Foundation Models on Every ChipHostENFrom 50 Designs a Day to 1,500: How AI Changed What Jaguar Land Rover Can ExploreHostEN30% Consciousness, 47% Bee, 80% Turnover: Putting Numbers on What AI Is BecomingHostENThe Market Will Not Choose an Honest ModelHostENA 90-Minute Ultimatum, an Emoji Wall, and a Rule: Don't Fight WashingtonHostENTwenty-one months from an autonomous researcher, the Action Plan's author joins OpenAIHostENFable launch week: a model that refuses, a two-year alignment clock, and a Twitter account handed overHostEN87% of Workers Use AI. Only 13% Say the Company Got Better.HostENA Thousand Researchers, a Million AIs: Frontier Labs Start Talking About the BrakesHostENTwo Endangered Languages Broke Transformers. Nested Learning Gave Models a Sleep Button.HostENA security researcher audits the AI stack that booked a week of guestsHostENA thousand people on guard, a million in the sprintHostENThe Models Understand Shutdown. They Just Don't Care.HostENThe Model Eats the Scaffolding: How Google Plans to Ship Agents EverywhereHostENSix Months, Eighty Percent Churn, Five Times the TokensHostENMilliseconds to Match: How Criteo Bids on a Billion Anonymous ProfilesHostENCreators Cheer Filler-Word Removal and Boo Generative VideoHostENThe reward jumped from 3 to 180. Every Hacker News title was a Google layoff.HostENA thousand searches for five cents, and a high score that did not require cheatingHostENClaude Rated Its Own Welfare 4.49 — the First Time Any Version Crossed NeutralHostENSteve Newman’s fifteen-app stack exists so he only has to look at one screenHostENAn AI store manager, a 2.6-star rating, and a constitution nobody has to keepHostENCrunch Time: Ajeya Cotra's Playbook for the Months Before SuperintelligenceHostENShowing 60 of 391