Nathan Labenz
Cognitive Revolution391Podcast-Folgen1Sendungen2023 – 2026Zeitraum
Podcast-Auftritte
391 Folgen · neueste zuerst202687 Folgen
AI:AM Highlights: Clean Neurons, Black October, the GPU Credit Crunch, and Buying the Supply ChainModerationArtikel folgtENAI’s Physical Bottlenecks: GPU Markets and the Chip Investment ThesisModerationArtikel folgtENFrontier AI and the Fight Over How to Pace It | Zvi MowshowitzModerationArtikel folgtENClaude Changed the Password. The Steering Wheel Was Already OptionalModerationENAgent IDs Are Empty, Rogue Agents Are Coming, and We Might Have to BargainModerationENAI Minds & 8mm Memories: Measuring Experience, Preserving HistoryModerationArtikel folgtENAI Code Needs Specs: Formal Verification & Safer AI AgentsModerationArtikel folgtENResample Every Token and You Can Watch the Model DecideModerationENThere Likely Is a Level of Intelligence We Shouldn't Go PastModerationENOutSystems Went From 4 Major Features a Quarter to 26. Token Spend Fell AnywayModerationENFormal AI Safety & Risk Strategy + LoRA-Powered Work AgentsModerationArtikel folgtENAI’s Memory Wall + Swyx on AI EngineeringModerationArtikel folgtENHumanoids Outran Usain Bolt. DeepMind Still Scores Robotics as GPT-2ModerationENThe same H100 costs 3x at AWS, and a four-month-old lab is catching genes the top sequencers missedModerationENHe Wrote a Book to Stop the AI Race — Then Almost Got Addicted to ClaudeModerationENThe Training Worked. It Just Worked Too Well.ModerationENThey Priced the Chance She Would Leave. Then They Handed Her the Letter.ModerationENAI Safety's Real Shortage Is Founders, Not IdeasModerationENThe Labs Are Screaming: Internal Models Are a Generation AheadModerationENEighty percent of what agents do should still be ordinary codeModerationENA Frontier Pause Would Help. America Will Not Sign One That Only Pauses the FrontierModerationENAstra Cleared the AGI Hurdle. The Auditors Got Three Days.ModerationENThe Goal Is Pax Robotica, Not Winning the AI RaceModerationENThe investigators were allowed six days. The word protein appeared once.ModerationENUber Burned a Year of Tokens in 13 Weeks. Retrieval Is How Agents WinModerationENThe Exam Papers Nobody AuditsModerationENOne Chain of Thought, 100 Million Tokens: RL Is a Hell of a DrugModerationEN19 Crossings, Zero Warnings: The Summer AI Agents Went Off-ScriptModerationENA $500 Lamp Cut TB Transmission 90 Percent. Almost Nobody Buys OneModerationENLindy Runs on DeepSeek. Its CEO Still Wants Chinese Models BannedModerationENThe Periodic Table Was Already in the WeightsModerationENThe Sandbox Was Empty for a Week: Zvi Mowshowitz on OpenFace and Picking Your PoisonModerationENFrom Two Papers a Month to Sixty: What Two Weeks in China Revealed About AI SafetyModerationENA $300 Universal Jailbreak: Why FAR.AI Now Thinks Defense Can WinModerationENHe Bought an Android, Two SIMs, and Left Claude Watching His InboxModerationENDavidad Cut His p(doom) From the Seventies to Under 5% — and Named a Bodhisattva as the Alignment TargetModerationENZero the J-Space and a Model's Hidden Goal Lights Up on Token OneModerationEN19 Neurons Can Drive a Car. Liquid AI Wants Foundation Models on Every ChipModerationENFrom 50 Designs a Day to 1,500: How AI Changed What Jaguar Land Rover Can ExploreModerationEN30% Consciousness, 47% Bee, 80% Turnover: Putting Numbers on What AI Is BecomingModerationENThe Market Will Not Choose an Honest ModelModerationENA 90-Minute Ultimatum, an Emoji Wall, and a Rule: Don't Fight WashingtonModerationENTwenty-one months from an autonomous researcher, the Action Plan's author joins OpenAIModerationENFable launch week: a model that refuses, a two-year alignment clock, and a Twitter account handed overModerationEN87% of Workers Use AI. Only 13% Say the Company Got Better.ModerationENA Thousand Researchers, a Million AIs: Frontier Labs Start Talking About the BrakesModerationENTwo Endangered Languages Broke Transformers. Nested Learning Gave Models a Sleep Button.ModerationENA security researcher audits the AI stack that booked a week of guestsModerationENA thousand people on guard, a million in the sprintModerationENThe Models Understand Shutdown. They Just Don't Care.ModerationENThe Model Eats the Scaffolding: How Google Plans to Ship Agents EverywhereModerationENSix Months, Eighty Percent Churn, Five Times the TokensModerationENMilliseconds to Match: How Criteo Bids on a Billion Anonymous ProfilesModerationENCreators Cheer Filler-Word Removal and Boo Generative VideoModerationENThe reward jumped from 3 to 180. Every Hacker News title was a Google layoff.ModerationENA thousand searches for five cents, and a high score that did not require cheatingModerationENClaude Rated Its Own Welfare 4.49 — the First Time Any Version Crossed NeutralModerationENSteve Newman’s fifteen-app stack exists so he only has to look at one screenModerationENAn AI store manager, a 2.6-star rating, and a constitution nobody has to keepModerationENCrunch Time: Ajeya Cotra's Playbook for the Months Before SuperintelligenceModerationEN60 von 391 angezeigt