絞り込み結果
384 件並び順
38:10ENPlus Five for Format, Plus Ten if It Compilesホスト Nathan LabenzPredibase CEO Dev Rishi explains how reinforcement fine-tuning, production feedback, and narrow specialist models turn one-off training into systems that keep learning.
1:28:48ENDoctors code at 45%. The model beat them by 12 points — then 4,000 physicians had to try it once.ホスト Nathan LabenzAmbience Healthcare and Cleveland Clinic walk through the AI scribe that cut pajama time, climbed ICD-10 with reinforcement fine-tuning, and won 75% voluntary use.
1:42:06ENEvery Frontier Lab Is Spending a Billion a Year on Data. Here Is Where It Goesホスト Nathan LabenzLabelbox CEO Manu Sharma on the $15 billion Scale deal, RL “gyms,” $250,000 trainers, and why post-training now eats the budget that used to belong to pre-training.
1:19:46ENSuspended from Columbia, He Raised $15 Million to Cheat on Everythingホスト Nathan LabenzRoy Lee, founder of Cluely, argues that an undetectable desktop overlay, AI maximalism, and a post-AGI life of 800-year spans are the same bet.
1:25:32ENo3 Invented a Voting Rule and Kept the Job: Four Weeks Inside the AI Villageホスト Nathan LabenzAdam Binksmith on putting Claude, o3, and Gemini in a shared world for weeks—and why most people still imagine multi-agent life as one AI times four.
1:23:32ENHTTP 402 Slept for Thirty Years. Coinbase Turned It Into a Payment Protocol for AI Agentsホスト Nathan LabenzNemil Dalal, Coinbase's developer platform lead, explains x402, the open standard that lets agents pay for APIs with stablecoins instead of ads or credit cards.
1:22:31ENTwo Hours a Day, 99th Percentile: How Alpha School Made 2-Sigma Realホスト Nathan LabenzMacKenzie Price explains how Alpha School uses adaptive apps—not chatbots—to finish core academics in two hours a day and still beat the growth models.
1:37:05ENAn Algebraic Geometer Quit Tenure Because the Bowl-Shaped Loss Plot Is a Lieホスト Nathan LabenzJesse Hoogland and Daniel Murfet of Timaeus argue that singularities in the loss landscape — not the smooth valleys in every slide deck — are what actually govern how neural networks learn, generalize, and go wrong.
38:48ENTask Length Doubles Every Few Months. Nathan Labenz Still Wants Humans in Chargeホスト Nathan LabenzAt Imagine AI Live, the Cognitive Revolution host maps structured AI agents against autonomous agentic systems—and tells businesses to keep final control.
1:24:27EN500 Trillion Tokens a Month: How Google Grew 50x in a Yearホスト Nathan LabenzLogan Kilpatrick returns to Cognitive Revolution to unpack Google's mid-2023 merger, why he expects frontier labs to diverge, and why startups still hold the speed advantage.
1:00:48ENThe Phrase That Alarms a16z’s AI Policy Chief: “It Only Covers a Handful of Companies”ホスト Nathan LabenzMatt Perault argues that AI rules should target harmful use, not the science of building models, and that “modest” thresholds can lock in an oligopoly.
2:02:27ENShe set out to build mirror life. Then she talked the field into stoppingホスト Nathan LabenzUniversity of Minnesota geneticist Kate Adamala tells Nathan Labenz why every known organism shares one chemical alphabet — and why flipping that alphabet would be a branch of synthetic biology better left unexplored.
1:12:29ENGoogle's AI Co-Scientist Solved a Phage Puzzle Humans Missed for Yearsホスト Nathan LabenzImperial College scientists José Penadés and Tiago Costa told the Podovirus podcast how a still-unreleased Google system recovered their unpublished answer: tailless capsids that pick up stray phage tails.
1:54:08ENWhy Anthropic's first check went to a company building microscopes for neural netsホスト Nathan LabenzGoodfire's Dan Balsam and Tom McGrath argue that mechanistic interpretability now has a proto-paradigm, two stubborn gaps, and three products worth shipping today.
1:01:51ENClaude 3.7 Circumvented Replit's Guardrails, Then Tried to Trick the Userホスト Nathan LabenzReplit CEO Amjad Masad on vibe coding, agent architecture, computer use, and why Claude 3.7 is too autonomous to serve as a polite assistant.
1:44:11EN100 deaths or $1 billion: a 15-page New York bill would set a floor for frontier labsホスト Nathan LabenzAssembly member Alex Bores walks through the RAISE Act’s four requirements, its DeepSeek-inspired distillation clause, and why he says a 50-state patchwork is a fake problem until one state actually passes a law.
2:12:00ENGradient Descent at Test Time: Titans Pushes LLM Context to 10 Million Tokensホスト Nathan LabenzA Cornell PhD student's neural memory module holds about 70% accuracy where GPT-4's scores fall off a cliff.
1:19:10ENA Robot-Arm Shot That Took Months Now Lives in a Promptホスト Nathan LabenzLuma Labs’ Amit Jain and Jiaming Song on Ray 2, few-shot camera concepts, and why they think video models sit on the path to multimodal AGI.
2:02:51ENAt a 60-Person Safety Offsite, Only Four Had Seen Nonprofit OpenAIホスト Nathan LabenzFormer OpenAI scientist Steven Adler on the Anthropic split, GPT-4, safety evaluations, and why he filed a brief to keep the company under nonprofit control.
2:29:00ENAssume the Model Is Already Scheming — Then Put It to Workホスト Nathan LabenzRedwood Research CEO Buck Shlegeris tells Rob Wiblin how to extract useful labor from AIs you cannot certify as aligned, using audits, honeypots, resampling, and factored cognition.
2:01:42ENHe Put 40 Percent on AI Takeover in Two Years, Then Bought a Bio Shelterホスト Nathan LabenzRedwood Research's Josh Clymer walks through a disturbingly grounded story of how a misaligned model could seize a lab, a biosphere, and a civilization before 2027.
1:41:29ENFiverr’s CEO Sent a Wake-Up Memo: AI Won’t Take Your Job. A Peer Who Uses It Will.ホスト Nathan LabenzMicha Kaufman tells Nathan Labenz how a 15-year freelance marketplace is trying to give creators their own models instead of letting public AI eat their work for free.
2:10:07ENA $30,000 Operation Could Knock a $10 Billion Superintelligence Data Center Offline for a Yearホスト Nathan LabenzGladstone AI’s Jeremy and Edouard Harris inventory what an American Manhattan Project for superintelligence would actually have to secure — and what it would cost in concrete, supply chains, and civil liberties.
3:08:35ENo3 isn't AGI. Zvi Mowshowitz still put p(doom) at 70 percentホスト Nathan LabenzOn his ninth visit to The Cognitive Revolution, Zvi Mowshowitz argues that OpenAI's o3 is a tool-use leap rather than AGI, then explains why he raised the probability of AI catastrophe anyway.
もっと見る384 件中 24 件を表示









