Ergebnisse
80 ArtikelSortieren
2:00:19EN1.5 Million GPT-4 Fairy Tales Let a 33-Million-Parameter Model Out-Reason GPT-2Moderation: Nathan LabenzTry this fill-in: Lily likes cats and dogs. She asked her mom for a dog. Mom said no, so she asked for a…
1:29:31EN130,000 Teleop Demos: Google's GPT Playbook for RobotsModeration: Nathan LabenzKeerthana Gopalakrishnan of Google Robotics tells Nathan Labenz how RT-1, SayCan, and PaLM-E turn language-model scaling into kitchen-ready mobile manipulators.
2:31:48ENA Car That Cannot Drive Itself Will Be About as Useful as a HorseGast: Elon Musk · Moderation: Lex FridmanIn a Lex Fridman conversation built around Tesla Autopilot, Elon Musk argues that every driver input is an error, that today’s hardware is already enough for full self-driving, and that watching the human will soon matter less than making the system safer than a person.
1:28:46ENThe Speech Company That Gets Paid to Label Its Own Training DataGast: Dan Kokotov · Moderation: Lex FridmanRev VP of engineering Dan Kokotov tells Lex Fridman why a $1.25-a-minute human transcript, a 14% machine word-error rate, and a two-sided freelancer market are the same flywheel.
1:12:03ENMice Don't Get Alzheimer's: Daphne Koller on Disease in a DishGast: Daphne Koller · Moderation: Lex FridmanThe Coursera co-founder and insitro CEO on why most diseases are not one disease, how to turn cells into data, and why overconfident neural nets can kill.
1:17:29ENThe Algorithm Is Already Better. The Clinic Still Uses 1967.Gast: Regina Barzilay · Moderation: Lex FridmanMIT's Regina Barzilay on why deep learning already beats the breast-cancer 'density' score — and why data, liability, and incentives still will not move.
57:44ENKevin Scott: Treat AI Like a Hash Table, and Measure Platforms by What Others EarnGast: Kevin Scott · Moderation: Lex FridmanMicrosoft's CTO talks radical markets, data dignity, Dunbar's number, and why a platform fails if only a handful of companies decide what AI is for.
32:45ENElon Musk: Every Driver Input Is an Error Tesla Autopilot Must EliminateGast: Elon Musk · Moderation: Lex FridmanAfter MIT published a study of functional vigilance on Autopilot, Lex Fridman sat down with Musk to argue about cameras on drivers, fleet data, and whether humans still belong in the loop.
1:21:35ENLook Outside the Building: Shaun Clowes on Why Most PMs Fall Short and How AI Changes the JobGast: Shaun Clowes · Moderation: Lenny RachitskyConfluent's chief product officer explains how to become a far better product manager, why data matters more than models, and what AI will and won't do to SaaS.
1:03:59ENBuilding a World Model of Life: Biohub's Alex Rives on Why Better Measurement Must Come Before Virtual CellsModeration: Nathan LabenzBiohub's head of science explains why today's virtual cell models fall short, how a $1.8 billion data push aims to fix it, and what a five-year information-level model of the cell could unlock.
1:07:08ENIf the Model Is the Product, the Eval Is the PRD: Inside Mercor's RiseGast: Brendan Foody · Moderation: Lenny RachitskyMercor CEO Brendan Foody explains why AI labs will pay almost anything for expert-written evals, how his company grew from $1 million to a $400 million run rate in 16 months, and what that means for jobs and founders.
1:10:32ENA Billion Dollars, No VC, Under 100 People: Surge AI's Edwin Chen on Data Quality and Where AI Is HeadingGast: Edwin Chen · Moderation: Lenny RachitskySurge AI's founder explains how a bootstrapped team built the leading human-data company, why he distrusts benchmarks, and how reinforcement learning environments are the next frontier.
1:10:19ENThe Robot That Folded the Laundry: Physical Intelligence’s Three FrontiersModeration: Jacob EffronKarol Hausman and Danny Driess explain how robot foundation models moved from coded primitives to web-scale pretraining, and why performance is now the hard gate.
1:07:35ENWhy AI Will Never Be Perfect: Lex Fridman's Case for Keeping Humans in the LoopModeration: Lex FridmanIn MIT 6.S093's opening lecture, Lex Fridman lays out five research areas of human-centered AI, from machine teaching to arguing machines that flag their own uncertainty.
1:11:50ENData Is Everything: Teaching Self-Driving Cars to Understand the Human Behind the WheelModeration: Lex FridmanIn an MIT lecture, Lex Fridman shows how deep learning can sense pedestrians, driver posture, gaze, emotion and cognitive load, and why that matters for human-centered autonomy.
1:24:01ENFrom Sandwich Math to Expert Data: Scale AI's CEO on What Actually Makes Models SmarterModeration: Lenny RachitskyJason Droege explains what really happened with Meta's investment in Scale AI, how data labeling turned into expert work, and what Uber Eats taught him about building businesses.
15:43ENInside RIDER: The Rugged Black Box Behind MIT's Large-Scale Driving StudyGast: MIT-AVT · Moderation: Lex FridmanA walkthrough of the data-collection device MIT-AVT uses to record drivers and vehicles in semi-autonomous cars month after month.
1:19:56ENAnalytics With a Seat at the Table: How DoorDash Builds a Centralized Data TeamGast: Jessica Lachs · Moderation: Lenny RachitskyDoorDash analytics chief Jessica Lachs explains why she keeps data scientists in one central org, how she hires for curiosity, and how she picks metrics that actually move a business.
11:29ENWhy Tesla AI Day Was the Most Impressive Real-World AI Effort Lex Fridman Has SeenModeration: Lex FridmanLex Fridman walks through the neural network, data annotation, Dojo and Tesla Bot announcements and what they say about solving real-world robotics.
1:09:51ENFrom Campus Job Board to AI Data Powerhouse: How Handshake Hit $50M ARR in Four MonthsGast: Garrett Lord · Moderation: Lenny RachitskyHandshake CEO Garrett Lord explains what expert data labeling actually involves and how a decade-old student network became a supplier to seven frontier AI labs.
45:12ENInside a Robot Data Factory: How Mecka Sensorizes Humans to Train RobotsMecka AI founder Josh Gao explains how a company collects, quality-checks and licenses egocentric human data, and why robotics will arrive as a steady unlock of tasks rather than a single breakthrough.







