Sutton's New Lab Bets on Real-Time Learning
burny_tech · x · 2026-07-14
Richard Sutton and Khurram Javed spun off from Keen to found Oak Lab.
They argue that intelligence should primarily come from runtime experience rather than memorizing everything through massive offline pre-training. Based on the big world hypothesis, they believe the world is too vast, making it inherently inefficient to try and learn everything before training.
They are also advancing a batch size one algorithm: the model updates based on single real-time experiences instead of relying on massively curated datasets and huge pre-training compute. The team aims to build a trillion-parameter agent capable of real-time learning and planning within a 20-watt power budget, claiming it could save orders of magnitude in compute and energy compared to existing methods.
Related event: Turing Award Winner Sutton Leaves Keen to Found OakLab(10 posts)→
More from AGI Musings
- LukeW: Technology Limits Are Temporary — and AI Skeptics Are Already Feeling It — LukeW · 2026-10-27
- GiveDirectly pilots an optional AI chatbot alongside cash aid in Rwanda — NathanpmYoung · 2026-10-27
- Ex-Anthropic employee essay: the org optimizes for power 'constrained by being seen as the good guys' — RichardMCNgo · 2026-10-27
- Investor: model version numbers are a mirage, only datacenter margins are real — StewartalsopIII · 2026-10-27
- WSJ Op-Ed: For Anthropic Share Offering Investors, a 10% Extinction Chance May Be a 90% Portfolio Win — davidmanheim · 2026-10-27
- WSJ: for Anthropic investors, a 10% extinction chance may still be a 90% portfolio win — peterwildeford · 2026-10-27