The Era of Pre-Training is Evolving
deanwball · x · 2026-07-12
The core argument is that pre-training scale was previously aimed at better fitting human language, but its primary role now is likely to provide the neural capacity for post-training capabilities like sample-efficient reasoning and long-horizon planning.
The author believes this makes "buying cutting-edge capabilities purely by scaling compute" much harder, as low-hanging web-scraped data is largely depleted. The necessary post-training data is more akin to continuously generated "learning by doing" micro-experiments from numerous diverse reward environments and user-agent OODA loops. The text notes this data is incredibly expensive, comparing its value to a "$6 billion SpaceX / Cursor acquisition." Finally, it emphasizes that beyond model distillation, it's nearly impossible to jump-start a frontier agent from scratch directly to the finish line.
More from AGI Musings
- François Fleuret: Only Two Long-Term Futures — No Super AI, or Staying Fully Human With It — francoisfleuret · 2026-09-11
- IG reel debunking the 'winning the AI race against China' fallacy hits 500k likes — louisvarge · 2026-09-11
- Post-AI World Leaves No Room for Learning on the Job — rachittshah · 2026-09-11
- Researcher questions AI safety eval firm, citing 'blatantly sloppy' security and monitoring — Kyrannio · 2026-09-11
- AI researcher memes agent-swarm tinkering with He Jiankui's embryo-editing quote — dejavucoder · 2026-09-11
- nabla_theta: happy to be wrong if the AI utopia arrives with little ex ante risk — nabla_theta · 2026-09-11