John Schulman: distillation is the main force fighting AI centralization

himanshustwts · x · 2026-09-12

Dwarkesh Patel released a new episode with John Schulman, Claire O'Neill, and Beren Millidge — researchers at "openish" companies — on what's actually happening at the frontier. Schulman's key claim: anything learnable via RL can be distilled easily because it's a small number of bits, making distillation the main counterweight to centralizing forces in AI.

Topics include: steelmanning the case against RSI, what's driving Chinese labs' progress, how automated AI researchers will be trained, whether long-horizon RL elicits AGI, the sim-to-real gap, how much progress data explains, why RL works so well, Move 37 and entropy collapse, and rapid-fire timelines.

Original post →

More from AGI Musings

AGI Musings channel →