Ethan Mollick: RL still counts as a win for human language
emollick · x · 2026-11-21
In a exchange with @JacksonKernion, Ethan Mollick argues that RL works because it starts from a model that learned to reason from human text, and the reasoning it reinforces is still in words — so he still rates it a win for human language. Kernion counters that RL is more centrally important to progress in math.
Related event: Mollick: Next-Token Prediction's Success Is a Win for the Humanities(4 posts)→
More from AGI Musings
- Gary Marcus mocks 'AI psychosis' hype over AI-solved math problems with satirical test — GaryMarcus · 2026-11-21
- NYU/UT Austin study: all 9 frontier models write text harder to process than human writing — windx0303 · 2026-11-21
- 250,000-page Riemann proof is more like an iPhone than magic: the AI math comprehension debate — nptacek · 2026-11-21
- After AI Cracked Navier–Stokes, Millennium Problems Fell Fast—Why Did Tao Miss It? — burny_tech · 2026-11-21
- Neuroscientists warn top journals are full of 'polished garbage' systems neuroscience papers — KordingLab · 2026-11-21
- Forecasters put 85% odds on AI solving another Millennium Problem by 2027 — NathanpmYoung · 2026-11-21