François Fleuret challenges LLM skeptics: what does AlphaZero say about the interpolation argument?
francoisfleuret · x · 2026-10-09
Prominent AI researcher François Fleuret poses a pointed question on X: to those convinced that 'LLMs cannot invent stuff, they interpolate, they do not extrapolate, they cannot be creative' — what is your perspective on AlphaZero? AlphaZero famously surpassed human knowledge through self-play and is often cited as proof that superhuman creativity can emerge without mere interpolation; Fleuret uses it to challenge the popular claim about LLM limits. The question drew substantial discussion.
More from AGI Musings
- Biotech Founder: Rejecting Open Source Is Demanding a Faster Relay Race While Refusing to Pass the Baton — Ghost_Pilot_MD · 2026-10-09
- Elliot Glazer slams AGI-bet critics who refuse wagers and never admit being wrong — NathanpmYoung · 2026-10-09
- Daron Acemoglu asks whether AI can create as many jobs as it will replace — mariotelfig · 2026-10-09
- Nick Bostrom: today's AI may already be conscious, Microsoft's denial lacks evidence — coherence · 2026-10-09
- Consciousness debate: provocative self-model definition fails the testability test — esrtweet · 2026-10-09
- India's AI predicament: students as RLHF workers, 15 GW gap by 2030 — abhiadesai · 2026-10-09