"LLMs can't think" is the unscientific claim, argues Plinz in debate over burden of proof

AlexTensor · x · 2026-09-26

A debate on whether LLMs can think. Plinz argues that claiming "LLMs cannot think" is itself unscientific: since LLMs reproduce outputs (scene understanding, translation, mathematical proofs, theory of mind) typically seen as requiring thought, proving a fundamental difference would require a formally rigorous definition of thinking and an argument about non-equivalence of causal structure.

Quoted motorhueso counters that the burden of proof lies on those claiming algorithms can think, since scientific understanding of thought has for centuries been tied to nervous systems, bodies and societies.

Related event: Plinz Argues Claiming LLMs Cannot Think Is Unscientific(2 posts)→

Original post →

More from AGI Musings

AGI Musings channel →