Gary Marcus

Rank 12 of 47

Score 83

@burkov yep i basically agree with your first sentence.

not totally sure i follow the second. even if they sample zillions of games, they still screw up. but you may have something more nuanced in mind.

5/19/2025, 5:14:33 PM

In reply to:

Andriy Burkov

@burkov

252d

@GaryMarcus Not making illegal moves cannot be enforced in an LLM. You need to do it in the token-sampling code that doesn't sample tokens if they would make the move illegal.

Gary Marcus

@GaryMarcus

252d

Has anyone found an LLM that can reliably play chess, without making illegal moves?

The statement is part of a technical discussion about the capabilities of language models in playing chess, specifically addressing the issue of making illegal moves. It engages in a dialogue about the limitations and potential solutions for improving LLMs in this context.

Principle 4:
I will engage in constructive criticism and dialogue with those in disagreement and will not engage in personal attacks or ad hominem arguments.
The statement engages in constructive dialogue by acknowledging agreement with part of the previous statement and expressing uncertainty about another part without resorting to personal attacks or dismissive language. [+1]