Remix.run Logo
▲ zesterer 7 hours ago

The real innovation here, IMO, is theorem provers. Without theorem provers, the LLM brute-force-by-bouncing-off-the-guardrails approach just doesn't work to begin with and the whole thing spins off into hallucination and dodgy reasoning, just as we see in non-trivial AI software. Once you've got some magic oracle that gives you a binary yes/no answer to every attempt at a proof, your search strategy can be enormously simplified.

▲j2kun 6 hours ago | parent | next [-]

I also thought that, but I learned that most of OpenAI's results in their proof dump did not use theorem provers in the loop. They translated the papers to lean afterward.

▲supuun 2 hours ago | parent | next [-]

If i understand it correctly, lean is very useful for post-training (RL) step of the models.

▲zone411 6 hours ago | parent | prev [-]

Right, that's how all these proofs (including a few of mine) are done. Given the large-scale AI efforts already underway to formalize proofs in Lean, there could be some very interesting gains once proofs are Lean-first. For now, though, formalization happens afterward.

▲famouswaffles 5 hours ago | parent [-]

I don't expect proofs are going to be lean first. There's a reason they already aren't.

▲jojomodding 6 hours ago | parent | prev [-]

It's not really the provers that are the innovation. They existed before AI. The innovation is that the computer can write the proofs.