[thfuran]

Alpha go and stockfish were specifically designed and trained to win at those games.

[Davidzheng]

And we can train models specifically at math proofs? I think only difference is that math is bigger....