From LLM experiments to a 4000+ ELO engine
Hey everyone!
Recently, I posted about the release of my chess game where you play against AI. Back then, I received quite a bit of "feedback" regarding the use of an LLM as the engine. While that approach was purely experimental—to see how well an LLM could "play" by understanding the context of the board and history—people kept asking about its ELO rating. Honestly, what ELO can you even give an LLM? :)
Well, I’ve got an update. I’ve now integrated Stockfish 18, a name well-known to the chess community. I haven't changed the game's title yet, but you can now test your skills against a real engine with an ELO of 4000+.
Here’s the link for those interested:
Chess AI
I’m more than happy to hear your thoughts and suggestions to improve the game. I’d appreciate any comments, and I’m especially looking forward to your constructive criticism!