[zornthewise]

Sure, the best chess/go players on this planet are not humans anymore and those games are full of novel situations with no known solutions.

[OJFord]

That seems pretty different though, they have relatively few relatively straightforward fixed rules, you can just compute possible outcomes for moves.

[ben_w]

The branching factor of Go prevents brute force calculation of the best possible moves. It isn't even obvious how to write a function that evaluates how strong any given board position is. The AI had to learn functions to do that.

[worrycue]

Stock Fish and AlphaGo aren’t LLMs - which is the type of AI ChatGPT is.

[anonylizard]

Deepmind is trying to combine "Alphazero" style algorithms with LLMs to build Gemini.

[worrycue]

We will see if they succeed and create a better AI that doesn’t relies solely on memorising answers.