AlphaGo has got nothing to do with LLMs though. It's a combination of RL + MCTS. I'm not sure where you are seeing any relevance! DeepMind also used RL for playing video games - so what?!