Show HN: nanoAlphaZero – Train a grandmaster-level chess model in 24h with TPUs

By tdoubleu · 2026-08-19 · 3 points · 0 comments

Hello HN, I built a complete, game-agnostic implementation of AlphaZero in JAX. repo: https://github.com/wtedw/nanoAlphaZero demo (NN + MCTS run locally in your browser): https://nanoalphazero.wtedw.com It uses no human data, can train grandmaster-l…

Open the full discussion on BetterNews