7 votes

New paper from DeepMind and world champion Vladimir Kramnik uses the AlphaZero self-learning chess engine to explore nine variants on the rules of chess