Mastering chess and shogi by self-play with general reinforcement learning (blog.acolyer.org)