The Maia engine seems promising. I've never really looked into it myself. But my understanding is that unlike AlphaZero or similar engines which use convolutio

The Maia engine seems promising. I've never really looked into it myself. But my understanding is that unlike AlphaZero or similar engines which use convolutional neural networks to predict a position's objective value and a probability distribution over legal moves and are trained by self-play which allows them to "learn" objectively best strategies, Maia's networks are trained on human moves taking rating into consideration. So humans typically don't gain much from playing traiditional engines, since they play at levels beyond mortal comprehension and make too deep moves from outer-space, Maia would be more down-to-earth. https://lichess.org/blog/X9PUixUAANCqFRSh/introducing-maia-a-human-like-neural-network-chess-engine https://lichess.org/blog/X9PUixUAANCqFRSh/introducing-maia-a...