This is true too, we are adapted to our environment, in particular the things that we do automatically (System I). Playing chess well is a combination of both

This is true too, we are adapted to our environment, in particular the things that we do automatically (System I). Playing chess well is a combination of both conscious and unconscious skills. However when deep learning systems play, it is all the unconscious, automatic application of statistical rules. They are playing a very different game from the human chess game. Because there is no abstract reasoning involved here, these systems cannot apply the lessons learned from chess to another board game, or to something completely different in life, which humans can and do. So even though they are much stronger than human players, they aren't strong in the same way.