Comment by andai

>Calling the second system a “next-move predictor” would be strange. It is not trying to predict what move appeared next in a dataset. It is trying to choose a move that wins.

Sounds like the next 1000 years depend on how carefully we define "winning".