It’s absolutely stunning that AlphaZero did this well so quickly.

Oh, and it took AlphaZero only four hours to “learn” chess. Sorry humans, you had a good run.

This is another case where the algorithm just learned by playing.