Move 37
In the second game AlphaGo played a move that commentators first took to be a mistake. It later turned out to be brilliant. For many people it was the moment a machine showed creativity.
A human victory
Lee Sedol won the fourth game with an exceptional move 78. It is the only game a human has ever won against this version of the program. In 2019 he retired, saying that artificial intelligence cannot be beaten.
What came next
AlphaGo Zero learned to play only against itself, without human games, and beat the older program 100:0. The same approach later helped with AlphaFold, see Deep learning and neural networks.