AlphaGo defeats Lee Sedol in Seoul

node

DeepMind's AlphaGo won four of five games against Lee Sedol in Seoul between 9 and 15 March 2016; Lee won the fourth game. Go had been the standing example of a game where search alone could not substitute for judgement, and the expectation before the match was that a decade or more remained. The result matters to this Journey for a narrower reason than the one usually given: it demonstrated capability in a closed, fully observable domain with an unambiguous scoring rule and unlimited synthetic training data. Every one of those properties is absent from the domains where the consequential claims are now made, so the match is evidence about what learned search can do and almost no evidence about anything outside a board.

A node is not a place. Drawing it on a map would assert something about the world that no stored fact supports.