I’m following the match-up between the world’s best player of the game of go, and a computer from DeepMind*, an artificial intelligence software house in London that was bought by Google in 2014. Go is played on a 19×19 grid of vertical and horizontal lines with black and white checkers. The number of games that can be played on it is enormous: The Economist’s article says a rough-and-ready guess gives around 10170. (Keep in mind there are only an estimated 1080 particles in the observable universe). Anyway : it’s 2-0 for the computer so far, but humans need not despair. General-purpose machine intelligence remains a long way off, says the article.
*Not to be confused with Deep Blue, the chess machine that beat Gary Kasparov in 1997.
Update : Sunday 3/13. I see it’s 3-1 for Deep Mind’s AlphaGo program. The South Korean Go master Lee Sidol won a game against the machine, at least denying it a clean sweep in the 5-game match-up.