It Seems AlphaGo AI Is Going to Beat Every Human at Every Game of Go
Tonight is the fourth of five games in a Go match between man and machine.
Go is an ancient Chinese game that requires the player use more than moves, they must use instinct. That’s why it was a wild thing when a British company DeepMind created an AI that can play, and apparently, beat humans at Go. Google currently owns the program. This week marks the second competition in 2016 in which the computer will go up against a human opponent. And three days deep into the five game match, it’s looking like humans are in trouble. A lot of trouble.
From his spot at the Four Seasons Hotel in Seoul, South Korea, 18-time world champion Lee Se-dol lost all three games as of late last night. Tonight’s match starts at 11 p.m. and it’s likely to be a four-hour massacre given that last night’s game required the computer to really think hard and human, and it did.
DeepMind’s founder, Demis Hassabis, said he was “a bit stunned and speechless” at the outcome. Lee feels a little like a disappointment right now. The Verge reports that he “believes that he had no chance in the first game, missed opportunities in the second, and succumbed to pressure today.” Details on the first and second games verify that he’s never going to win.
This past January, AlphaGo whopped European champ Fan Hui in each of the five games they played. SpaceX and Tesla’s Elon Musk, who is notoriously wary of AI, even tweeted about the impressive AlphaGo achievements, “Experts in the field thought AI was 10 years away from achieving this.”
So scary, it’s good, because AlphaGo is so good it’s scary.
Tuesday, March 15, Korea time, or March 14 in the States, will mark the final game of man mind v. machine mind. Watch the fourth game tonight live at 11 p.m. EST here.
