A new kind of AI

Old but very cool news (from late 2017): In Just 4 Hours, Google's AI Mastered All The Chess Knowledge in History.

“I always wondered how it would be if a superior species landed on Earth and showed us how they played chess,” grandmaster Peter Heine Nielsen told the BBC. “Now I know.”

But this news didn’t make as big a splash as it should have, I think. The important thing to understand here is that this is not the same kind of AI that has beaten human chess players in the past. AI has won at chess in the past by brute force, by the computational capacity to churn through all the possibilities. It’s closer to a party trick than real intelligence. For decades, all we’ve seen from chess-playing AI is more of the same, more brute force.

This is something new. This time AI won at chess by “learning” it, basically from scratch, better than anyone or any AI has ever learned chess before… in a few hours.

And that’s going to change the world as it gets applied to other challenges.