For the first time, a supercomputer beat a Go master.

 OK, there are a few caveats. It was a 9x9 game, and Go is more commonly (and challengingly) played on a 19x19 board. The computer decisively lost against the same Go master on that size of board. And it was just the first officially sanctioned such victory, and only the first of those if you restrict it to 'non-blitz' performances. 

Still, it was a breakthrough in AI programming. The (originally) Chinese game of Go is played on a simple grid with undifferentiated black and white stones and a set of rules much simpler than those of chess, yet it is judged to be a much more complex game than chess.

It's been more than a decade since  IBM's Deep Blue beat chess grandmaster and world champion Garry Kasparov, but Go mastery has so far remained a human skill.

Here are a summary of the match, some information on the methodology,  and an IEEE Spectrum article on cracking Go.