Ran into this article and its fantastic title today, though the news is a few days old. Computer Go programs have recently notched up wins against professional players, something that was inconceivable just a couple of years ago.

It's important to keep in mind that the wins in question are handicap games played against lower-tier professionals (something like, say, Irina Krush giving knight odds and losing). However, this is notable because a) computer Go has been so poor that mediocre players routinely beat bots, and more importantly b) some Go players have looked down on chess and argued that the fact that their game remains unconquered is a sign of its superiority.

The approach the programmers are turning to, Monte Carlo analysis, is also very interesting. Chess is by contrast easy to program, since writing a function to evaluate a position is if not simple then at least straightforward: you have material balance, static advantages/disadvantages such as isolated or doubled pawns, dynamic issues such as space control, etc. In Go, with its abstract notions of "influence" or "thickness," this has been much more difficult to model.

Anyway, a good read.

I'm not in position to judge (I'm playing one game per with SmartGo and is kicking my ass big time in even games), but the Go programmers will have to go to great lengths to produce a software that could a pro in an even match. Unlike chess, go comes with strategy and tactics in equal parts. And if tactics are hard to grasp, go strategy will seem completely alien to a chess player.