I used to play Go at university and after, but rather dropped out when I had children. Recently I’ve started playing online again, though its a somewhat inhuman way of playing. One of the themes of those times was that computers were rubbish at Go, in contrast their chess performance (one of the problems is that its pretty hard to evaluate a go position, particularly at the start, by any kind of counting. In chess (I think) you can count pieces and locations pretty quickly). So of course we Go players took that as clear proof that Go was a more interesting game :-)
The Economist has a little article updating this a bit: here asserting that “Monte Carlo” methods enable progams to play well on small boards. It still sounds rather like brute force to me, but (guessing from the story) rather than doing all branches of the tree to a certain depth you randomly prune the tree but follow some branches to the end? Full 19×19 boards should still be safe for a while, though.