For comparison, the first time a computer did this in chess was in 1996, when Deep Blue beat Garry Kasparov in the first game of a six-game match, before Kasparov turned around to win three and draw two. But the next year, Deep Blue and Kasparov played again, and this time Deep Blue won the tournament, 3.5 to 2.5. By 2006, computers could reliably beat humans at chess.
Go is a much more complex game than chess, especially from a computer's perspective, and this level of competition was thought to be a decade or more away. DeepMind's success in this is significant not just because of the milestone, but because it illustrates the tremendous power of the new foundational techniques in AI that they have been exploring. These techniques are likely to have a huge effect on how machine learning works over the next several years, and together with the major advances in neural networks in the past few years (many of which happened at Google, DeepMind's sister company in Alphabet), we should expect huge strides in the abilities of computers to observe and understand the world around them in the years to come.
Congratulations to the #AlphaGo team!