I think that the context matters. While his actual statement is now false, he was really talking about a "solution" to Go, i.e. an algorithm that can compete with any opponent (and back then, Go programs couldn't "even" beat humans). Google's algorithm is (probably) nowhere near a "solution" to Go, but an algorithm that can beat currently-living human opponents. I.e. it is quite likely that a rather simple algorithm would still beat Google's program, only that people's minds don't (or can't) employ that algorithm when they play.