I've heard this a few times. In 2014 I was doing an MSc in Intelligent Systems ("AI" after the Winter) and Go was discussed in class in the context of Russel and Norvig. I don't remember the tutor saying that beating a grandmaster (? do they have grandmasters in Go?) was "10 years away". I remember him saying that Go was the last of the classic board games where humans still dominated machines because it requires intuition.
So, can you say where the "10 years away" quote comes from? Is it an actual quote? Do you know someone who actually said beating [a top human player] in Go is "10 years away" at some time before 2015?
>In early 2014, Coulom's Go-playing program, Crazystone, challenged grandmaster Norimoto Yoda at a tournament in Japan. And it won. But the win came with caveat: the machine had a four-move head start, a significant advantage. At the time, Coulom predicted that it would be another 10 years before machines beat the best players without a head start.
I'm sure there's more to find, but of course google now biases towards the articles about AlphaGo actually winning.
More to the point, when you commented above that "The answer was "10 years away" for decades, right up to 2015", did you have the Wired article in mind? I mean to say, did you read that article in 2014 and think that machines dominating Go is still 10 years away or is it more something you found with a search yesterday? Did you think in 2014 that machines dominating Go was 10 years away?
What I really want to know is what this "prediction" means. Was there really some kind of consensus on "10 years"? How seriously was this taken? It's all so vague and anyone could have said anything and meant anything.
But, ok, seriously, yes I am an armchair futurist, and also used to play go with my uncle (I am terrible at it), so since chess fell I've been waiting for go to fall too.
I am a believer in exceptional people and teams entering a space and turning established thought on its head. Another example would be spaceX. If you'd have said, ten years ago, that we'd have reusable rockets, that land on their tails like a 1950s sci-fi movie, you'd have been laughed at. On a boat! Ha ha! Crazy! Boeing still don't believe it, given what they've just managed to roll out. I've personally built my career on doing what others have said is impossible.
So, seriously then, sure, if enough people have been saying that something is still 10 years away, then other people have been listening, noticing the opportunity, working on it quietly, and a solution could happen later today.
My perspective on this is that Go (I don't know anything about rockets), was a technological advance that should not have been surprising for anyone who knew the state of the art. One reason I don't know how to interpret Coulom's quote is that he must have been one of the few people who understood computer go as well.
You'd say, yeah, that's the point, but if you look at the AlphaGo paper it really has nothing new in it. There is no sudden breakthrough in scientific understanding, in algorithm design, in efficiency or accuracy of a search, or anything of the sort. All there is, is Monte Carlo Tree Search and neural nets. That's as old as bread and about as exciting.
The only reason why it was AlphaGo that dominated Go, and not some other system, is that it was created by a company with the resources of DeepMind, meaning Google. It had to be such a big company exactly because there was no real advance in the scientific understanding of the game of Go, and so progress could only be made by turning up the compute to 11.
The same thing happenned back in the 1980's with DeepBlue and chess, also: it had to be IBM because few others could make a minimax system that searched so far and deep, because few others had IBM's computing resources.
And you know, perhaps that's what Coulom meant when he said "10 years". That it would take 10 years for the scientific understanding of Go to advance to the point where a machine can beat a human. It just so happenned that turning on the firehose and spamming the dollars at the problem made it go away, so there's no need to understand anything anymore.
Same thing happenned with chess, also. Alan Turing, Claude Shannon, John McCarthy, all those AI greats, they thought that chess was a measure of human intelligence and they wanted to understand how humans play chess using our human intelligence. Well, it turned out that you don't need human intelligence to play chess, as you don't need it to play Go, and a dumb machine with a blind search can look deeper into a game tree and find better moves than a human. But as far as I can tell, predictions of "10 years" for chess or for go or for anything else, were trying to predict when the scientific understanding of those games and of human and artificial intelligence would have advanced far enough that we could make an intelligent machine capable of playing go like a human, but better. That's what McCarthy would have wanted- he once made a bet with a Grandmaster that "in 10 years" a machine would beat the grandmaster in chess; and lost the bet. Because McCarthy was thiking about understaning intelligence, not winning games. And he underestimated the difficult of the former and overestimated that of the latter. Coulom, in the Wired quote, probably made the same mistake, thinking along the same lines.
If so, then he wasn't wrong, because while we have machines that can play better Go than humans, we still don't understand how humans played better Go than machines for so long.
