Quote:
nterest in Go has been fading long before AI came to the party. ... Personally AI has had a positive impact on my interest in Go. And I can imagine that for others it has a negative one. But to gauge the impact on the whole go community, you need research, not whim.
Wow, there's a lot to unpack there
Not entirely sure what 'whim' (sudden change of mind) is supposed to mean there, but from the context I'm guessing it's meant to be a belittling term for "experience-based opinion." Let's take that first. We all know that most actions in our daily life have to be based on best guesses, that is based on past experience and opinion. It is the human condition. Sometimes we can add research to the mix, but generally that just means trusting somebody else's opinion. It is highly unlikely that any research would ever be done on AI's effect on go or chess interest, but if it were it would almost certainly start with some assumptions (quite possibly false) and then be a sampling of opinions of different people. If that's correct, I could argue that what I have done, by reading widely and asking for others' opinions is in fact 'research'. In the course of that 'research' I have come to the conclusion that opinion on L19 (with its top-heavy dose of mathematicians and programmers) is heavily skewed in favour of AI but opinion in the pro-based countries seems to tilt strongly the other way.
For those who have to fork out money, i.e. sponsors, I would suspect that they too would rely on experienced-based opinion, though I also suspect they would call it expert opinion. But it's opinion nonetheless.
As to the earlier fading of interest in go, I would not dispute that, but it's irrelevant here. There's no cause and effect. In fact it's a subtle form of the coin-toss fallacy. Thinking six heads in a row means a head is more likely on toss seven. Or here, thinking that continual fading interest in go means that more fading must be due to what went before, and not to the introduction of a new element (AI).
But the biggest issue is comparing, in the current context, the positive and negative impacts. I think we can agree they both exist, but that does not make them comparable. A positive effect on making go/chess more interesting for an existing player leaves the go/chess population X at value X. But a negative effect on an amateur go player is very likely to lead to a population of X - 1.
To make a genuine comparison, one would have to cite a positive case where a non-player is excited enough by AI to take up the game anew, i.e. X + 1. I'm pretty sure such people may exist but I've only come across one and he was more interested in the programming and I've come across many of the X - 1 persuasion.
Go and chess are susceptible to so many influences that interest in either becomes a Marmite issue. For example, in chess many people are put off the game by draws. Many are also put off by having to memorise reams of opening lines. In go, many people are apparently put off by what they regard as unclear rules. When I started go, many people were put off by thinking it was a Japanese game (post-war anti-Japanese sentiment was still strong then). Given the impact that issues like that (and there are more) have had on both games, I don't see why it should surprise anyone that a huge issue like AI can have both predictable and unpredictable consequences on different individuals.
It's not entirely germane, but I think the following quotation from an article I read just an hour or so ago is worth mulling over. It is from a report on the world chess championship Game 2:
"The difficulty posed for Nepomniachtchi was that he was no longer "choosing off a menu" but rather choosing moves blindly against a 3600-rated engine that was behind Ding's at-home preparation."
Depending on whether you are a bottle half-full or bottle half-empty man you could highlight the fact that the human bested the computer preparation, or you could say (as I do) that preparing at home with a team of seconds plus computers is not much different from an athlete taking drugs. But, however you look at this bottle, I honestly can't see that it is of any interest to non-chess players. And given Carlsen's opting out, presumably such chess is not of much interest to people like him.
I think the fundamental difference in this debate is actually nothing to do with logic. I think the real difference is between amateur players who want to be stronger (i.e. they are more interested in themselves and believe AI will help them) and those who are more interested in other people, i.e. the pros. There is no obvious reason to think that either group is intrinsically better than the other, but the latter group is, I would imagine, of far more interest to pros and their sponsors, and so to that extent that group is ultimately going to be more important for the future of the game. After all, where would pro soccer be without the armchair fans or the fans on the terraces, or the merchandise-buying classes, and instead they were no fans but just amateurs who play in the park at the weekend?