Are we getting better?

[kvasir]

kvasir: "Analysis has also become much more objective.": Yes, where application of mathematical theorems allow objectivity.

That is a very narrow view if you use the normal meaning of 'objectivity' but it could be reasonable if you are thinking in terms of philosophy and philosophy of science. Not sure if it can be considered reasonable even in philosophy of science. What I mean is if I analyze a joseki, you analyze the same joskei and a third person does the same. Now we all agree, does it really matter if it is ultimately only our opinion. Obviously it is possible to be very thorough, even play multiple trial games, ask others opinions and so forth; there are degrees of objectivity. It may just be the human experience that we can't fathom everything, but still we take what we can have.

"With the new AI programs it is much easier to analyze objectively.": No, because AI can make mistakes. Analysis with AI need not be objective. The advantage is rather that using AI allows to find more blunders, which human beings might have overlooked.

It is true that the programs do make mistakes and sometimes the point estimates of errors is suspiciously close to counting the immediate point difference of two moves as if nothing else matters. I think I have also seen a few cases were KataGo was off by a 0.5 point at the end of the game. You can still use the programs to speed up your analysis, the program corrects most errors quickly and you can play out sequences and use basic tewari principles to test for impossible results. Though, you are right in that the main usefulness is in quickly finding blunders (and also small errors). It would take hours to analyze games in the same detail that I can do in couple of minutes with my new laptop and it would take months to analyze a single game in the same detail that my new laptop can do in 20-30 minutes and one wouldn't really be able to do either without making many mistakes. There just wouldn't be any time to learn from it. Such tools are much more useful than I thought they would be and I have really only recently been realizing what this could do for my game, if I can really learn so much because the learning tools might out pace the student now.

Concerning my own experience as a 5 dan since 1998, my knowledge has increased dramatically since then and the 5 dan I was in 1998 would have no chance against the 5 dan I am today. So I think European dan play must have improved significantly since then. I just cannot say exactly how many ranks in terms of 1998 ranks. Might be 0.5 or 1.5 ranks - I do not know which. When comparing the skills of 3 dan opponents then and now, I think it is closer to 1.5 ranks overall improvement. When comparing the skills of 5 dan opponents, judgement is much harder because I might be prejudiced too much. When seeing the improvement of top Europeans, 1.5 ranks is realistic but does that boil down 1:1 to 6d and 5d players, too? I am unsure. At least 0.5 ranks for sure, but it could be more.

I would not suspect the same level of improvement since 1998 of top Asian professionals for the reasons further above though.

I have heard many similar antidotes. Norway comes to mind because with few new dan players through out the years everyone that improves ends up robbing rating points from their friend, later their friend improves and takes back the rating points and everyone is exactly where they started.

What I find interesting is that we have rating deflation but games like chess have rating inflation.

[TOTAL]

An interesting commentary by one of the co-creators of AlphaGo & Zero: https://www.youtube.com/watch?v=WXHFqTvfFSw

I was unaware Zero did not actually play with any top-ranked human at least officially. If people do not really have access to zero, my original question loses validity; it was the tabula rasa, self-reinforcing approach that I think would bring human playing to new levels.

At the same time, it might disrupt the culture, steal the spirit of the game.
If it finds out some part of the playing heritage less relevant than its totally synthetic creations, I can easily imagine how this could broaden the human horizons on the one hand and make top humans retire, as was the case with Lee Sedol.

[gowan]

It might be that some players have given up the game after it became clear that AI can be stronger than any current human player. I've said elsewhere that whether humans can't defeat AI players is really irrelevant. Human runners can't run a mile is less than three minutes and AI controlled self-driving cars would be able to go one mile in two minutes. Does this somehow disparage human capability?

[TOTAL]

I agree with the running analogy to a point. If it should turn out that the AI breakthroughs to come might turn the knowledge of go upside-down, as already hinted at with reference to 'some' classic moves and strategies in that video,
1) new generations of go geniuses might dominate the human go world with new, synthetic strategies. Note that on several occasions when AG played totally out of the blue but successfully, Lee Sedol was disconcerted as he lost the capacity to draw from his arsenal of strategies. It was broadly commented that he was not playing 'his game' (I forget the original wording). What mattered was that the best human player on Earth seemed unprepared for this type of eventuality - he did not understand what was happening.
2) if the breakthroughs to come are numerous and huge - and I am aware this is a condition, but I think we all agree, go is way deeper than what has been explored - it might be that those top talents who have a 'subscription' to the newest engines will have an advantage comparable, to get back to your metaphor, to sprinters having high-tech leg implants.

[RobertJasiek]

If it should turn out that the AI breakthroughs to come might turn the knowledge of go upside-down

Maybe partially but not all the knowledge of go theory! There is go theory established as mathematical theorems, which already state the absolute truth within the, sometimes powerful, scope of their presuppositions and can, in principle, never be improved. See e.g. http://home.snafu.de/jasiek/GoTheoryResearch.html