Thoughts on "The workshop lectures" by Yilun Yang

[paK0]

Triton had some nice things to say about them and I am considering buying them, but I would like some more opinions if possible and reviews on the net seem rather scarce.

- How did you like the individual books/the series as a whole?

- Do they hold up in the "AI" world?

- What player level are they targeting?

Thanks in advance.

[sorin]

- Do they hold up in the "AI" world?

Just to point out that this is a very high expectation - I don't think any material published before AlphaGo would "hold up".

On the other hand, I don't think it matters much: one can learn tons from pros, regardless of whether AI agrees with pros or not

[MikeKyle]

I really like these books. I think they're a little bit distinctive as go books, so I expect some people will disagree.

The author is not affraid to state a state a clear principle. Whenever you try to nail something down in go, there seems to be a host of exceptions, so I think a lot of books avoid being so clear cut, but Yilun Yang seems happy to commit to his statements.

- How did you like the individual books/the series as a whole?

The series is really easy to dip in and out of. Each book is made up of 4 or 5 lectures. If you're not looking for something specifically after a review or a discussion then the lectures are still just nice bitesize chunks to read through.

I found myself getting half the series from one source and then I enjoyed it enough to seek out the other half. I recomend getting the full set if you can.

- Do they hold up in the "AI" world?

[snip..] high expectation [snip..] I don't think it matters much [snip..]

I agree absolutely with sorin.
The author gives great insight in how to think about techniques and styles. The specifics might loose a little bit of relevance as tastes change, but the fundamental guidance will almost certainly still apply.

- What player level are they targeting?

I'm not sure exactly where they're targeted. I'm European 3/4k (KGS 1k if you're more familiar with that?) and I get a lot out of them. I have players at my club at around 15k (European) who I think would get something from them. Any lower may be a stretch, but I think it's good to pick materials that stretch you.

[EdLee]

The author is not affraid to state a state a clear principle.

I can live with doubt and uncertainty and not knowing. I think it is much more interesting to live not knowing than to have answers that might be wrong. -- RPF

[Kirby]

I like them. They are not very long, but fun to read.

[RobertJasiek]

I don't think any material published before AlphaGo would "hold up".

Neural net AI does not create explicit go theory so there is only very little competition between AI and written explicit go theory. AI has provided a few new josekis, shown as much flexibility as previously the most flexible human players, deep (simulated) reading and as accurate positional judgement as previously the most accurate explicit go theory on positional judgement has described. All that previous material holds up that has not been more restricted or less accurate than AI. E.g., I have no regrets about what I have written about positional judgement, endgame or fighting but wish I had anticipated the AIs' new 3-3 under 4-4 josekis. Of what I have seen from Yang Yilun about explicit go theory in other books, AI does not restrict its validity but my more detailed work shows limitations (not: refutations, except for one example) of his.

I quickly browsed through the workshop lectures and found that they would be for single digit kyus, mostly 9k - 5k maybe to 3k. I did not buy them because they are not for advanced dan readers. If there is a problem of Yang's writing, it is not AI but being often written for SDK more than dans, i.e., for progress beyond SDK, a reader also needs to consult other sources.

[paK0]

Thanks for your comments, i have decided to order the books.

- Sdk rank fits me nicely.
- I like clearly stated principles.
- Little chunks of knowledge are preferable, as I want to focus mainly on playing as much as possible right now.