Article - AI is rewiring how the world’s best Go players think

Discussion of anything and everything relating to chess playing software and machines.

Moderator: Ras

smatovic
Posts: 3598
Joined: Wed Mar 10, 2010 10:18 pm
Location: Hamburg, Germany
Full name: Srdja Matovic

Article - AI is rewiring how the world’s best Go players think

Post by smatovic »

AI is rewiring how the world’s best Go players think
https://www.technologyreview.com/2026/0 ... ers-think/

..with some words by retired Lee Sedol.

--
Srdja
chetday
Posts: 171
Joined: Sun Nov 18, 2012 9:18 pm
Location: North Carolina
Full name: Chet Day

Re: Article - AI is rewiring how the world’s best Go players think

Post by chetday »

Very interesting. Thanks for sharing the link, which led me to a different article, this one from back in March of last year on chess and Alpha Zero: https://www.pnas.org/doi/10.1073/pnas.2406675122
User avatar
towforce
Posts: 12879
Joined: Thu Mar 09, 2006 12:57 am
Location: Birmingham UK
Full name: Graham Laight

Re: Article - AI is rewiring how the world’s best Go players think

Post by towforce »

smatovic wrote: Fri Feb 27, 2026 3:18 pm AI is rewiring how the world’s best Go players think
https://www.technologyreview.com/2026/0 ... ers-think/

Kudos to the author, Michelle Kim: the article is a treasure trove of wisdom!

“Go has become a mind sport,” says Lee Sedol, who retired three years after his 2016 defeat to AlphaGo. “Before AI, we sought something greater. I learned Go as an art,” he says. “But if you copy your moves from an answer key, that’s no longer art.”
...

Even top players like Kim and Shin don’t understand all of AI’s moves. “It seems like it’s thinking in a higher dimension,” she says. When she tries to learn from AI, she adds, “it’s less about rationally thinking through each move, but more about developing a gut feeling—an intuition.”
...

In 2024, researchers at Google DeepMind extracted new chess concepts from AlphaZero, a generalized version of AlphaGo Zero that can also play chess, and taught them to chess grandmasters using chess puzzles.
...

“Top-tier players haven’t yet been able to deduce the general principles behind AI moves,” says Nam Chi-hyung, a Go professor at Myongji University. [Graham's interjection: today's NNs are finding a large number of surface level patterns rather than the deep underlying ones]
...

Even if AI is an opaque teacher, it’s a democratic one. It has supercharged training for female Go players, who have long been underdogs of the game... ...More broadly, AI has narrowed the gap between players by helping everyone perfect their opening moves.
...

Analyzing male players’ ["experts"] moves with AI has shattered their veneer of infallibility... ..."Now, I know that they make mistakes, and their moves aren’t always brilliant"
...

“I may be one of the strongest human players, but with AI around, I can’t be so arrogant,” he says. “AI gives me a reason to keep improving.”
Human chess is partly about tactics and strategy, but mostly about memory
ernest
Posts: 2056
Joined: Wed Mar 08, 2006 8:30 pm

Re: Article - AI is rewiring how the world’s best Go players think

Post by ernest »

smatovic wrote: Fri Feb 27, 2026 3:18 pm AI is rewiring how the world’s best Go players think
https://www.technologyreview.com/2026/0 ... ers-think/

..with some words by retired Lee Sedol.

--
Srdja
The link asks for subscription to be read... :cry:
smatovic
Posts: 3598
Joined: Wed Mar 10, 2010 10:18 pm
Location: Hamburg, Germany
Full name: Srdja Matovic

Re: Article - AI is rewiring how the world’s best Go players think

Post by smatovic »

ernest wrote: Tue Mar 03, 2026 4:47 am
smatovic wrote: Fri Feb 27, 2026 3:18 pm AI is rewiring how the world’s best Go players think
https://www.technologyreview.com/2026/0 ... ers-think/

..with some words by retired Lee Sedol.

--
Srdja
The link asks for subscription to be read... :cry:
Here, from Germany with Firefox, subscription boxes pop up but I can X them away, maybe try in browser reader-mode with Ctrl+Alt+R.

--
Srdja
User avatar
Eelco de Groot
Posts: 4701
Joined: Sun Mar 12, 2006 2:40 am
Full name:   Eelco de Groot

Re: Article - AI is rewiring how the world’s best Go players think

Post by Eelco de Groot »

Same here, can click them away, it is a free first story. Usually last resort is reading the http source, but please keep it secret. Restarting the computer maybe works. Or just coming back later.

Here is the translated page for Ernest if he can read Dutch maybe this is not blocked right now: https://www-technologyreview-com.transl ... tr_pto=sdn
Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan
ernest
Posts: 2056
Joined: Wed Mar 08, 2006 8:30 pm

Re: Article - AI is rewiring how the world’s best Go players think

Post by ernest »

Eelco de Groot wrote: Tue Mar 03, 2026 8:12 am Same here, can click them away, it is a free first story. Usually last resort is reading the http source, but please keep it secret. Restarting the computer maybe works. Or just coming back later.

Here is the translated page for Ernest if he can read Dutch maybe this is not blocked right now: https://www-technologyreview-com.transl ... tr_pto=sdn
Thanks Eelco !!!
I succeeded with your translation link, using the same Google Translate, but --> Engels (English)
Uri Blass
Posts: 11164
Joined: Thu Mar 09, 2006 12:37 am
Location: Tel-Aviv Israel

Re: Article - AI is rewiring how the world’s best Go players think

Post by Uri Blass »

towforce wrote: Fri Feb 27, 2026 6:24 pm
smatovic wrote: Fri Feb 27, 2026 3:18 pm AI is rewiring how the world’s best Go players think
https://www.technologyreview.com/2026/0 ... ers-think/

Kudos to the author, Michelle Kim: the article is a treasure trove of wisdom!

“Go has become a mind sport,” says Lee Sedol, who retired three years after his 2016 defeat to AlphaGo. “Before AI, we sought something greater. I learned Go as an art,” he says. “But if you copy your moves from an answer key, that’s no longer art.”
...

Even top players like Kim and Shin don’t understand all of AI’s moves. “It seems like it’s thinking in a higher dimension,” she says. When she tries to learn from AI, she adds, “it’s less about rationally thinking through each move, but more about developing a gut feeling—an intuition.”
...

In 2024, researchers at Google DeepMind extracted new chess concepts from AlphaZero, a generalized version of AlphaGo Zero that can also play chess, and taught them to chess grandmasters using chess puzzles.
...

“Top-tier players haven’t yet been able to deduce the general principles behind AI moves,” says Nam Chi-hyung, a Go professor at Myongji University. [Graham's interjection: today's NNs are finding a large number of surface level patterns rather than the deep underlying ones]
...

Even if AI is an opaque teacher, it’s a democratic one. It has supercharged training for female Go players, who have long been underdogs of the game... ...More broadly, AI has narrowed the gap between players by helping everyone perfect their opening moves.
...

Analyzing male players’ ["experts"] moves with AI has shattered their veneer of infallibility... ..."Now, I know that they make mistakes, and their moves aren’t always brilliant"
...

“I may be one of the strongest human players, but with AI around, I can’t be so arrogant,” he says. “AI gives me a reason to keep improving.”
1)"AI has narrowed the gap between players by helping everyone perfect their opening moves."

It is correct if we assume all players have the same memory but I think it is logical to think that top players have a better memory and can memorize more opening moves relative to weaker players with the same training time and in this case top players can still get an advantage from the opening against weaker players.

2)I wonder what is the reason for difference between chess and go because in chess we do not see a reduction of the gap between males and females.

I wonder if go engines helped humans to improve relatively more than chess engines.
In this case what is the reasons:

possible reasons(only a guess and do not know if a correct guess):
a)Playing correctly the opening is more important in go relative to chess.
b)There are too many equal moves in chess and it is not the case in go so in go it is easier to study opening 20 moves forward relative to chess.
User avatar
towforce
Posts: 12879
Joined: Thu Mar 09, 2006 12:57 am
Location: Birmingham UK
Full name: Graham Laight

Re: Article - AI is rewiring how the world’s best Go players think

Post by towforce »

Uri Blass wrote: Thu Mar 05, 2026 4:35 am2)I wonder what is the reason for difference between chess and go because in chess we do not see a reduction of the gap between males and females.
With apologies to the global Go community if I'm wrong, but my reading of the article is that the game of Go was less accessible to women than it was to men in some countries in which Go is a big game - and especially at the top level, where the teaching methodology was for the student to live in the house of the master.

I wonder if go engines helped humans to improve relatively more than chess engines. In this case what is the reasons:
The most important thing about Go is that it's a more complex game than chess: in most positions, there are a lot more choices of move than there are in chess. This is what makes it a less "computer friendly" game than chess in the first place: the game tree grows too large too quickly.
Human chess is partly about tactics and strategy, but mostly about memory
Uri Blass
Posts: 11164
Joined: Thu Mar 09, 2006 12:37 am
Location: Tel-Aviv Israel

Re: Article - AI is rewiring how the world’s best Go players think

Post by Uri Blass »

towforce wrote: Thu Mar 05, 2026 10:22 am
Uri Blass wrote: Thu Mar 05, 2026 4:35 am2)I wonder what is the reason for difference between chess and go because in chess we do not see a reduction of the gap between males and females.
With apologies to the global Go community if I'm wrong, but my reading of the article is that the game of Go was less accessible to women than it was to men in some countries in which Go is a big game - and especially at the top level, where the teaching methodology was for the student to live in the house of the master.

I wonder if go engines helped humans to improve relatively more than chess engines. In this case what is the reasons:
The most important thing about Go is that it's a more complex game than chess: in most positions, there are a lot more choices of move than there are in chess. This is what makes it a less "computer friendly" game than chess in the first place: the game tree grows too large too quickly.
I know that go is more complex relative to chess.

I do not know that go is less computer friendly game than chess.
It may be the opposite and because go is so complex humans simply have no idea how to play it when it is different in chess.

I understand that In go humans had no idea what is the best move in the opening and maybe they became 200 elo stronger thanks to learning opening moves from engines(I do not know)