The nickname given to Shin Jin-seo in go circles is “Shin Artificial Intelligence.” The highest praise a human player can receive is that he plays like a machine.
This nickname first appeared in 2020, when a Korean reporter combined “Shin” and “artificial intelligence” into one phrase. Four years later, it has gone from a joke to a factual description.
On October 15, the final of the 4th Quzhou Lanke Cup World Go Open concluded in Quzhou, Zhejiang. Shin Jin-seo 9-dan defeated his compatriot Byun Sang-il 2–0 to win the Lanke Cup for the second time. This was his ninth world major championship, surpassing Ke Jie’s eight and tying Korean predecessor Cho Hun-hyun, putting him third on the all-time list, behind only Lee Chang-ho’s 17 titles and Lee Sedol’s 14.
Looked at one by one, those nine trophies stretch from the 2020 LG Cup to this year’s Lanke Cup. Over six calendar years, there has been one every year. He waited 19 months for this ninth title.
His path of growth also doesn’t quite fit the usual sports story. He learned go from his father at age five, qualified as a professional at 12, and set a Korean record as the youngest ever to do so. Since 2019, he has remained world No. 1 in rating for a long time. From 2021 to this year, he helped South Korea’s team win six straight Nongshim Cups, running up 21 consecutive cross-edition victories, 16 of them from the pressure-packed main board.
What supports these results is more than 10 hours a day spent with AI. This May at the Chunlan Cup, the overlap between one of his moves and AI’s recommended choice reached 88%. Earlier, The Paper gave a lower figure: his everyday move overlap with AI advice was 37.5%, while the average for professional players was 28.5%.
I went back through the ten-year history of human-versus-machine go and saw a rather cold line. In 2016, AlphaGo beat Lee Sedol; in 2017, it beat Ke Jie. Humans lost twice, and then stopped comparing themselves, switching to learning instead. This July, Shin Jin-seo played against the open-source AI KataGo in Seoul under handicap rules, giving him two stones, and he lost. He said he had never won in training when giving two stones, and would need to give three stones to have a chance.
The authority in go has already changed hands. In the past, to judge whether a move was good, one had to ask a top 9-dan. Now a group of players gathers around a laptop and waits for the AI to speak.
My view is that this is not just go’s fate; it is the preview trailer for all human-machine competition. Once a field has a standard answer, the value of humans shifts away from “thinking better than machines,” and Shin Jin-seo has shifted to execution. He memorizes the AI’s answers and, under the pressure of the clock, places them on the board. That does not sound romantic, but it is the only way humans are winning championships right now.
I’m not sure how long this system can hold. The agreement rate between players and AI is still climbing, and when it approaches 100%, championships become a test of memorization. At that point, whether go is still interesting will be another question.
Something is happening in the crypto market that looks almost the same. The authority in trading decisions is also shifting from humans to machines, and automated strategies are beginning to take over order execution. Decentralized AI networks talk about splitting computing power and models so anyone can plug in, while rendering networks talk about renting out idle GPUs. Both point to the same question as go: when machines calculate more accurately than you do, what do you have left?
$TAO $RNDR $BTC #币安广场 #申真谞 #围棋 #人工智能 #human-machine battle
This nickname first appeared in 2020, when a Korean reporter combined “Shin” and “artificial intelligence” into one phrase. Four years later, it has gone from a joke to a factual description.
On October 15, the final of the 4th Quzhou Lanke Cup World Go Open concluded in Quzhou, Zhejiang. Shin Jin-seo 9-dan defeated his compatriot Byun Sang-il 2–0 to win the Lanke Cup for the second time. This was his ninth world major championship, surpassing Ke Jie’s eight and tying Korean predecessor Cho Hun-hyun, putting him third on the all-time list, behind only Lee Chang-ho’s 17 titles and Lee Sedol’s 14.
Looked at one by one, those nine trophies stretch from the 2020 LG Cup to this year’s Lanke Cup. Over six calendar years, there has been one every year. He waited 19 months for this ninth title.
His path of growth also doesn’t quite fit the usual sports story. He learned go from his father at age five, qualified as a professional at 12, and set a Korean record as the youngest ever to do so. Since 2019, he has remained world No. 1 in rating for a long time. From 2021 to this year, he helped South Korea’s team win six straight Nongshim Cups, running up 21 consecutive cross-edition victories, 16 of them from the pressure-packed main board.
What supports these results is more than 10 hours a day spent with AI. This May at the Chunlan Cup, the overlap between one of his moves and AI’s recommended choice reached 88%. Earlier, The Paper gave a lower figure: his everyday move overlap with AI advice was 37.5%, while the average for professional players was 28.5%.
I went back through the ten-year history of human-versus-machine go and saw a rather cold line. In 2016, AlphaGo beat Lee Sedol; in 2017, it beat Ke Jie. Humans lost twice, and then stopped comparing themselves, switching to learning instead. This July, Shin Jin-seo played against the open-source AI KataGo in Seoul under handicap rules, giving him two stones, and he lost. He said he had never won in training when giving two stones, and would need to give three stones to have a chance.
The authority in go has already changed hands. In the past, to judge whether a move was good, one had to ask a top 9-dan. Now a group of players gathers around a laptop and waits for the AI to speak.
My view is that this is not just go’s fate; it is the preview trailer for all human-machine competition. Once a field has a standard answer, the value of humans shifts away from “thinking better than machines,” and Shin Jin-seo has shifted to execution. He memorizes the AI’s answers and, under the pressure of the clock, places them on the board. That does not sound romantic, but it is the only way humans are winning championships right now.
I’m not sure how long this system can hold. The agreement rate between players and AI is still climbing, and when it approaches 100%, championships become a test of memorization. At that point, whether go is still interesting will be another question.
Something is happening in the crypto market that looks almost the same. The authority in trading decisions is also shifting from humans to machines, and automated strategies are beginning to take over order execution. Decentralized AI networks talk about splitting computing power and models so anyone can plug in, while rendering networks talk about renting out idle GPUs. Both point to the same question as go: when machines calculate more accurately than you do, what do you have left?
$TAO $RNDR $BTC #币安广场 #申真谞 #围棋 #人工智能 #human-machine battle