‘Shin-telligence’ beats an actual AI in two-stone handicap Go match

작성일 07-20

World No. 1 Go player Shin Jin-seo, whose style of play is likened to artificial intelligence, defeated the elite AI KataGo Sunday after losing an earlier match.<div class="image"> <img src="https://imgnews.pstatic.net/image/640/2026/07/20/0000091448_001_20260720164216266.jpg" alt="" />Go champion Shin Jin-seo plays against the Go-playing AI KataGo at Korea Economic TV's studio in Jung District, central Seoul, on July 19. KOREA BADUK ASSOCIATION </div> Shin Jin-seo, the Go champion nicknamed “Shin-telligence” for his AI-like style of play, defeated an actual AI &#8212; the Go-playing AI KataGo &#8212; in a two-stone handicap match on Sunday. Shin, aged 26, beat KataGo by 4.5 points playing Black after 290 moves in the second game of the Ssen Math-Hankyung Kishin Match at a Korea Economic TV studio. The victory avenged his loss in the first game on Friday. KataGo is a top-tier AI reportedly capable of weighing 26,000 variations within a 20-second time limit. Despite the head start, the match unfolded like a horror movie, with Shin constantly being pursued by an unyielding force single-mindedly bent on catching him.&#160; Shin was given a two-stone handicap, roughly an 18-point advantage on the board. KataGo spent nearly five hours relentlessly pushing to close that gap. The margin narrowed as play went on, and Shin repeatedly found himself one misstep away from losing his lead. But he never let KataGo pull ahead, and ultimately defeated the AI. KataGo opened the second game with an unusually long and complex jeongseok, also known as&#160;joseki &#8212; a standard sequence of moves &#8212; of the so-called AI jeongseok patterns introduced to Go since AlphaGo’s debut a decade ago. Such sequences are so long and intricate that a single deviation can badly damage a player’s position. KataGo, however, had not factored in one piece of practical information: Its opponent was Shin, whose style so closely mirrors an AI’s that he has earned the nickname “Shin-telligence.” Shin followed a sequence of more than 50 moves without error, and by the time the jeongseok ended, a quarter of the board had already been settled. Having narrowed the scope for complications early, Shin then entered the middle game with less pressure, breaking from his earlier promised strategy. <div class="image"> <img src="https://imgnews.pstatic.net/image/640/2026/07/20/0000091448_002_20260720164216571.jpg" alt="" />Go champion Shin Jin-seo plays against the Go-playing AI KataGo at Korea Economic TV's studio in Jung District, central Seoul, on July 19. KOREA BADUK ASSOCIATION </div> “I will avoid fighting as much as possible,” Shin had said before facing KataGo. On Sunday, though, he initiated a fight against the AI &#8212; and won it. When a close battle broke out in the center during the middle game, Shin faced a choice: Connect his Black stones and wait for a later opportunity, or cut through White’s large group and go on the attack. Winning the fight would widen his territorial lead, but combat also meant more variables &#8212; and a lower win probability on KataGo’s calculations. At the critical moment, Shin chose to attack. His 102nd move split White’s large central group in two. A prolonged battle followed, but that single move secured his advantage. “There were countless moments when I wanted to push back or attack,” Shin said afterward. “Each time, I held back and endured it. When the fight broke out in the center, I checked the clock and saw I had time to spare. I would not have made the cut if I had been short on time. I decided to fight while minimizing my losses.” The third game will take place on Tuesday.&#160;Another win would earn Shin a combined 250 million won ($168,000) in prize money and a Genesis G90 sedan.

등록된 댓글이 없습니다.

로그인한 회원만 댓글 등록이 가능합니다.

먹튀폴리스

‘Shin-telligence’ beats an actual AI in two-stone handicap Go match

멤버랭킹

관련자료

멤버랭킹