It Should Have Been AlphaBaduk
Ten years after AlphaGo and Lee Sedol — the romance of Go learned from books, two moves that chilled me, and a small regret about the name.
It’s been ten years since the match between Google-DeepMind’s AlphaGo and the master Lee Sedol, and various bits of news drift past. Let me put down a few stories around Go.
Go learned from books, and the romance in it
In middle and high school I saved my allowance to buy Go magazines and read them. I never played people much in earnest, but I kept, for a long while, the hobby of following game records — replaying them by hand, move by move. There were the game records in the newspapers, and it was fun to walk through a full record one stone at a time.
Oddly, I was bad at playing actual people. I hated collapsing to one or two cheap tricks, and losing wounded me more than it should. I don’t know my exact rank, but later, on internet Go, I hit 1-dan on some site and stepped away. In the official amateur scale I was probably around 3–4 kyu.
The masters Cho Hunhyun and Lee Changho, who appear in the recent film The Match, were of the same era, and the monthly Go magazines carried a lot of the Japanese games. Since mine was Go learned for show, the stories outside the contests were more fun — Cho Chikun, who played as if his life were on the line; Fujisawa Shuko, who said he only needed to win four games a year; the always-winning Lee Changho and Cho Hunhyun I cheered for out of patriotism. But for show, or for romance, the first name out is Takemiya of the Cosmic Style. The one move I still remember is move 69 in the first Fujitsu Cup final. A shoulder hit against a star point.
In the first-ever world-tournament final, a move that’s in no book? Cosmic Style or not, I wondered whether that was even possible, and the way he crushed White afterward was the highlight too. I didn’t watch it live, but even with my shallow skill I remember wondering how such a thing was possible — the most memorable scene, to me, among moves a person has ever played.
Early machine Go
When I joined the computer-engineering department in 1993, a professor told me chess AI — Deep Blue, or Deep Thought — played better than ordinary people, and around then a program called Badajanggi, running on Windows 95, played Korean chess better than most people I knew. By the time I graduated, they said it had beaten the chess champion.
But for Go, even Eunbyeol, said to be the strongest of its day, was rated only around 2–3 kyu — the number of cases is absurdly large, and I spent a long time thinking the intuition people had built up over the ages really was something. In the thick of company life, knowing it only as a kind of pastime, ten years ago that day came.
The AlphaGo match
I was working at Google when there was a big event in Korea, and though I was at headquarters, I’d liked and cheered for the strongest of the age, master Lee Sedol, so I watched with interest — and, in a how dare a mere machine spirit, I bet on the human. The result, as everyone knows, settled into humanity salvaging one final win, and my bet, sad as it was, let’s call it the heart of a supporter.
Get Chaesang Jung’s stories in your inbox
Join Medium for free to get updates from this writer.
Two record-breaking scenes. The first is move 37 in game two, a shoulder hit on the fifth line.
By the analyses, after this move there was no chance to turn it around. It touches, in a sense, the example above — and thinking that what we’d call romance when a person does it was a move a machine spat out, I remember it being eerie. As a bonus, from then on I came to think every move AlphaGo stamped out must have a reason. Replaying the records of games one and two, I remember being very afraid.
The next scene is the famous move 78 in game four.
A move in no book, and on the other hand a move hard to believe a person made. Hard to call it a fluke, hard even if you credit deep thought and prediction — and that it was a kind of deliberate bug, exposed by Monte Carlo tree search plus pruning running up against blowing out memory, I now fairly agree with.
As a fan, I’d handed the master I looked up to like an idol, the strongest of his age, the title of humanity’s last winner — but on the other hand I wondered what pain I was handing a man who’d made this his lifelong work, and his decision to retire, and the different life after, I cheer for too.
What I think ten years on
As of 2026 it’s gotten hard to find anything a machine doesn’t do better than a person. That it’s fine teaching material goes without saying, and ten years on it will go into history… but if it had only been AlphaBaduk, I think it would have been a touch better. Go’s Chinese romanization is Weiqi, the Japanese Go, the Korean Baduk — and though the strongest players were Koreans, I felt the regret of having somehow missed the move.
Seeing that Demis Hassabis recently went to Korea to play a special game with the current strongest, master Shin Jinseo, left a bitter taste. He is the father of AlphaGo, true — but an event bringing in a Nobel laureate for a teaching game against the world’s strongest… I have a regret that Google Korea could have paid it a bit more attention.
Part of IT Thoughts, 2026 — notes on where the industry is going, from a first-person engineering view. See the full series →
Adapted from my Korean essay on Brunch: brunch.co.kr/@chaesang/178
Originally published at https://ieumtech.net on September 9, 2026.
