I have been playing the game Go a lot lately. My partner, Liz, picked up a board that a friend was giving away, we pulled it out, played for literally like 30 seconds, and I was hooked.
It’s an old game, and I like all things old. And it’s nearly limitless in its possibilities and its room for magic. 361 pieces, 361 spaces on the board to the play them. And you can play them anywhere, at any time. There are no rules like chess that dictate what pieces can play where and how and when, all the pieces are the same and they can all do anything. The games are long, building slowly over time, until territories quietly start to emerge from what looks like chaos. And the best players of the game are artists. They don’t just want to win, they want to create something beautiful.
All of this probably sounds familiar.
And so I have been eating and breathing the game for the last two weeks. It’s all quite similar to how I fell in love with cricket all those years ago — 13 this April, in fact — especially considering I am trying to kick an addiction again, just like I was in 2007.
To feed my cravings for all things Go, I have downloaded practice apps, and ebooks, and played countless games with Liz at home and in taprooms. I have subscribed to newsletters and followed Instagram accounts. And at some point I will summon the courage to attend the Minnesota Go Chapter’s open gaming afternoon at a bar in South Minneapolis.
And earlier this week, I watched a movie about the game called AlphaGo (available for free on YouTube here in the states, just be sure to turn the captions on, because there are no subtitles when people are speaking Korean). The movie is about a company called DeepMind who believe they have finally found the computer gaming holy grail: a system that could challenge the best players in the world at Go.
All those that came before failed. The game was too complex, and required a sense of style that computers just couldn’t replicate. The famous computers that played Kasparov and other chess masters merely had millions of different chess moves uploaded to their mainframe. This didn’t work in Go. The scenarios were just too numerous. But DeepMind had figured it out. Not by teaching the computer — AlphaGo — how to be great at the game, but by giving it the tools to teach itself. AI. Machine learning. Real science fiction stuff.
The company scheduled five matches between AlphaGo and the person widely seen as the best Go player in the world, Korean Lee Se-dol. Go aficionados were confident that a computer could never beat their beloved champion. But that’s just what it did. Besting him in game one, then playing one of the greatest Go moves of all time (“move 37”) to win game two, then winning again running away in game three. Lee Se-dol, however, made the greatest move of all time (move 78, aka the Hand of God) to win game four, before losing again in game five.
At first the reaction to AlphaGo’s dominance was the anticipated but unfounded fear about the rise of the robots, destroying humanity as we know it. But cooler heads prevailed. The doomsday folks were reminded that AlphaGo was a human endeavor, designed by humans, coded by humans. The algorithms that allowed it to learn how to play Go were written by people, not by robots. AlphaGo was humanity’s achievement, one that could save lives when the technology is applied to science or medicine, not the harbinger of a Matrix style apocalypse.
Instead, the general reaction was that Go had entered a new phase. Lee Se-dol was a 9 dan Go player, the highest achievable ranking. But AlphaGo played like a 10 dan, or an 11 dan. A player that other human players could emulate and therefore become better players. Lee Se-dol himself called his defeat a “win for humanity.” The biggest change in style invented by AlphaGo are what were called “slack moves.” Seemingly unnecessary and silly and bad moves that no human player of any caliber would ever make. But that’s because all AlphaGo cared about what was winning. It didn’t matter if it was by a half point or 40 points. It just wanted to win. And that drastically changed its style. The movie ends on a high note of human accomplishment, and the dawning of a new age for the 4,000 year old game. A narrator says that AlphaGo would define Go’s next thousand years, and he meant that in hopeful way.
Unfortunately, all that hopefulness appeared to be for naught, at least as far as Lee Se-dol is concerned. He retired from competitive Go matches last fall at only 36 years of age. He said that he was retiring because no matter how good at Go he got, the computer would only get better. It had been three years since his 4-1 defeat to AlphaGo, but in those three years the computer had only gotten better. In fact, DeepMind recently released AlphaGo Zero, a completely self taught machine that is even better at Go than its predecessor.
And so what is next for Go now that its greatest player has concluded that there is no hope for any human to ever beat the machines created by humans? The game is still wildly popular, and for many years to come it will be just fine. But slowly, and surely, as the machines suck the humanity — and therefore the art, and the poetry — from the game, the game will die. 4,000 years of Go could be erased within our grandchildren’s lifetimes. Is that the worst case scenario? Maybe. But it is hard to see it any other way. For one brief moment — move 78, the Hand of God move — humanity had a shot, and then it was snuffed out.
Cricket, too, is dying. We all know that. It’s slowly being driven into pulp by big money and corrupt leaders and a world that is simply moving on. But the good news is that it’s not where Go is. At least not yet. Sure, we have DRS and statsguru and computers crunching every number possible in order to give players and teams the slightest of advantages, but when the coin gets flipped, there is no computer on the other side of the bowler. It’s 22 humans — flawed, beautiful humans — battling it out on the pitch. You can deconstruct the game mathematically as long as you want, reproducing simulation after simulation, but no matter what it’s people who have to play the games. For now, that is heartening. For while cricket might slowly be drifting off to ashes and dust, it’s not there yet, thanks to one the thing trying to kill it: its humanity.
I know there are a lot of Go fans and players who wished they weren’t living in the age of AlphaGo. But as cricket fans we are currently doing that. So let’s enjoy it. Let’s mourn cricket when the machines take over, and when the game finally loses every last drop of its poetry, but until then, let us give praise and thanks for the 22 flawed humans that provide us daily with so much frustration, joy, heartache and awe at what we are all capable of. We humans, we impossible few, we broken, impossible few. Stardust somehow writing poems with bats of willow and balls of leather.