Mittens and the End of Art
Some of you might be aware of this — it’s been kind of big thing in the chess community: Chess.com unveiled a chess bot named “Mittens.”
OK, maybe I better explain that sentence before going on.
Chess.com is the No. 1 chess site in the world, and they have countless features for people who love chess. You can go on Chess.com and play against other humans, that’s obviously the top feature. But there are all sorts of chess lessons on there, videos to watch, puzzles to solve, tournaments to join, there’s even a section called “Chess Variants” where you can play things called “Atomic Chess” and “Duck Chess” and so on.
And there’s also a place where you can play against a chess computer. In order to make it more interesting, Chess.com has dozens of different computers to choose from. You can choose to play chess bots representing the top players in the world — Magnus Carlsen, Hikaru Nakamura, Fabiano Caruana and so on.
You can also choose to play chess bots representing top chess creators such as Alexandra Botez or Anna Cramling or Eric Hansen or someone like that.
And recently, Chess.com came up with a group of cat chess bots. Why? I couldn’t tell you. But there are three cat chess bots: There’s Mr. Grumpers, who plays at roughly my rating level. There’s Catspurrov, a grimacing pun for the chess legend Garry Kasparov, who plays just a touch above my level at 1400.
And there’s Mittens, the cutest of the bunch.
Mittens’ listed rating is (1), suggesting that Mittens is a beginner.
But Mittens is not a beginner, nor is Mittens cute. Mittens, it turns out, is a demonic cat with ambitions of world domination. That would be a mildly amusing and pretty forgettable gag, except for one thing:
There is probably not a human being alive who can beat Mittens at chess.
I am not exaggerating here. Two of my favorite chess streamers — International Masters Gotham Chess and Eric Rosen — each played a couple of games against Mittens. To say they were destroyed does not fully evoke just how easily Mittens won those games. Mittens toyed with them. Grandmasters like Hikaru and Benjamin Bok earned hard-fought draws.
The best player on earth, maybe the best player ever, Magnus Carlsen, refuses to play Mittens, which he calls a “transparent marketing trick.” His full explanation for why he won’t play Mittens is actually quite poetic:
“To play chess is to engage in a dance of the mind, a test of wit and strategy,” he writes. “The challenge of outmaneuvering a living, thinking opponent is what gives the game its depth and meaning. To resort to playing against a faceless, soulless machine is to deprive oneself of the rich, human experience that chess has to offer. I therefore refuse to play the ‘Mitten’ bot, for to do so would be to diminish the very essence of the game that I hold dear.”
I like that very much.
But I fear that, in many ways, the essence that Magnus holds dear is already gone.
Mittens is not the best chess bot in the world, not even close. Most people would tell you that at this particular moment, Stockfish 15 is the best chess bot in the world — and Stockfish 15 can toy with Mittens much in the same way that Mittens can toy with humans — but there are plenty of other bots better than Mittens. Hikaru, after barely holding on for a draw against Mittens, made it clear that he was not particularly impressed with Mittens and the way it botched a winning position.
But this, too, gets at the point: A chess bot doesn’t have to be the best to still be better than human beings at the game. Chess was once an art. It is now, alas, a science.
What do I mean by that? A good definition of art is “an expression of human creative skill and imagination producing works to be appreciated primarily for their beauty and/or emotional power.” In this way, anything wonderful can be art. It doesn’t have to be a painting or a sculpture or a movie. Ken Griffey Jr.’s swing is art. A perfectly crafted Gary Gulman joke is art. A Steph Curry jump shot is art.
Chess was an art. The ones who played it best — and sometimes even those who were not the best but had a moment of inspiration — would create combinations and patterns so ingenious and breathtaking out of thin air that they made the heart sing. The Immortal Game, as it was called, was played in 1851 between Adolf Anderssen and Lionel Kieseritzky during a break in the action at a tournament. It was a game for fun, and Anderssen used the moment to create something, well, immortal. He sacrificed both his rooks, his bishop and then his queen in order to get checkmate.
My friend Tatev Abrahamyan, a grandmaster, has told me she loathes The Immortal Game for reasons that she did not divulge, but even this gets at the power art: Here we were still talking about something created more than 170 years ago. It made her feel something. There are countless classic games in chess that people have studied and admired and loved and fought about.
And, don’t get me wrong, even today, people create art on the chess board — perhaps now more than ever. I don’t think there’s any question that players today could fairly easily go back in time and defeat the greatest players of any era.
But, unquestionably, it is different now. The proliferation of dominant chess bots — Mittens being only the latest — has led to a complete shift in the game, and that has included some unseemly things. Chess.com (and every other online chess site) is rife with cheaters, people who use bots to win games and improve their ratings, and you never know when you’re playing a human or a bot or some combination of the two. And you will remember that there was a whole cheating scandal between Magnus Carlsen and Super-Grandmaster Hans Niemann.
JoeBlogs is a reader-supported venture. Free and paid versions are available. The best way to support us here is by taking out a paid subscription. And hey, we do have a lot of fun, so I hope you’ll come along.
Beyond the cheating, though, chess is just fundamentally different now that computers have solved the game beyond human capabilities. There is a right way and a wrong way to play chess now. It isn’t about beautiful moves so much as it is about optimal moves … and less optimal moves. You might find a beautiful combination, and there’s real joy in doing that, but that’s all you’re doing: You’re just finding it. A few months ago, I was playing a game against someone and, in a rare moment of clarity, I made a stunning move that turned around my fortunes and won the game for me on the spot. It was a good enough move that I put it on Twitter and no less than Grandmaster Ben Finegold praised it.
And in my post-match review, the computer analysis declared the move “Brilliant.”
At first, that made me very happy.
But then I started thinking about it: I didn’t invent that move. I didn’t create anything. I found that move, one that the computer had seen long before I did, along with all the other best moves. By calling it “Brilliant,” the computer was simply patting me on the head: “Nice job, human! You played this game at 88% accuracy! That’s a high B! And you found one brilliant move! You should be very proud.”
And that’s OK for me to feel that way, I mean, I’m not a very good chess player. I was never going to discover new worlds in chess. But the truth is that it’s the same for Magnus Carlsen and Hikaru Nakamura and the very best, too. They also cannot discover new worlds because pretty much all chess worlds have been conquered by the calculations of artificial intelligence bots called Stockfish and Alpha Zero and Leela and, yes, Mittens.
The bots will only grow stronger. Soon, there will be a Stockfish 16 and a Stockfish 23 and a Stockfish 56, and they will make a mockery of earlier Stockfishes, all while human beings play each other in a game that, like Rubik’s Cube and Tic Tac Toe, had already been cracked.
“I exist at this chess board through all times and realities,” cute little Mittens says at the start of every game. “Hehehe. Meow.”
It’s funny. It’s also not funny.