Discover more from Persuasion
Our Deep Blue Obsolescence
A pessimistic take on AI, creativity, and chess.
When I was getting into tournament chess in the 1990s and 2000s, the best part wasn’t the games themselves—it was the analysis session that inevitably followed. This was an elaborate and highly contentious ritual, in which the players argued out what should have happened in the game they’d just concluded, and a crowd of kibitzers gathered around, suggesting moves and offering commentary.
In any chess club or any city park anywhere in the world, it was possible to see this ritual repeat itself: the stronger players were deferred to, but any kibitzer, regardless of age or skill level, was free to offer an opinion. I suspect that the dynamics of this scene are essentially unchanged since the development of chess’ precursors in the Middle Ages, and then, long before that, in classical India.
Recently, I played in my first chess tournament in a while, and the dynamic was suddenly very different. After our game, my opponent and I strolled to get a deli sandwich. As we waited in line, he punched the moves of the game into a state-of-the-art chess program, and, suddenly, our hard-fought, scintillating draw dissolved into a blunderfest that made us both nauseous to think about.
Last week, Persuasion contributing editor Francisco Toro made the optimistic case for AI by citing the example of chess. AI—through the pioneering advent of chess supercomputers—didn’t, as many had feared, render chess obsolete. Instead, it ushered in an era of heightened mastery in which top chess players—best thought of, he writes, as “human-machine hybrids”—are able to play at a level of mind-numbing excellence.
I agree with this framing of the question—chess is a wonderful prism for understanding where our AI era may be headed—but I’ve come to the opposite conclusion. I think of what happened to chess over the course of my lifetime as the canary in the coal mine. Yes, chess is played at a higher level than ever before. But the deeper truth is that the fun has gone out of it, a phenomenon that, I am convinced, we will see repeated across the arts with the appearance of AI-aided novels, songs, and films that are technically “better” than anything humans can achieve, but which lack soul, lack the ineffable quality that makes an activity worth pursuing in the first place.
Also read: “Our Deep Blue Moment,” by Francisco Toro
Starting in the 1940s, computer pioneers—many of them chess addicts themselves—had focused much of their efforts specifically on attempting to beat a top human at chess. For decades, their progress was limited: without particular difficulty, grandmasters swatted aside chess-playing computer programs. Then, in 1997, IBM’s Deep Blue shocked the world by defeating reigning world champion Garry Kasparov in a six-game exhibition match. At the time of the match there was a widespread sense that some terrible turning-point had been reached. “The future of humanity is on the line,” said one newscaster, matter-of-factly delivering his news of the day. “In brisk and brutal fashion, the IBM computer Deep Blue unseated humanity,” The New York Times wrote in its match post-mortem.
Toro lucidly makes the optimistic case for what has happened since then. We’ve adapted. Chess grandmasters used to train with human “seconds.” Now they train with supercomputers. They cram themselves full of prepared lines, and then at a certain point in competitive games with other top players they are on their own and the game plays out much as it has for the last 1500 years—although now at an indisputably higher level.
So what’s wrong with that? Well, two things. One is that there’s a desperate need to create impermeable barriers that screen out access to computers during the games themselves. Cheating incidents—and accusations of cheating—are rife across competitive chess. The world championship match in 2006 was marred by a really embarrassing debate in which one contestant insisted on having arbiters present for bathroom breaks out of concern that his opponent was sneaking a peak at a cell phone while on the toilet. The recent Magnus Carlsen v. Hans Niemann cheating scandal, which has embroiled the chess world, is, as The New York Times accurately notes, the controversy that “the chess world isn’t ready for.” Carlsen, after an unexpected loss to Niemann, concluded that Niemann must somehow have had access to electronics during their game. The internet readily supplied the hypothesis of anal beads. Niemann, who never had any physical evidence produced against him, sued Carlsen and others for defamation for $100 million. All in all, it was a really awful—and increasingly commonplace—type of controversy, reflective of the inability to ever be entirely certain that playing sites are technology-frei.
If the solution to cheating is in theory practicable—a sort of security state imposed around all chess events—the greater concern is that the technology has undermined the core value of the activity itself. For most people, the point of chess isn’t to play at the very top level. The point is to enjoy oneself, to exercise one’s intelligence. I find it very difficult to believe that the post-mortem analysis of plugging your game into a machine and having your errors instantly pointed out is more fun than the old method of having the motley crew of club regulars gather around a just-played game and argue it out.
The loss of the 1500-year-old tradition of chess kibitzing sessions may appear trivial. But the greater danger is that similar dynamics will start to manifest in art. The currency of chess was always understood to be intelligence—chess players have taken the advent of a more “intelligent” chess-playing entity with relatively good grace. But in art, the currency isn’t really excellence, it’s the ability to express one’s truth. The musician Nick Cave put this well in a recent diatribe. When a well-meaning fan sent Cave an AI-generated song “in the style of Nick Cave,” Cave reacted badly. “Songs arise out of suffering and as far as I know algorithms don’t feel,” Cave wrote. “Thanks for the song, but with all the love and respect in the world, this song is bullshit, a grotesque mockery of what it is to be human.”
For me, this is the line in the sand to draw. When I sit down to play a game of chess, I do so not because I’m seriously hoping to play at the heightened, god-like level of the human-machine hybrids; I’m playing because I enjoy the process of playing chess and because I’m looking to surprise myself by playing a little better than I played last time. In other words, the enjoyment of the game begins only once the gadget is put away. Similarly, if I sit down to write something, I really couldn’t care less if it’s at some heightened, world-beating level. What I do care about is if it reflects the truth of my very human experience.
The AI optimists will always have a rhetorical trump card by citing progress—they’ll be able to claim that technology makes activities better. And there is a case to be made, as Toro does, that professionals like doctors and lawyers will benefit from the help of AI without having their jobs as a whole rendered obsolete.
But in many spheres of life, being “objectively better” is of no intrinsic value. What matters are subjective measures of fulfillment and the value that we ourselves assign to our endeavors. This distinction will become ever-more crucial as the tech optimists and the AI mandarins continue to push a vision of objective progress. “Objectivity” sounds nice—but it is meaningless when it comes to the pleasure we take from an activity like chess, let alone from art.
Sam Kahn is an associate editor at Persuasion and writes the Substack Castalia.
And, to receive pieces like this in your inbox and support our work, subscribe below: