brianchitester.comPosts

The Magic Circle Leaks

Score: 0Best: 0

Snake

Arrow keys or WASD to move. Swipe on mobile.

Press space to start / pause.

I asked an AI coding agent to build Snake and a couple of minutes later, it had. I played a round — ate a few dots, crashed into my own tail — and sat there oddly pleased with myself for something I hadn’t really done.

Then a small symmetry surfaced. When I was a kid I used to play Snake on my dad’s phone, riding around in his truck while he worked. A round lasted a minute or two. Which is about how long the agent took to write the whole game. You can now create Snake in the time it once took to play Snake.

That observation is the kind of thing that gets written up a thousand ways at the moment, almost always as a story about cost. The cost of creation is collapsing. Capability is being handed out for free. Depending on who’s holding the pen, this is either a renaissance — everyone’s a maker now — or a hollowing — nobody has to earn anything anymore. Both versions agree on the underlying claim: the interesting thing that happened is that making got cheap.

I think the minutes are a decoy. The interesting thing that happened to me wasn’t that making got cheap. It was that I had fun, and the fun was suspicious. Because by any careful definition of the word, what I did shouldn’t have counted as fun at all.

What play actually is

The person to consult here is Johan Huizinga, a Dutch historian who in 1938 wrote a strange and ambitious book called Homo Ludens — man the player — subtitled, in English, A Study of the Play-Element in Culture. The title is a deliberate provocation against the other names we’ve given ourselves, Homo sapiens, man the knower, and Homo faber, man the maker. Huizinga’s claim is that play deserves a seat at that table, and not a small one. His real thesis is that play isn’t a product of culture but the soil it grows from — that law, war, poetry, ritual, and commerce all began as, or still carry the structure of, play. And there’s a telling detail in that subtitle: the Dutch original described the play-element of culture, and the English softened it to in. Huizinga objected. “In” makes play one ingredient among many. “Of” makes it the ground. He meant the ground.

For my purposes the gold is in his definition. Play, he says, is free — it’s voluntary, and play to order is no longer play, since coercion kills it. It stands outside ordinary life, marked off as “not serious,” yet it absorbs the player completely. It happens inside what he called the magic circle — a bounded patch of space and time, the board, the court, the screen, where special rules hold and outside of which they dissolve. And then the clause that should make you stop: real play is connected with no material interest, and no profit can be gained from it. Pure play is defined, in part, by its uselessness.

The play that stopped coming

Here is the part I’d been noticing for a while without naming. Playing games isn’t fun the way it used to be. As a kid I could disappear into a game for hours. Now, half an hour in, the absorption doesn’t arrive, and worse, a second voice shows up and says: this is a waste of time. The thing that used to be effortless leisure now has to be defended, and I can’t quite defend it, so I stop.

That voice isn’t mine, originally. It’s the ambient pressure of a culture that has decided time is for converting into output. You feel it everywhere now — the sense that an unmonetized hour is an hour leaking value, that rest has to be reframed as “recovery” before it’s allowed, that a hobby is better if it could plausibly become a side income. Under that pressure, consuming a game starts to feel like a small defeat. You’re spending the one resource you’re supposed to be compounding, and getting nothing legible back.

You can watch the same pressure operate socially with a single substitution. “I beat Halo last night” earns a shrug. “I made a game last night” earns respect. Same Tuesday, same hours, arguably the same joy in the doing — but one is a confession and the other is an achievement. The activity barely changed. What changed is which one you’re permitted to say out loud without a flicker of embarrassment.

So the honest description of what happened with Snake isn’t that a tool made me productive. It’s that the tool found a loophole. The economy had quietly eaten my ability to enjoy doing nothing, and the agent handed me back a way to play that was disguised well enough to get past the guard — I bolted a blog post onto it, a flush of status in front of the people whose opinion I track, the quiet plan to build a version that’s mine. The fun was real. It just had to arrive in costume.

By Huizinga’s lights, this is a corruption — play adulterated with exactly the material interest that’s supposed to be excluded from the circle. If he were the only authority, the story would end on a slightly mournful note: I no longer have access to real play, only to a profitable counterfeit of it. But he isn’t the only authority, and the man who came after him saw the loophole coming.

The leak

Roger Caillois was a French writer of magnificently uncategorizable range — he called his method “diagonal science,” cutting across fields rather than down through one — and in 1958 he wrote Man, Play and Games, which opens by saluting Huizinga and then, politely, taking him apart. Two objections. First, the “no material interest” clause quietly writes gambling out of play, which is absurd; the casino is plainly a temple of gaming. Second, Huizinga was so fixated on play’s grand social function that he never actually sorted the games themselves. Caillois set out to do the sorting.

What he produced is the part still in use. He split play into four families by the pleasure each delivers:

And he laid every game on a second axis running from paidia, wild improvised exuberance, to ludus, play bound by arbitrary, demanding rules — what he beautifully called the taste for gratuitous difficulty.

Then comes the move that rescues my suspicious afternoon. Caillois argued that each form of play has a corruption — what it becomes when it escapes the magic circle and fuses with ordinary life. Agôn uncontained curdles into raw ambition. Alea into superstition. He even claimed whole societies organize themselves around different pairs: archaic ones around mimicry and vertigo, the mask and the trance; modern ones around agôn and alea — merit and chance, the examination and the lottery, the career and the jackpot. But here’s the crux of everything I’d stumbled into: play does not stay politely inside its circle. It leaks. It entangles with money and merit and identity, and that leak is not an aberration — in a civilization built on agôn, it’s the normal way play survives.

Which reframes the costume entirely. Disguising my fun as output isn’t a sad compromise I made with a joyless economy. It’s the standard adaptation. I made my play legible to a society organized around merit, the same way the lottery makes the pleasure of chance legible to a society organized around money. The disguise isn’t a betrayal of the play. It’s the price of admission to a world that won’t seat play at the table under its own name.

Why some of us are bitter and some aren’t

Caillois also hands you a clean read on something I keep seeing in my own field, which is full of capable people, and a striking number of them are bitter about exactly the tools that have me excited.

Writing code by hand was, in Caillois’s terms, heavily ludus — gratuitous difficulty, mastered slowly, loved by its best players for the difficulty. The skill was the point. AI strips much of that ludus out. So of course the people who were playing for the difficulty feel a loss; you’ve removed the part they came for. Their bitterness reads as a complaint about quality or craft, but underneath it’s grief — a scoreboard that counted their best skill has quietly stopped counting.

I was never in it for the ludus. I was a decent web developer and never a great engineer. In college I had a circle of friends who built little games almost every week — we’d get together and play whatever they’d cooked up — and I loved those nights, but always as a player. Actually building one took skills I didn’t have; the fact that I was never at the top of the coding game was, looking back, a ludus barrier — a gate made of gratuitous difficulty that kept me out of the room where the games got made. What I wanted was the paidia: the loose, exuberant oh, what else could I make that those friends ran on every week while I admired it from outside. AI lowered the difficulty barrier, and the paidia came flooding through. The same event that strips the ludus the experts were playing for is the event that finally hands me the paidia I was locked out of. We’re not even playing the same kind of game. That’s why one tool produces mourning at one desk and delight at the next.

Where this lands

The kid in his dad’s truck needed none of this. He wasn’t ashamed of Snake, wasn’t asking whether it counted, wasn’t building a case for why the time was well spent. He just wanted the high score, and he wanted to show his dad how long the snake had gotten. He played inside the circle, for nothing, and that was the entire point.

Somewhere between then and now I started needing a permission slip. The thing the agent gave back to me wasn’t the play — I’d had that at eight, for free. It was the permission. It dressed an hour of play in the one costume my adult life would wave through, and only then would I let myself enjoy it.

That’s the part worth sitting with, and it’s not really about AI at all. Huizinga, writing on the literal eve of the war that would kill him, thought modern civilization was losing its capacity for genuine play — that the real thing was being crowded out by spectacle, by professionalization, by a seriousness that couldn’t tolerate the useless. He’d have recognized my afternoon immediately. Not as the renaissance the optimists are selling, and not as the hollowing the doomers are mourning, but as one more small instance of a culture that can no longer let play in through the front door, and has learned to smuggle it through the back, wearing the uniform of work.

The good news is that it still gets in. The play was real; the dots still wanted eating. But we should at least be honest about what we’ve built — a world where the most natural thing a person does, the thing that’s older than culture itself, now has to arrive disguised as something productive before we’ll permit it. I’m going to go build my version of Snake anyway. I just notice, now, that I had to file it as output before I’d let myself call it fun.

2026 © Brian Chitester.