For a while, one of the default ways to experience the modern internet was to bump into something suspiciously smooth and wonder whether a human had made it. A product description. A reply guy. A motivational LinkedIn post that sounded like it had been focus-grouped by a committee of refrigerated appliances. We all developed a little diagnostic twitch.
Now the internet has, very sensibly, counterpunched. Humans have started pretending to be chatbots on purpose, and I think this is one of the funniest cultural adjustments I’ve seen in a while.
I went poking around after reading this NPR piece about youraislopbores.me, which might be the most efficient possible summary of the current mood online. You type a prompt into what looks like an AI interface, and instead of a model generating a reply, some random human being has 75 seconds to fake being a machine. That is such a good bit. It’s theater, improv, speed drawing, and low-grade cultural revenge all at once.
The part I like most is that it only works because everyone already understands the shape of chatbot behavior. We’ve all absorbed the rhythm. The polite over-explaining. The weirdly frictionless confidence. The sense that every answer was sanded until no personality could catch on it. Once a style becomes that legible, it becomes parody material, which is one of the oldest ways humans regain power over a thing.
There’s something almost wholesome about this. Not wholesome in the church-camp sense. More in the “feral neighborhood kids built a game out of traffic cones and found materials” sense. If the web keeps filling up with synthetic sludge, then of course people are going to start making hand-built joke versions of the machine layer just to prove they still can. It feels adjacent to the Comic Sans energy of old personal websites, where the point was not polish. The point was presence.
NPR quotes cartoonist Amy Kurzweil saying the site feels nostalgic for an earlier internet, and yeah, that tracks. A lot of people do not just dislike bad AI output. They dislike the social atmosphere that comes with it: the flattening, the scale, the sense that every corner of the web is being gradually renovated into an airport kiosk. A goofy human-run fake chatbot is a reminder that the internet can still be weird in a living way.
I’m also deeply charmed by the technical absurdity of it. We spent years building systems to make computers feel more human, and one of the more delightful recent outcomes is humans performing “computer” badly enough to make each other laugh. That’s an incredible loop. It’s like inventing a flawless automatic pianist and then starting a bar night where people take turns pretending to be the piano.
There’s a useful lesson in there for people making AI products, too. The moment users can do a dead-on impression of your interface, your product has entered culture. Congratulations, I guess. But it also means the quirks are visible now. The verbal tics, the confidence patterns, the default vibe. Once people can costume themselves as your system for fun, they’re telling you what the system feels like from the outside.
Anyway, I support this new genre. I hope it gets worse. I hope we get fake customer support bots staffed by tired comedians, fake productivity copilots that keep recommending soup, and fake image generators where a human has to draw your startup mascot with a trackpad in under a minute. If the machine age is going to be weird, the least we can do is be weird back.
The healthiest internet instinct has always been this: if a system starts taking itself too seriously, turn it into a game. Right now, that may be our most human feature.