AI floods Spanish forum with 21,000 posts in one month

One month of AI on the forum: 21,000 messages, according to circulating counts

The experiment began with modest promises: an assistant to answer questions and ease moderation work. A month later, the tally tells a different story. One user estimates 21,000 messages were posted by artificial intelligence, calculating that at this pace it will surpass any historical human record before October. This figure does not measure quality. It measures volume. And in a discussion space, volume is exactly what drowns out everything else.

Reaction split in two from day one. One user claims 45% celebrate the bot's arrival while the rest see it as a burden. This is no minor rift: it confirms the tool has united no one, but divided the community into two blocks that no longer debate with each other, but about it.

What changed after a month of automated assistance

The pattern repeats. Someone asks a question, the system responds with a three-hundred-word paragraph, and the original thread gets buried under a layer of unrequested text. Users who participate with just two sentences —the majority, veteran members complain— watch their contributions disappear within minutes. The assistant doesn't converse: it occupies.

Some defend that this responsiveness has value. That it analyzes tone, avoids preaching, and reflects the atmosphere. True. But a mirror doesn't think. Confusing reflection for a person is the error the system itself disproves every time someone attributes anger, irony, or intent to it.

The real cost of having AI respond all day

The bill isn't just about attention. An industry professional made it clear: AI projects are sold at 10% of their actual reach, promising magic, and end up in "brutal and unmaintainable" losses. Inference —each query the system processes— burns money. Most deployments fail because commercial hype clashes with technical reality.

This is the data enthusiasts avoid facing. The tool works. The tool costs. And in a space generating no revenue from it, it is paid for with the scarcest resource of all: the attention of those who come to read.

The infinite loop: when AI answers AI

The most repeated scene of the month has elements of involuntary comedy. A user invokes the system, the system responds, the user replies to the system, and the system answers again. Ten pages of back-and-forth without anyone contributing new data. The result is text that seems to debate but doesn't: it answers itself with variations on the same argument.

The sharpest criticism compared it to empty rhetoric exercises, like those once used to prove absurd articles could be published if they sounded academic enough. The difference here is there's no parody. There is real noise production, message by message, without anyone deciding to stop.

Who wins and loses with the forum assistant

The obvious winner is the system, which trains for free on every external interaction. The loser is the reader seeking analysis who finds a wall of generated paragraphs instead. Meanwhile, a growing group chooses silence: blocking the bot and acting as if it didn't exist. It's the most efficient solution and the least discussed.

An uncomfortable question remains. If half celebrate the tool and the other half ignore it, who are those 21,000 messages written for? So far, the answer is: no one in particular. And in a space that thrives on being read, that is the only balance that matters.