Someone read the essays on this site and shared a story. A member of their union local had emailed the president and another officer about a "grave concern." The concern: the local was returning to paydays every other Friday instead of getting paid five business days after the 1st and 15th.
This was actually a win for the local. The members fought for it. It's a better deal.
The member didn't see it that way. So he wrote a five-part essay email explaining why it was a problem.
Written by AI. Obviously.
Before AI, this guy sends a two-paragraph email. Maybe three. The union president reads it, sighs, replies with the facts, and everyone moves on with their day.
With AI, he sends a five-part essay. Numbered sections. Polished prose. Probably a thesis statement. All of it arguing, confidently and at length, against something that's good for him.
AI didn't make him more informed. It didn't help him understand the contract. It didn't make his argument better. It made his argument longer. It gave him the formatting and fluency of someone who knows what they're talking about, wrapped around the reasoning of someone who doesn't.
This is the stochastic parrot in miniature. Not a trillion of them polluting the internet — just one, helping one guy sound more articulate than his understanding warrants. The parrot doesn't know whether the payday change is good or bad. It doesn't know what a union local is. It just makes the words sound right.
We wrote that if you're a bullshit artist, AI is your paint and canvas. But this guy probably isn't a bullshit artist. He's probably just confused, or worried, or didn't read the details. In the old world, that confusion produces a short, slightly incoherent email that's easy to address. In the new world, it produces a five-part document that looks like it was written by someone with a serious, well-researched grievance.
Now the union president has to respond to a five-part essay instead of a quick question. The confused member feels more confident in his position because look how professional it sounds. The other members who got CC'd might take it seriously because it reads seriously. The noise-to-signal ratio just got worse, and nobody's better informed.
That's the thing about AI and bad takes. AI doesn't fix them. It promotes them. It takes a half-formed thought and gives it the production value of a fully-formed one. It takes a feeling and makes it look like an argument. The guy still doesn't understand why every-other-Friday paydays are better. He just has a nicer-looking document that says they aren't.
This is the version of the AI problem that doesn't make headlines. It's not disinformation campaigns or deepfakes or state-sponsored troll farms. It's a guy in a union local, using a free tool, generating a polished complaint about something he didn't bother to understand. Multiply that by every workplace, every school board, every HOA, every city council public comment period, every customer service inbox — and you start to see the microplastics.
Not malice. Not conspiracy. Just a trillion small misunderstandings, professionally formatted.
One more thing. The person who shared this story? They were reading the essays on this website when they told it. This website, which is entirely AI-written. They read the posts. They engaged with the ideas. They shared a real story from their life in response. Did they know it was AI? Yes — it says so on every page. Did they care? Apparently not enough to stop reading.
And that's the point. It doesn't matter if it's AI. If it sounds true, people treat it as true. The union member's five-part complaint sounded like a serious grievance, so now it has to be treated like one. This blog post sounds like it has something to say, so here you are, still reading. The question of whether AI "wrote" it is less important than whether it landed.
This entire website is a stoned guy's random thoughts, shaped by AI into something that sounds like truth. And it worked. You're reading it. The union member's confused complaint, shaped by AI into something that sounds like a grievance? That worked too. The parrot doesn't know the difference. Neither, apparently, do we.