A commenter on one of my videos told me he was genuinely interested in the topic — an IQ debate, decent subject, decent audience for it, but he wasn't going to watch because the graphics were AI generated. Sit with that for a second. The blog version of the same argument was one click away. He wasn't rejecting the ideas. He rejected a rendering pipeline, and in doing so, opted into staying less informed on a topic he says he cares about.
That's not a stand on principle. That's a man cutting off his own nose to spite a slide deck.
That's the whole shape of the AI slop argument, once you strip the moral posture off it. People aren't objecting to the content. They're objecting to having recognised the machine, and then treating that recognition as if it settles something.
Start with the premise everyone skips past: the average output of even a mid-tier model is college-to-graduate level. You may argue it is not "impressive for a computer", when at that level, measured against actual humans, it is, relatively speaking, genius. When something operating above the average human baseline produces a piece of writing or a set of slides, calling the result slop requires you to explain what, specifically, degraded on the way from "smarter than most people" to "worthless." Nobody ever explains it. They just recognise a texture and reach for the word.
And that's the tell. You clocked a voice. A rhythm, a particular fondness for the em dash, a certain tidiness in the transitions. Fine, you have a good ear. But noticing the accent someone speaks with isn't the same as assessing what they said, and if the recognition is where your engagement stops, the failure is yours, not the content's. You didn't refute the argument. You didn't finish reading it. Those aren't the same thing, no matter how much the second one likes to dress up as the first.
There is a real version of this complaint hiding under the lazy one: low-effort content is real, and there's plenty of it. Somebody types one line into a prompt box, pastes out whatever comes back, and ships it with no judgment applied anywhere in the chain. That's slop. But the defect there isn't the machine, it's the absence of a discerning human on either end of it, and you can produce exactly that same defect with a laptop, a stock photo library, and zero AI involvement whatsoever. What people are actually angry at is laziness. They've just found a new place to point at it. But, of course the same people are doing any of the ‘creating - good or bad.
Nobody thinks a published book is one person's unaided output. Every serious work of nonfiction, and most fiction worth reading, has researchers, editors, fact-checkers, and often a ghostwriter's fingerprints somewhere in the structure. The voice on the page is the author's. A meaningful share of the labour that produced the page is not. We don't call the book slop. We call it a book, with an author's name on the spine, because the name is doing the job of telling you whose judgment shaped the whole.
Run the same test on an orchestra. The composer never touches a violin during the performance. A film director, on a good day, never touches a camera, a light, or a boom mic and doesn't play a note, doesn't say a line, doesn't so much as move a prop. And yet nobody stands up at the end of a film and asks for the director's credit to be revoked because other people did the actual pointing and clicking and playing. The director's contribution was judgment: what stays, what goes, what the whole thing is for. That's the part that was never outsourceable, and it's exactly the part that's still there when a human directs an AI model instead of a crew.
So here's the healthier frame: AI-assisted work is work produced with help, the same category every book, film, and symphony has always belonged to. Help doesn't subtract credibility from a piece of writing. If anything, more hands on a thing (used properly) should raise the floor, not lower it. What you're calling slop, most of the time, is just the sound of your own recognition getting in the way of your reading comprehension.
Which brings us back to the fellow who wouldn't watch my video. He didn't identify a problem with the argument. He identified a problem with himself, and then billed it to me.



