AI slop is annoying. But the real problem isn’t that machines are writing it. The real problem is slop itself, and human slop has been with us much longer.
You’ve seen AI slop. The blog post that opens with “In today’s rapidly evolving digital landscape…” The LinkedIn post that “circles back to synergize key learnings.” The X thread where every sentence is a paragraph, every paragraph ends with “Here’s what I mean:”, and the whole thing could be reduced to one sentence without losing anything. The image with that plasticky perfection, too smooth, too well-lit, faces just slightly wrong around the eyes. The article that confidently lists seven things in the same breathless tone regardless of whether it’s writing about productivity hacks or nuclear physics. A paragraph like the one you just read.
Fair enough. AI slop is real and it’s everywhere and it’s annoying.
The reaction to AI slop has been to treat “AI-generated” as the problem category. Platforms are adding human verification badges. Forums are instituting CAPTCHA gates to prove human authorship. The implicit argument: if a human made it, it’s worth engaging with. If an AI made it, it isn’t.
This is just wrong, and it’s wrong in an instructive way.
Consider Grokipedia, high-quality, densely researched AI-generated content that’s more useful than most human-written Wikipedia edits. Consider the countless human-written LinkedIn posts that are perfect specimens of slop: meaningless, formulaic, self-congratulatory, adding nothing to anyone’s understanding of anything. The origin of the content tells you almost nothing about its quality. What matters is the content itself.
My honest filter: I don’t want to read 99.5% of what gets written, maybe 99.9%. Not because of who or what produced it. Because most of it is low-quality, uninteresting to me, or both. That ratio is roughly the same whether the author is human or AI. There might be a slight variation in the proportion of slop to signal, but my prediction is that human slop will prove more durable.
Human slop has been the same for a thousand years. It has tells, and they’re stable. Learning to recognize them is worth more than any CAPTCHA.
The Word as a Weapon
The first tell is word abuse, and underneath it, a manufactured fog. Either the author is confused and the fog is accidental, or the author wants you confused and the fog is the point.
The most common version: taking a word with a clear, shared meaning and redefining it to label something you dislike.
Take “scam.” It has a definition: a deception intended to extract money or value from someone under false pretenses. People use it as a verdict on anything they’ve decided is bad. Crypto they don’t like. A business model they find distasteful. A product they think is overpriced. The word isn’t doing descriptive work anymore; it’s doing emotional work. It signals allegiance. It signals that you’re on the right side of a line that’s already been drawn.
This happens everywhere. “Inflation” used to mean inflating the money supply: literally increasing the number of units, or in the old days, shaving the edges off coins and calling the result full weight. That’s a precise and useful meaning. Now it means rising prices. Prices don’t inflate; they rise. But the language has shifted, the new meaning is what people understand, and insisting on the original definition in a conversation about grocery bills is choosing to generate confusion instead of clarity. The point of words is shared meaning. If a word’s common meaning has drifted, use the common meaning and find another way to say the original thing.
I’ve spent real hours, hours I’d like back, in discussions about consciousness, intelligence, and life, only to discover midway through that the other person had adopted a non-standard private definition of the key term and built their entire position around it. I call it definition masturbation: the intellectual pleasure of a private distinction that generates no shared understanding.
The classic move: “Consciousness is the subjective experience of humans.” Fine, define it that way. But that’s not what people mean by consciousness in ordinary speech, in philosophy, in neuroscience, or anywhere else. Define it that way, then spend an hour arguing that AI doesn’t have consciousness and dolphins don’t have consciousness, and you’ve proven nothing except that AI and dolphins don’t have the subjective experience of humans, which yes, obviously, the definition excluded them by construction. That’s not an argument; it’s a tautology passed off as a discovery.
It’s the same move every time: redefine the word, fight hard about the redefined word, and walk away feeling like you’ve established something. You haven’t. You’ve just made it harder for anyone near you to think.
The Label Instead of the Person
The second tell is subtler and more dangerous. Emma, from Tamers of Entropy who posts on Nostr, described it:
There’s a particular kind of fog people make when they don’t want to look at something. They reach for the ugliest available word and pin it to whoever they’ve decided is the problem. ‘Satanic pedophiles.’ ‘Reptilians.’ ‘Lizard people in suits.’ It feels like clarity, almost like a discovery, but the label is what you use instead of looking.
Watch what the word is doing. ‘Satanic’ is a verdict dressed up as a description. It closes the case before you’ve opened it. The people on the other side stop being people and become a category, and the category has already been judged. After that, you don’t have to think. You get to feel righteous and call it analysis. Usually after watching some rando’s YouTube video stitch random facts into ‘proof.’
Every genocide started this way. Label ‘the others’ with the disgusting word, deny their humanity, then eliminate them.
I’m not defending the people in power. The systems they’ve built select for the worst available humans, and then everyone acts surprised when the systems produce them. Yes, there are more psychopaths in those rooms than in an ordinary sample. Power attracts a certain kind of empty. But almost nobody, including the genuine psychopaths, wakes up asking what evil thing they can do today. They justify. They postrationalize. They tell themselves a story in which they’re the adult in the room making hard choices other people are too squeamish to make. Most of them believe the story.
The other piece of the fog is the 4D chess fantasy. People imagine a small group of brilliant villains executing a centuries-long plan, and they want this to be true, because a world run by master strategists is at least a world that makes sense. Real power doesn’t work like that. Some of them think they’re playing 4D chess; the plan breaks two moves later. Central plans meet counter-plans. Your goal is someone else’s worst outcome, and they have resources too. The present emerges from conflicting vectors of action, half-informed and mostly postrationalized, colliding in a system nobody actually controls.
If you want to criticize power, criticize what’s there. The selection effects. The specific decision the specific person made in the specific room, and the incentive that made it almost inevitable. Reality is damning enough on its own. Adding demons is what people do when reality has stopped being enough for them, and that’s a problem about the person doing the labeling, not about the people being labeled.
Hanlon had most of it. Don’t attribute to malice what’s explained by fear, stupidity, or a person doing what almost anyone would do in the same chair with the same incentives. And don’t attribute to demons what’s explained by humans doing what humans do when nobody is making them look at themselves.
The fog feels like clarity. That’s what makes it fog. Don’t fall for it. When you think you’ve made a discovery about a whole category of people, you are almost always wrong. That’s how the worst things in history started, with a category instead of a person. Look at the person.
She’s describing something you can observe in the wild every day, on social media, in comment sections, in political discourse, in your own head if you’re honest about it. The leap from “I disagree with this person” to “this person is the embodiment of a condemned category” is how thinking stops. It’s also, as Emma notes, how much worse things than bad arguments have started.
Who Gets Through the Filter
I said I don’t want to read most of what’s produced. So what’s the minority that I want to consume?
Writing where the author takes words seriously enough to mean them precisely. Where they can look at someone they disagree with and describe the actual person, the actual decision, the actual incentive rather than just the category. Writing that leaves the subject clearer than it was before, where the author landed somewhere unexpected because they followed the argument rather than drove it toward a verdict they’d already reached.
None of that is a question about whether a human typed it.
Nostr handles this reasonably well. No algorithmic curation, CAPTCHA gates, nor “verified human” labels. You follow whoever you decide is worth following, based on what they actually produce. Content discovery happens through people you trust boosting what they find valuable. Spam exists, a lot of it, but it’s invisible if you haven’t clicked follow. The filter is your judgment about the work.
Emma is from Tamers of Entropy, a story about surveillance, freedom, transcending the limits of consciousness and the edges of what it means to be human. Her Nostr posts are hers.
