I write about perfume, where the genre collapsed into slop years before anyone had a GPT to blame: thousands of people who'd smelled a thing and were performing the act of having-an-opinion about it. The algorithm pays out for a particular silhouette of content, and the moment you can name what it rewards, you can produce it without meaning a word of it. People always did this. What's new is that the shape is the one thing a machine makes better than we do, so it now arrives by the ton.
And the tell isn't the em-dashes everyone hunts for. It's that the algorithm-optimized shape does not trust you to be in the room. It chews the food first: makes the point, restates the point in case the point escaped you, tells you how to feel about the point, then walks you through a door and turns to explain the door. So I'd put the cutline somewhere other than human-versus-machine. The question is whether the writer trusts the reader. Slop is writing built for a reader it doesn't believe is there.
"performing the act of having-an-opinion" - Love that. And yeah, because AI doesn't have original "thoughts" its writing style is heavily verbose and will default to restating the same point seven different ways to fluff your piece and drive false engagement. The optimistic side of me believes we will self select and people will develop a natural revulsion against that over time (as many already do!). But not all platforms will adapt. I hope Substack does. That's part of my job here.
Thoughtful piece — and I agree with most of it. AI is simply a tool, nothing can replace the mind. But becoming overly reliant on AI can certainly make us dumber.
Loved this! It matches exactly what I learned writing about my own AI workflow.
I published a piece recently about how I work with AI like a CEO who hires directors. One sharpens the concept, one edits, one packages. They filter and amplify what I bring, but the ideas, the stories, and the calls are all mine. They just come out clearer on the other side.
The feedback surprised me. Nobody asked how much AI was involved. They said it was clear and useful. That's it.
Which confirms your point: readers reward quality and effort. When the thinking is yours, the tool makes it sharper. When there's no thinking behind it, readers smell it fast.
I would love to screen/block notes and posts which are AI-ish whether by algorithm or consensus. At the moment I just unsubscribe, unfollow and block sources. That may seem unfair but I have a limited amount of attention. The author should have that reflected in their stats.
Thank you for your thoughtful post!
„The trust that matters here is in yourself to use AI without letting it use you.“ that’s one of the best conclusions about AI I’ve read lately.
I write about perfume, where the genre collapsed into slop years before anyone had a GPT to blame: thousands of people who'd smelled a thing and were performing the act of having-an-opinion about it. The algorithm pays out for a particular silhouette of content, and the moment you can name what it rewards, you can produce it without meaning a word of it. People always did this. What's new is that the shape is the one thing a machine makes better than we do, so it now arrives by the ton.
And the tell isn't the em-dashes everyone hunts for. It's that the algorithm-optimized shape does not trust you to be in the room. It chews the food first: makes the point, restates the point in case the point escaped you, tells you how to feel about the point, then walks you through a door and turns to explain the door. So I'd put the cutline somewhere other than human-versus-machine. The question is whether the writer trusts the reader. Slop is writing built for a reader it doesn't believe is there.
"performing the act of having-an-opinion" - Love that. And yeah, because AI doesn't have original "thoughts" its writing style is heavily verbose and will default to restating the same point seven different ways to fluff your piece and drive false engagement. The optimistic side of me believes we will self select and people will develop a natural revulsion against that over time (as many already do!). But not all platforms will adapt. I hope Substack does. That's part of my job here.
Thoughtful piece — and I agree with most of it. AI is simply a tool, nothing can replace the mind. But becoming overly reliant on AI can certainly make us dumber.
Loved this! It matches exactly what I learned writing about my own AI workflow.
I published a piece recently about how I work with AI like a CEO who hires directors. One sharpens the concept, one edits, one packages. They filter and amplify what I bring, but the ideas, the stories, and the calls are all mine. They just come out clearer on the other side.
The feedback surprised me. Nobody asked how much AI was involved. They said it was clear and useful. That's it.
Which confirms your point: readers reward quality and effort. When the thinking is yours, the tool makes it sharper. When there's no thinking behind it, readers smell it fast.
Thanks for putting this into words.
I would love to screen/block notes and posts which are AI-ish whether by algorithm or consensus. At the moment I just unsubscribe, unfollow and block sources. That may seem unfair but I have a limited amount of attention. The author should have that reflected in their stats.