Context
I have ADHD, and I write with AI. Those two facts are connected, and this piece is about how. Not “AI helped me polish this.” The article you’re reading did not exist as a finished thing until an AI sat down with me and built the structure around it. The usual word for that from the outside is slop, or, aimed specifically at people like me, neuroslop.
TL;DR: My ADHD means some tasks, like turning a scattered thought into a finished piece of writing, put a specific strain on exactly the resource I’m short on. AI tools remove that barrier. For a neurotypical writer they’re a convenience. For me they’re the reason a thought becomes a shipped piece of work instead of a fragment in a notes app forever. Same tool, completely different stakes.
Slop to you, accessibility to me
The slop criticism is usually fair. A lot of AI-assisted writing really is low-effort output nobody reviewed, produced because it was easy to produce. I’m not arguing against that criticism in general, most of what gets called AI slop earns the label.
But the label carries a hidden assumption: that the person could have written the piece themselves, just a bit slower. That assumption doesn’t hold for me. The barrier was never knowing what to say. It was the blank page itself, and the executive-function cost of turning a scattered mental state into the first three bullet points that make a piece of writing exist. That cost isn’t there for everyone in the same way, and it doesn’t scale down with “try harder.” It’s the same shape as needing glasses to read fine print: fine for most of the day, and a hard wall the moment the task demands exactly the thing you can’t do unaided.
The point is narrow: some tasks put a specific, predictable strain on exactly the resource ADHD affects, and a tool that corrects for that strain is doing accessibility work whether or not anyone calls it that.
Three things this actually looks like
Not an abstract claim. Here’s what it looks like in practice:
- Structuring from zero. I hand over a mess, and get back a skeleton I can build on. The blank page was the wall; the skeleton is the door.
- Finishing what attention broke. A thought gets interrupted halfway through, the way thoughts do, and instead of staying a half-sentence in a drafts folder forever, it gets picked back up and completed.
- Turning raw stream-of-consciousness into structure. A chaotic voice note or a wall of unpunctuated thinking goes in, and something readable comes out the other side, without losing what I actually meant.
None of these are about the AI being smarter than me. They’re about the AI correcting the exact point where my own process used to fail.
It’s bigger than writing
This isn’t only about articles. The same pattern shows up across how I work: writing code, getting through a work task, turning a half-formed idea into something that actually ships. In every one of those, the AI is functioning less like a tool and more like an extended working memory, the extra context I can’t hold onto myself. I haven’t measured this rigorously across those areas the way this project measured the reading-side research (see the neuroinclusive writing checklist built for this same content pipeline). The pattern is consistent enough across code, work, and writing that I’m confident it isn’t a writing-specific quirk.
Why the slop label sticks right now
Right now, anything tagged AI-generated gets treated as slop by default. That reflex says more about where we are than about what the tool is. Two kinds of literacy are still immature: creators who reach for the tool without doing the thinking it still requires, and readers who haven’t yet built the pattern-matching to tell that apart from someone using the same tool to get a real thought past a real barrier. Both sides are new at this. The label is a snapshot of an early, poorly-understood moment, not a fixed property of the output.
The honest part I won’t dress up
The AI drafts real sentences. Whole paragraphs, including ones in this article. Pretending it only “helps with structure” while I supposedly write every word myself would be its own kind of slop, a lie about how the thing actually got made.
So here’s the real division of labor. The AI is fast hands. What makes the piece mine is everything around the drafting: the direction, the judgment about what’s true and what’s off, and the dozen rounds of “no, that’s not what I meant” that reshape a draft until it says the thing I actually think. This exact article went through that, more times than I’d like to admit. The argument is mine. The calls about what stays and what’s wrong are mine. The typing, increasingly, isn’t, and I’ve stopped thinking that’s the part that mattered.
There’s a meme I like: a boss riding in a cart, a leader pulling one alongside the team, an introvert dragging the cart alone, and then “ADHD introvert with AI” flying a fighter jet. It’s a joke, but it lands for a reason. I’m not pretending I’m running on foot next to everyone else. I’m flying something. Flying is still a skill. It’s just a different one than running.
The close
Call it neuroslop if you want. What reads as low-effort from the outside was, for me, the thing that got a real thought out of my head and onto the page at all. A barrier removed, by someone who did the part that was actually mine to do.
