I’m going to use the structure I hate to tell you about the structure I hate:
The problem isn't the structure itself, it's that AI uses it all the time, everywhere, for everything.This pattern has been making me want to delete the entire internet for the past two years.
Once you notice it, it feels omnipresent. LinkedIn posts, corporate blogs, here on Substack, over and over. Even in my own drafts, when I’m not paying attention, like some kind of linguistic virus I picked up from editing too much AI slop.
I’m not gonna give you real examples because you know what they look like. I’d rather spend time explaining why this keeps happening and give you some quick fixes so you can move on with your day.
First, just to be clear: you can totally use this structure when you’re writing. People have used it for a long time to fix a wrong idea, make a real point, or clarify something confusing.
AI doesn’t do that. It uses “not X, but Y” because it sounds smart. And when you see it three or four times in a short article, you know a human didn’t think about what they were saying. The repetition ruins the reading experience.
Blake Stockton nailed it in his post: AI “tries to sound sophisticated without being sophisticated.”
AI uses it because it sounds like you’re being thoughtful without thinking. It feels nuanced but says nothing. And the pattern is everywhere in the training data (academic papers, TED talks), so the model learned that this is how you sound smart.
You’ll see it in different versions:
The em dash version is another pattern people complain about, by the way. But it doesn’t come close to being as annoying as these repetitive structures. In this case, it’s just standard punctuation — people use it all the time, and it doesn’t interrupt your reading the way “not X, but Y” does.
If you’re using AI to write, you have three options.
You can try telling the AI to avoid this pattern. Blake Stockton has a couple of solid prompts you can use:
"Avoid any sentence structures that set up and then negate or expand beyond expectations (like 'X isn't just about Y' or 'X is more than just Y' or 'X goes beyond Y')"“Avoid any sentence structures that set up and then negate or expand beyond expectations (like ‘X isn’t just about Y’ or ‘X is more than just Y’). Instead, use direct, affirmative statements. Feel free to be creative with your sentence structures and expression styles.”Fair warning: prompting doesn’t always work. Save it to Claude’s memory or ChatGPT’s custom instructions if you want, but don’t be surprised when it shows up anyway.
The pattern runs so deep in the training data that you might need to repeat the instruction every single time you generate text.
Add it to your editing checklist. Searching for “just” will catch many instances, but you also need to look for:
This method is annoying, but it’s reliable.
If you’re a human writer trying not to sound like a bot, the fix is simple: be specific.
You can still hear the pattern, sure. But at least now it’s saying something real instead of corporate nothing.
The “not X, but Y” structure isn’t some writing sin you need to confess. But AI’s mindless overuse turned it into a red flag for lazy content.
If you’re going to use it, make it mean something. Make it specific. Make it yours.
Or to put it another way: it’s not about banning this structure, it’s about using it like you actually have a point to make.
(Ugh. Still annoying.)