I don’t use AI because I want it to think or write for me. I use it because it helps me stay with my own thinking long enough to turn it into something useful.
I have ADHD — mostly the inattentive kind — so my brain doesn’t always move in a straight line. Ideas arrive out of order. Connections show up before the structure does. I’ll know what I mean before I can explain it clearly.
I also deal with what I call delayed intelligence: I often understand what I actually meant five minutes, five hours, or five days after the conversation where I needed the thought.
That gap — between knowing something and being able to articulate it — used to cost me a lot of time and frustration.
AI narrowed that gap.
The blank page was never really the issue. The harder part was catching a half-formed thought, keeping it in view long enough to work with it, and turning it into something another person could follow.
For me, that’s got more to do with working memory than productivity. The scattered thoughts, the threads that disappear mid-thought, the idea that felt clear thirty seconds ago and is now somewhere behind three other ideas — that’s ADHD doing what it does.
AI helps with that. Not by generating content, but by reducing cognitive drag. It gives scattered thinking a place to land so I can actually do something with it.
The real value isn’t prompts. It’s conversation.
I bring a messy idea to ChatGPT or Claude and ask: what am I actually saying here? What’s missing? Where’s the stronger argument? The first answer is rarely the answer. But it gives me something to react to, and that reaction is where my actual thinking shows up. I push back, correct it, reject parts of it, sharpen others. Sometimes the AI finds something I hadn’t said clearly yet, something I’d forgotten, or an angle I hadn’t considered.
That back-and-forth forces structure. Not because the AI imposes it, but because explaining a half-formed idea to anything — even a language model — requires you to give it a shape.
They’re not interchangeable, and I don’t use them that way.
ChatGPT handles research, fact-checking, structure, and stress-testing. If I want a second opinion on an argument or need to pressure-test a claim before I commit to it, that’s where I go. It’s good at examining the idea from the outside.
Claude handles long-form drafting, article flow, tone, and tightening language once the idea has a shape — the iterative editorial work, multiple passes, voice calibration.
For visual mockups and video, pulling YouTube transcripts or summarising long-form content, I use Google’s AI tools.
I still decide what’s true, what sounds like me, what deserves to stay, and what needs to be properly designed.
AI can make unfinished thinking sound finished. That’s the trap.
A clean draft doesn’t mean the idea works. A better sentence doesn’t mean a better argument. AI will organise weak thinking just as readily as strong thinking, and the result looks credible enough to fool you into thinking you’re done.
So I still ask: is this true? Is this mine? Would I actually say this? Did we find the point, or just polish the fog?
That last question is the one that matters. If I can’t answer it, the draft isn’t ready — regardless of how good it reads.
For me, that’s the real value of AI. It gives my thinking somewhere to land, then helps me stay with it long enough to turn it into work I can stand behind.
PS: The em dashes are grammatically correct in case you’re wondering.
If you like this content, here are some more ways I can help:
Cheers!
This article is AC-A and published on LinkedIn. Join the conversation!
