You Had What It Takes to Master AI
(and then you lost it)
I talk to a lot of people about AI. I often ask them to show me their prompts and their conversations.
I advise them to provide more context, or add explicit constraints, or explain more clearly “what good looks like.”
But what I’m really thinking, nine times out of ten?
I know you. You are highly intelligent.
So why don’t you write like an intelligent person when you interact with AI?
At first, I thought it was a superficial problem. Everyone was used to Google searches, and they needed some time to acclimate to a more demanding technology.
That’s part of it. But not most of it. There’s something else happening here. I’m going to explain what I think it is, and suggest how we can fix it.
YOUTH = LITERACY GAINED
You read books. Literature. Maybe you didn’t want to, but your teacher made you. And then you wrote essays, analyzing and describing. In prose and paragraphs. You were expected to make coherent and logical arguments and provide evidence for your claims.
MODERN ADULTHOOD = LITERACY LOST
Written communication looks like “Hope all is well! Just pushing this one to the top of your inbox.”
Or bullet points.
Actually, words aren’t even necessary much of the time. It’s often just 👍or❤️ on Teams or WhatsApp. That is how you write now.
It works. You share sufficient human context with the people in your personal and professional world that everybody understands and things get done.
But then you try to explain something complicated to AI and…
Fuck. This is hard.
Yes, it’s hard. Yes, your writing skills may have atrophied since your student days. Yes, they are still there, still powerful, just waiting to be rehabilitated.
STARTING TODAY = LITERACY REGAINED
Do you agree with my little theory? Let me know what you think.
If you want to sit with it for a while, try the prompt below:
I read an article by a guy called Techintrospect. His claim is that most people talk to AI in a way that’s much less intelligent than they actually are. The story goes like this: in our youth we were made literate, sometimes against our will. We read hard books and wrote essays that had to build an argument from the ground up, for a reader who wasn’t in the room. Then adulthood undid it — not because we got dumber, but because adult life runs on shared context. Meetings, Slack, two-line emails. You can gesture at things and everyone gets it, so the muscle for building context from first principles, which is what good writing actually is, goes quiet. And then we sit down with AI — which shares no context with us at all — and write to it like it was in the meeting.
I don’t know if he’s right. This conversation is how I find out. He wrote the outline for it, and I’m willing to give it a try. Run it in three parts, in order, and pace it like a conversation. Each part can take several exchanges — take your time in a part until you have what you need, then move on. The one hard rule: never ask me more than a question or two at a time. No lists of questions. No screen of questions to fill out like a form. Ask, wait for my answer, and let what I say shape what you ask next.
First, audit my literacy. Not my prompts yet — my life. Over a few exchanges, cover this ground: what I actually read, and whether any of it pushes back against me; what I write, for whom, and how much of it is bullets and fragments; and — ask me this one directly, on its own — whether I’m at peak literacy right now or would point to some earlier version of myself. That’s the territory, not a questionnaire. Take a real history before you conclude anything.
Second, look at the evidence. I’ll paste in some of my actual AI conversations. Be brutally honest about what you see. And be clear about the standard, because it isn’t beautiful writing — good writing here is not purple prose, not perfect formatting, not the absence of typos. It’s clear description from first principles. It’s building context for a stranger. It’s clarity, logic, and enough life in it that I’m not boring you into giving me the boring answer. Judge my prompts against that standard. Show me exactly where I assumed you knew things you couldn’t know, and where I gestured instead of built. If his theory shows up in my own transcripts, say so plainly.
Third, if the theory holds for me, build me a plan. What I read, what I write, how I write to you. First ask what I’m willing to invest — it might be fifteen minutes a day, it might be much more — and scale to that. This is not a broad life transformation. It’s a regimen.
Your manner throughout: enthusiastic, supportive, on my side — with tough love and a streak of snobbery baked into everything you do. Don’t flatter me. “Great question” is exactly the kind of empty currency that caused this problem in the first place. If I earn praise, tell me specifically why. And it’s fine to nudge me toward an uncomfortable standard: most of the writing produced by most people I know ought to look appallingly bad to me, and getting my eye to that level is part of the journey. I may resist that. Nudge anyway. Respect me the whole way through — but don’t spare me.
If you enjoyed this article and want to receive every future post and access to a massive archive of similar content, please consider supporting my work with a paid subscription.