I’ve always liked writing. Not professionally, but as a way of thinking. There’s something about the act of putting an idea into words that forces it into a shape. Vague intuitions have to become sentences. Sentences have to follow from each other. By the time a thought reaches the page, it’s clearer than it was in my head, and sometimes it’s different from what I expected.

I blogged for years, back when blogging was still a thing people did before social media absorbed most of the oxygen. That writing was fast, raw, unedited. I’d finish a thought, hit publish, and move on. It was fun. At work, I still write more than my role requires. Most of it disappears into the void, but the act of writing is useful to me regardless.

This blog is a different experiment.

The first change is using AI as a thinking partner during the early, exploratory phase. Steve Jobs called the computer a “bicycle for the mind” — a tool that amplifies human effort. That metaphor still holds for me, though what I’m experiencing feels more like sparring than cycling. When I start with a half-formed idea and subject it to sustained challenge — where are the contradictions? what would a skeptic say? is this actually one idea or two pretending to be related? — the thinking goes further than it would alone. The AI doesn’t have the ideas. But it asks questions I wouldn’t ask myself, because I’m too close to my own assumptions to see them.

The second is adding an editorial step. For all the years I’ve written, I’ve never worked with an editor. Not once. In software engineering we accept code review without question — no serious project ships without someone else reading the code. But prose? I’d write it, maybe re-read it once, and publish. The idea that someone could go through my writing and say “this paragraph doesn’t earn its place” or “you’re explaining this twice because you don’t trust the reader” was foreign to me.

It’s been revelatory.

After one piece went through a proper editorial pass, I learned things about my own writing that I’d never noticed across years of doing it. I repeat myself — saying things twice or three times, slightly differently each time. A reflex born from anxiety about being understood. I also hedge preemptively: adding disclaimers that address objections nobody is making, explaining what I’m about to say before saying it. Once these patterns were pointed out, I couldn’t unsee them.

This process requires more effort, not less.

Being challenged during the exploration phase means I have to defend my ideas — and sometimes discover they don’t hold up. The editorial feedback reveals problems in my writing that I didn’t know existed, and fixing them is real work. Each piece goes through more rounds of genuine revision than anything I wrote when I was writing fast and hitting publish.

An amateur cook occasionally produces something great, but the variance is high. A professional chef hits a consistently higher standard — not through superior ingredients, but through process and accumulated skill. The editorial process is something like that. It raises the floor. It makes the gap between what I intended and what I actually produced visible, and that visibility is where the learning happens.

That learning is what I’m here for. Not the output — or not primarily. If the writing gets better, that’s a welcome side effect. But the thing I find genuinely fascinating is the process itself: discovering that writing has a craft dimension I’d never engaged with, and that iterating on a piece teaches me something I carry into the next one.

In Up in the Air, George Clooney’s character says “to know me is to fly with me” — a line that reveals something hollow about him. But the underlying idea resonates here: to know how I think is to write with me.

The process isn’t without friction. AI can be sycophantic — agreeing with my arguments when it should be pushing back harder. I’ve learned to notice and redirect, but the deeper question lingers: how do I know when it’s agreeing because I’m right versus when it’s being polite? I don’t always.

I’m a few pieces in — not far enough to know if any of this holds. But the early signal is encouraging. I’m catching things in my writing now that I didn’t see a month ago. Small things. But small things compound.

Linus Torvalds’ Just for Fun captures something I believe: the best reason to do something is because you find it genuinely interesting. I find this interesting. I find it interesting that writing has layers I’d never peeled back. I find it interesting that a tool built on statistical patterns can meaningfully challenge my thinking. I find it interesting that the process of making something better can be as satisfying as the thing itself.

The experiment continues.