A Memory — Before It Had a Name
In the early 2000s, I wanted a PDA.
I'm not certain I ever owned one — but I remember the wanting clearly enough that it might as well have been real. The promise of a device that could keep your calendar, your notes, your contacts. A small screen that was, in theory, the beginning of an assistant. Something that would think with you, not just store for you.
What I imagined was more than what it offered. I wanted it to understand context. To anticipate. To help me reason through things, not just retrieve them. The gap between what it could do and what I wanted it to do was too wide, and eventually I stopped imagining and accepted the limitation.
Twenty-something years later, the gap closed.
A Mirror — What the Work Actually Feels Like
I didn't plan to become anything AI-native. Theo moved in that direction — gradually through 2025, then decisively around end of Q1 2026. The shift happened at work before it happened personally. But a conversation in February with a friend who thinks about AI more seriously than most people I know pressed something loose. He didn't tell me what to do with it. He just made it harder to stay on the sidelines.
What followed was a period of figuring out what this thing actually is, in practice. The word I keep returning to is mirror — not a clone, not an oracle. A mirror that occasionally surfaces angles you hadn't considered, and leaves the work still yours.
I've been working on a poetry book. The process of co-editing with AI produced something I didn't expect: connections between poems written years apart, threads I hadn't consciously built, but which were apparently there. The AI didn't write the book. It helped me see what I'd already written.
This blog exists, in part, for the same reason. I've always had ideas. I've also always had a very effective internal mechanism for convincing myself those ideas weren't worth saying. Imposter syndrome is real — I've carried it since Davos 2023, sitting in rooms asking myself what I was doing there. The resolution I eventually came to was quiet: you are where you're meant to be . Tawakkal. Not naivety — surrender to the design.
AI didn't fix that. But it removed enough friction that ideas stopped dying quietly before I could write them down.
A Slope — What It Amplifies
The first real project I built in Claude was a trading bot.
I spent time on it. Learned things. Then, at some point, paused — and asked the question I should have asked earlier: why am I reinventing the wheel? The tooling already existed. My capital was limited. The real reason I was building it, examined honestly, had more to do with the excitement of building than with any clear answer to what for .
I shelved it.
What that experience left behind was harder to resolve than a paused project: the awareness that AI will follow your lead — wherever you're pointing, at whatever speed you're willing to move. It doesn't ask why. It doesn't hesitate. It amplifies what's already there. A builder with an unclear intent doesn't get unstuck. He gets more efficient at going nowhere.
The guardrail isn't the tool. It's you — your values, your intent, your willingness to keep asking the question the tool won't ask for you.
A Reckoning — The Obligation
That February conversation stayed with me longer than I expected.
My friend didn't tell me I needed to engage more seriously. But there was something underneath what he said — an allusion, not a directive — that I keep returning to. The idea that if enough people who think carefully about these things don't engage, the space fills with people who won't. That stepping back, out of caution or reservation, is itself a choice with consequences.
Fardu Kifayah. A collective obligation — if enough people fulfil it, the burden lifts from the rest. If none do, all bear it.
I still have reservations. Privacy, self-sovereignty, what it means to depend on something this capable. Those don't go away. But I'm increasingly convinced that the answer isn't to stay on the outside. It's to be someone who engages honestly — who asks the hard questions from within, not from a distance.
Cautiously optimistic. That's where I land.
Not because the risks aren't real. But because you are where you're meant to be.
And apparently, this is where I am.
IN.
Previously — Building in Public, Building in Private