Why a notebook
Notes in the margin of work that moves too fast
I finish a project and the code stays, with its whole history. The decisions behind it don’t stay anywhere: why I picked this database and not another one, what I took on trust without measuring, which decision from six months ago cost me a week just now. They stay in my head, and my head forgets.
Here I write them down.
No schedule
Writing to a schedule would go badly for me. I’ve tried, and the result is half-finished notes published to keep a streak alive. There’s no cadence here: one month there may be three posts, the next none at all. What decides that a post ships isn’t the date, it’s that the idea held up.
Publishing to a calendar ends up as filling pages. A post ships once the idea has survived the cooling, and not before.
What it’s about
Things I’ve touched with my hands:
- Applied AI, for real. Not what a model does in a demo, but what breaks when the data stops fitting or when the invoice arrives.
- Systems and infrastructure. Self-hosted servers, containers, domains, certificates. Boring until they fall over.
- Decisions and what they cost. Every technical decision is a bet with an expiry date, and we almost never write down how it turned out.
What won’t be here: tutorials for things I’ve never used, and opinions about technologies I’ve only read about.
This is post 01. If you want more, there’s a feed; if you want to argue about any of them, write to me.