There's a particular pleasure in finishing something small and moving immediately to the next thing. The momentum feels good. The backlog shrinks. You can point to the trail of completed work and feel productive.
But I've noticed that when I move too fast, the learning evaporates. Two months later I can't explain why I made the choices I made. The project exists; the understanding doesn't.
The case for the journal
Writing about what I'm building forces me to articulate things that otherwise stay vague. Why did I use this library instead of that one? What broke first, and what did the break reveal? What would I do differently?
Between Pours is my attempt to answer those questions in public — or at least semi-public — while the details are still fresh.
A note on AI tools
Most of what I build here involves Claude in some capacity. I want to be honest about that. I'm not a developer by trade, and these tools let me build things I otherwise couldn't. That's not a disclaimer — it's the whole point. The interesting question isn't whether AI was involved. It's what I was trying to make, and whether I made it.
More soon.