The two hours I spent making something nobody asked for

I spent a whole Saturday building a small thing for myself. A little tool that solved an annoyance in my own workflow.

Nothing grand. No plan to release it. No thought about whether anyone else wanted it.

It is now the thing I use most. And when I mentioned it in passing, three people asked for the link.

That is the pattern I keep forgetting. The best things I make come from solving something that genuinely bothers me, not from imagining what other people might want. When I build for a hypothetical user, I get cautious and generic.

When I build for my own irritation, I get specific and useful.

The problem is that building for yourself feels selfish, and building for a market feels responsible, so I keep doing the responsible thing and getting boring results.

Do you build for yourself first, or for the user you imagine? And which one has actually produced your best work?