Why I keep building tiny AI apps

A short reflection on the value of small experiments, why I publish them as subdomains, and what I learn by making things no one asked for.

There is a special kind of learning that only happens when you ship something that doesn't need to exist. No deadline, no stakeholder, no roadmap — just a question you can't stop thinking about and a Sunday afternoon.

For me, those questions usually involve AI in some form. Not because AI is the only interesting thing to build, but because it is the most surprising collaborator. It breaks in ways I don't expect, succeeds in ways I don't deserve, and forces me to clarify what I actually mean by 'works well.'

Each subdomain is a notebook I happen to have left open on the internet. DSPView taught me about audio and UI feedback loops. Tune taught me about embeddings and taste. Tram taught me that transit data is a small novel. Texas42 taught me that games are mostly about pacing and expectation.

I publish them because unfinished, thoughtful work can be more useful than a polished resume. If a potential employer lands here, I want them to see how I think, not just what I have shipped.

The next tiny app is probably already forming. I don't know what it is yet, but I know it will start with a question, a few bad sketches, and a domain I buy before I am ready.