This site has robots.txt. It has llms.txt. It seemed only right to also have somewhere that makes the human contribution legible — not in the technical sense, but in the authorial one.
The name humans.txt is in deliberate dialogue with the other two. All three exist here. That completeness says something without requiring a manifesto: human contribution is worth marking explicitly, in a world where it’s increasingly easy to make it invisible.
What it isn’t
It’s not a blog in the conventional sense — posts for SEO, update announcements, that kind of thing. It’s closer to a working journal: reflections on the process, things that surprised me, positions that are still forming. Essays occasionally. Short notes probably more often.
The attribution question
When content is the product of a session rather than solo writing — which will often be the case — attribution matters. The > **[Jeff]:** / > **[Claude]:** convention already used in design documents in this project is the model: dialogue made visible, neither voice collapsed into the other.
The mechanism: I write something. The AI responds within that session. The result is committed together. The asymmetry is honest — I write first, the AI responds — and that’s fine. It reflects what’s actually true about how this work happens.
Why this matters
This project has a position on human-AI collaboration — not stated as a thesis but embedded in every structural decision. AI coding assistants as first-class users. Sanitization before any AI-facing surface. Documentation structured for both human and machine readers.
humans.txt is the last piece of that: making it visible that there’s a human here too, with a perspective worth reading, who isn’t trying to disappear behind the tooling.
The goal isn’t to argue for that position. It’s to demonstrate it in the way the site is built.
The format for this section will emerge from a few real examples. Designing the container before knowing what goes in it produces containers that fit nothing.