A studio for the questions we can't stop thinking about.
Hypoth holds a small group of AI products. Each one tests a single hypothesis about how people work, learn, or read — written down before any code, kept honest after the launch.
The interesting AI products won't come from people guessing what AI is good at. They'll come from people asking specific questions and being willing to be wrong in public.
Hypoth is the operating model for that. One sentence hypothesis per experiment and an extensive vision based on that hypothesis written before the code. A defined window with a defined success bar. A continuation note if it holds, a post-mortem if it doesn't.
Every experiment begins as one written sentence committed enough to test against a defined bar. If we can't write it, we don't build it.
One hypothesis per product. Strong opinions about who it's for and what success looks like. No platforms, no suites.
Whether the hypothesis held or didn't, we write down what happened. The honest post-mortem is part of the work, not a footnote.