The Studio Method
How the studio actually works.
Not a venture studio. Not a consultancy. Something built for the specific problem of incubating software that lives next to AI.
What we look for.
A pattern, not a category. We start with a real friction point that an AI workflow keeps hitting, then ask whether the right answer is a piece of standalone software — not a feature inside a model, not a prompt, but a tool that earns its own existence.
If the answer is yes, we incubate it. If the answer is no, we write about the friction instead and move on.
How we build.
Two to four people, working internally, on six- to twelve-week cycles. Real users in the first month. Documentation is part of the deliverable, not an afterthought.
We respect deep work hours. Most days have at most one meeting. The studio runs async-first because the work needs uninterrupted time to be any good.
What happens after.
Some projects stay in the studio as long-lived utilities. Some spin out as their own ventures. Some get open-sourced. Some get retired with a thank-you note and a public post-mortem.
The studio is the parent in every case. Spin-outs carry a small ownership stake back to the studio so the next incubation has runway.
The incubation lifecycle
From friction to ship to spin-out.
A friction point in an AI workflow that keeps showing up.
A small team. A sharp tool. Real users inside a month.
Stay, spin out, open-source, or retire with notes.