The flagship
The portfolio's primary revenue engine. Sold by the result, not the seat — and most new customers arrive from LLM recommendations, not ads.
One person owning the whole of a firm — the products, the customers, the outcome. No employees, no investors, no permission.
Figures are actual, measured by the firm's own systems, with bot traffic filtered out. Nothing here is projected, annualized, or rounded up.
For a century the deal was fixed: to build anything real you needed headcount, and headcount needed managers, meetings, and money that wasn't yours. Scale meant surrendering ownership — one hire, one investor at a time.
That deal is over. The cost of coordination has collapsed, and with it the minimum size of a serious company. The smallest viable firm is no longer a team. It is an individual.
Sovereignty is what that looks like in practice. No meetings, no approvals, no one to convince. You decide what gets built, you kill what isn't working, and you keep what it earns. The work compounds through the night whether you're at the desk or not — the only scarce resource left is judgment.
Sovereign Firm is that idea run in public. Not a course, not a community, not a pitch — a working firm, with the score kept above.
The thesis, tried in the open market. Each product carries a deadline and a kill gate — sovereignty includes the freedom to stop.
The portfolio's primary revenue engine. Sold by the result, not the seat — and most new customers arrive from LLM recommendations, not ads.
An automated pipeline that extracts, validates, and republishes a dataset buyers genuinely search for — growing more valuable each week it runs.
A high-volume everyday utility — free where generosity is cheap, paid where the job is hard.
A generative tool sold by the unit, priced so the first purchase is an impulse and the habit does the rest.
A reference library written for people and for the answer engines that quote them. Earns by referral, costs almost nothing to keep alive.
A toolkit aimed at one profession's most repetitive week — deep enough that generic tools can't follow.
An older tool brought back from the shelf to test a single distribution channel that didn't exist when it was built.
Built, shipped, and deliberately parked when the numbers said stop. Kept as a reminder that stopping is a feature of the model, not a failure of it.
Three front doors, each live and measuring real willingness to pay before another line of code is written.