01
Domain first. AI second.
The moat is the operator who knows where the model should be pointed, not the model itself. Ten years running the problem beats ten weeks training on it.
Principles
Twelve principles I run companies by.
Written down so I don't have to explain them twice — and so anyone considering working with me knows exactly what they're signing up for.
These are not aphorisms. They are the operating principles I've stress-tested across two companies, four Inc. 5000 years, two patents, and a decade inside healthcare. Some of them cost me money to learn. All of them save me time now.
If you work with me — as a co-founder, an employee, an investor, an advisor, a partner — these are the defaults. I'll break any of them for a good reason. I won't break them for a convenient one.
02
Outcomes, not software.
If the customer can't tell you the outcome you delivered without opening a dashboard, you sold them software, not outcomes.
03
Ship in weeks, not quarters.
Five days, four nights, three people is not a stunt; it is the correct pace when the vision is settled and the tools are ready.
04
The chaos is the opportunity.
If existing systems weren't failing, the market wouldn't need a new one. Do not run from chaos. Run into it.
05
Regulated industries are a moat, not a barrier.
Compliance is the cost of entry that keeps followers out.
06
Team small. Taste tight.
Three people who know beats twenty who don't. Hire slowly, ship fast, defend the standard.
07
Build distribution before you need it.
The essay you publish this month is for the person who Googles you the day your product ships.
08
Say no more than yes.
Every yes is a meeting, a project, a distraction. The healthy default is no.
09
Compound in public.
Nothing you do in private compounds. If it matters, publish it.
10
Write to think.
If I can't write the argument down in a paragraph, I don't understand it. Every hour writing pays for itself three times: clarity, distribution, permanence.
11
Long horizons. High pace.
Not urgency versus patience — both. The decade is the horizon; the five-day sprint is the pace.
12
The next version is always the right version.
Whatever we shipped last week is already wrong; attachment to the past version is the tax founders pay for pride.
Coda
These principles will change. Not because they are wrong, but because the problems will get harder, and the principles will have to keep up. When they do, this page will change with them — and the versions will be dated, so anyone who wants to see how the thinking evolved can trace it.