The site for The Mothership with Daphne Oz launched Oct 1. About two weeks of design and development, and five days from first commit to launch, built with Claude Design, Claude Code, Sanity, Flightcast, CookieYes, Resend and Netlify. Here is what that took, and why it wasn’t possible a few years ago.
I never spent a dollar marketing my own company. A decade in, every new client came from the last one. Reputation compounds while you sleep, and it cuts both ways.
“You don’t need to know how it works, just what it does” is the fastest way to become useless in this field. In the AI era, judgment goes stale in months. My rule: no opinion on anything I haven’t used with my own hands.
I’ve signed the front of a paycheck and the back of one. The biggest difference between running a company and running tech inside a big org isn’t the money, it’s how a decision gets made.
A tooling committee took four months to approve something I’d already been running productively for four months. The gap wasn’t bureaucracy, it was two different questions treated as one.
Six weeks, a working build, and almost zero usage. The lesson wasn’t “talk to your users.” It was that I preferred my imagined user to the real one, because mine agreed with me.
After 10+ years running my own business, the hard part of a corporate leadership job wasn’t having a boss. It was learning that a good idea has to survive other people’s calendars.
Three "best model ever" launches in ten days. If you re-evaluate your whole stack every time, you never ship. Decide once how you test a new model, and every release becomes a 30-minute experiment.
· 1 min read
Something you’d like me to write about? Let me know.