Writing
Notes on building AI and SaaS products: how I work, working with founders, and lessons from my own studio products.
The spec comes before the code
Why a sourced, locked spec beats jumping straight into code with AI, and how a requirements brief plus deep research makes AI coding faster and cheaper.
Read
Humans keep launch authority: reviewing AI-built releases
When AI writes code faster than teams can review it, approval becomes the bottleneck. How BlastRocket keeps people in charge of shipping.
Read
I build it, or I lead your team: choosing how to get to live
The two ways I take a product to live, building it myself or leading your team, with the trade-offs and how to choose between them.
Read
Compare, don't diagnose: designing health software families trust
How I design health software that shows change over time without alarming families or overstepping clinicians, using Together Through as the example.
Read
The rules my coding agents work under
The written rules my AI coding agents follow: no production data, every change through a preview, evidence for every claim, and people making the calls.
Read
What selling a startup taught me about building one
Lessons from selling Stacks to EBSCO: focus, documentation, owning your IP and building a product someone else can run.
Read
One app, many chapters: multi-tenant design for programs
How I built Mentor Intelligence so independent mentor program chapters share one app and one database without ever seeing each other's data.
Read
Test every preview with real users
User testing at AI speed: real users try every preview on real tasks, fixes go worst first, and shipping loops back to testing instead of ending it.
ReadPage 7 of 8 · 60 articles