Writing

Writing

Notes on building with judgment: how I measure it, why it matters, and what holds up under scrutiny.

• June 30, 2026 · 5 min read

  I told companies not to build software. Then building got cheap.

  In 2020 I argued that unless software is your technological base, you should buy SaaS, not build. The framework was right and still is. The reason behind the advice is what AI inverted: the cost of building collapsed, and with it the safe default. Build vs buy stopped being a cost question and became a judgment one: what deserves to be owned now that owning is cheap?

• June 30, 2026 · 6 min read

  Building is cheap now. Here's how I measure what isn't.

  AI made building fast and cheap, which makes 'we shipped fast' a useless claim, because fast is meaningless without how hard. So I measure build velocity and software complexity straight from the git history, with a rubric anyone can reproduce. The result: a platform that took roughly 19,766 hours the old way now reaches the same complexity in about 168.

• May 1, 2020 · 11 min read

  SaaS vs in-house software development

  A 2020 piece on a question most companies get wrong: is software your technological base, or just a tool that supports it? The framework still holds. The default answer it reaches, buy SaaS instead of building, is the one AI later inverted.