2026
When building is cheap, the decision is the product.
From Decide Then Build — the operating sequence for the AI era.
AI made execution cheap without making deciding easy. Teams can now generate interfaces, agents, and decks in an afternoon. The failure mode moved upstream: shipping the wrong thing, faster.
Decide Then Build is the sequence I use on HealthTech and EdTech work. Name the decision. Name the constraint (clinical, regulatory, ethical). Then build the smallest artefact that would change that decision. Not a pitch. A working loop.
The books on the shelf walk the same argument through real ventures. They go live when a Gumroad URL is published — not before.