Most of my work happens in the gap between a product decision and the thing actually existing. Eight years designing products, three building them — which means I can argue about a flow in the morning and have it running by the afternoon, instead of writing a spec for someone else to interpret.

The tools change, the approach doesn't: get close to the real problem, build something rough enough to react to, then cut until only what matters is left. A working prototype settles arguments that a mockup never will.