Roadmap decisions rather than dates

Why we picked it · the editor's summary

Larson's example is a credit-card application redesign that was two weeks of work and two months of calendar, because a dozen ambiguous decisions sat unresolved along the way. His claim is that with AI tooling, empowered teams, and cheap prototypes, execution is no longer the constraint, so the roadmap should be an ordered list of decisions and the internal question how fast good ones get made. External dates stay, for coordination, and stay conservative. The hedge is real: this works with strong risk management and experienced technical oversight, and he says so. We would push back on one thing. The two-month version had a schedule; the decision-first version needs a reviewer with authority, and that person is the scarce input.