Delivery used to be the bottleneck; now an agent fleet can ship a quarter's roadmap in a weekend. That makes Cagan's line between output and outcome the sharpest filter a founder has today.
Marty Cagan wrote Inspired to explain why most product work fails: teams get handed feature lists and are graded on shipping them, while the real risks sit elsewhere. He names four — whether customers will use the thing, whether they can figure it out, whether it can be built, and whether it works for the business — and insists they be tested with cheap prototypes before engineering commits, not discovered after launch. Empowered teams receive problems to solve rather than features to build, and discovery runs continuously alongside delivery instead of as a phase that ends.
Agents collapse the cost of delivery, which sounds like liberation and is actually the trap: a feature factory running at machine speed produces unwanted software faster than any human org ever managed. What agents cannot hold is value and viability risk — knowing whether anyone wants this and whether the business survives it stays the founder's job, and Cagan's four risks tell you exactly where to spend that judgment. His prototypes-over-specs instinct also gets cheaper: have an agent build the throwaway version this morning, put it in front of a real customer this afternoon, and only then let the fleet build it for keeps.
- Grade your agents on outcomes, not shipped features
- test value risk with an agent-built prototype before committing the fleet
- keep value and viability judgment human — you are the empowered team of one.
