The author misses a step in the engineering loop that’s important though. Plans almost always involve hidden assumptions and undefined or underdefined behavior that implementation will uncover.

His whole point is two mental models and a model delta. Exactly what you just described.