A specification is a guess dressed up as a plan. A prototype is a fact. The fastest way to learn whether an idea works is to build the crudest version that can fail honestly.
An overnight prototype is not about cutting corners. It is about compressing the distance between a question and an answer.
Most ideas die on contact with a real prototype, which is exactly the point.
Cheap Failure Is the Feature
The value of building fast is failing cheaply. A bad idea killed in a night cost you a night.
Teams that prototype relentlessly are not reckless. They refuse to fall in love with untested ideas.