I’m a fan of metaphor and productive ambiguity, and so I like this improv approach to product development.
Some improv scenes are initiated with a generic line and performers extract the game organically. e.g. "I can't believe it's midnight" is an intriguing start to a scene but there's no obvious game. In contrast, some improv scenes are initiated with strong game right away. e.g. initiating the scene with "No, you're an accountant, you can't just become a lion tamer". Both ways can lead to hilarious scenes.Source: Your product is a joke | The Paperclip
Likewise, some products are initiated with a rough idea. This is in the camp of Eric Reis' model, where you’re lean, get feedback, and iterate quickly. The idea is to treat the path to product market fit as a series of experiments with hypotheses. In contrast, there is Keith Rabois' model, where you have a strong vision from day 0 and not much changes from then. The idea is that you have a master plan from the start, and you get heads down on executing it. Check this post by Casey Winters comparing these models with far more nuance.