Nothing tells you whether a product will be a hit quite like seeing a user react to a prototype. Suddenly the conversation switches from hand waving to practical suggestions. Prototypes reveal the unspoken requirements quickly and effectively. The faster you get to a working prototype, the higher your chances of success.
Only the engineers can build the product, and they can only build the optimal platform for that product if they understand the long term direction. Hiding behind a “just fulfil the spec” philosophy will lead to problems implementing new features in the future. And, yes, prototypes should be built to be thrown away; otherwise the team will hold on to them too hard and not hear constructive criticism.
Outcomes can be prototypes, early versions and ultimately final products.
Iterative and agile.