It was mostly "friends or people who thought the idea was cool". The handful people who were most adamant about the idea were somewhat savvy Twitter users, so I envisioned others having the same passion for the idea.
Ways to deal with this sad fact:
* Ask users "how much would you pay for this?" Follow up with "seriously, if I build it, you'll commit to paying that much?" Tends to bring them back to reality, but doesn't make sense for all products.
* Look for external evidence for user pain. If there is no tool on the market for doing X, but users manage to sorta-kinda cobble together four other products to sort of do X poorly, you can believe that they want X.
* Build quick little prototypes, as cheap as possible, and see what users actually do with them. (You've just done that -- good job :))
This can be overcome by explaining to them up front, that you are about to invest your soul in the effort and that they will be doing you a disservice if they are not honest and that failure later will be harder on you. When they realize that they are actually saving you a larger amount of future grief then they tend to be more honest.