This is an interesting problem I've been faced with recently, a problem to which I yet don't have a good answer. I feel you guys must have run into it before and would be able to advise me best.
Say you're working with other people, and you're trying to come up with cool ideas / features for a startup. Lots of ideas come up, and a good chunk of the team seems to like a particular concept, even if you think it's not that great. In fact you feel that based on all you know and your personal experience, the idea is pretty doomed to failed, but the rest of the people are really excited about.
You now have the option of either spending hours debating people about why their idea sucks (potentially coming off as the "uncreative guy" who's all about curbing people's enthusiasm / destroying ideas and not really gaining their favor by crushing their dreams) or you shut up and let them go on with it onto the development / prototyping phase, potentially wasting months and implementation on something dumb. You don't want to be the guy who's no longer invited to discussing ideas cause he'll just rip them all to shreds, and so your opinion will no longer be considered.
What's the best way of handling this? I'm all about thinking outside of the norms, challenging conventional wisdom and letting people brainstorm their... guts out, but at some point you need to really see what makes sense, as it will cost a lot of time and effort to attempt to turn that idea into reality.
I'm not sure this helps you much with your situation. In retrospect we should have spent the time doing market research on some of our other ideas which might have been more commercial. Perhaps this is something you can do with your pals? Follow Steve Blank's advice from "Four Steps to the Epiphany" and start talking to your potential customers before you do much serious product development.
While I agree with steventruong that no-one can predict what will make money, there are psychological factors to consider: if you really think the idea is a waste of time, you won't fully commit to it and that might make it more likely to fail...