The flipside is that a lot of developers don't have enough hobbies to draw from with which to start a business. One thing we can all build is a developer-focused tool, but we are also some of the hardest customers to please, since we would only pay for something that is sufficiently difficult and useful that we wouldn't want to just build it ourselves or look for an open source solution.

This is something I have had to tell myself - go out and have a life as if you've already "made it", then when you find something you can fix that you yourself would prefer over the alternatives, that's your startup. Unfortunately, I think that means we have to be patient about ideas - you're not going to notice the real problems in a given area after having tried it once (unless your solution is a newbies guide or something).