There are two kind of ideas for startups: 1) There is a problem, and current solutions are ugly. Think how to fix it and sell your fixture, and you have an ide

There are two kind of ideas for startups: 1) There is a problem, and current solutions are ugly. Think how to fix it and sell your fixture, and you have an idea. Many programmers when faced with some problem will quickly solve it with some script/whatever magic and not even consider that they could turn that solution into a service/product. How to get such ideas: use more stuff, always look out for problems / poor experience. 2) Those ideas that start with "It'd be cool to do X". Think twitter. I've no idea how to filter out good ideas of this form from bad ones, but with enough iteration I think a lot of even dumb ideas can fly. I've started game dev recently. http://chopperdefense.com http://chopperdefense.com is my first game. I did not know what kind of gameplay or graphics I wanted when I started it. There even were no helicopters in the game initially. But with enough iterations I think I made a pretty decent game. While I was doing it I got a lot more game ideas which I duly noted and also iterated on. "To a game developer, everything looks like a game". Top idea in your mind, as PG says.