But that's not how it works.
For web and mobile products, the act of building or prototyping is where you do most of the learning. I believe that you can't come up with great ideas in a vacuum. You build some mocks, you write some code, and you discover a bunch of things that you could improve on. That's how you come up with the 10x improvements that startups need to succeed. reMail was a great example for this: It took 3 iterations around the same space to come up with a product that got attention and users - it was almost a year into the company when it hit me that you could store and search all your email on your iPhone. No one talks today about the two products we built before the one that succeeded.
Ideas are easy[*]. Give me some smart friends to brainstorm with and we'll come up with several great ideas today that will be $100M+ businesses in a few years - and a lot that won't be. It's about finding the good ones and the ones that you're passionate about working on for a long time.
There's another common thought I hear voiced often: That you have to come up with an idea that no one else has done before - a green field idea. That's wrong. Competition in a space is often a sign that there's something valuable there. Competition is energizing. A lot of the ideas I'm thinking about exploring already have established players or startups - it's just that I think I can execute much better than them.
[*] This used to say "Ideas are dime a dozen" but that upset some people and I changed it to a phrasing that more accurately states what I mean. In the next sentence, I added "- and a lot that won't be" for the same reason.