Back when I was running my own consultancy I had this belief that if I made something "cool" then it would become a hit almost (as I understand now, in retrospect) by magic. And it was easy to just move on to something else when the magic didn't work.
I've spent some time working in a bureaucratic organisation for nearly 3 years now and, despite all the obvious downsides, this has been a positive experience because I now think I have more of a "business head" to think with. (The regular salary helps a lot too!)
Coming up with ideas for web apps is, I think, fairly easy for the the type of person who hangs out here on HN. Not all of those ideas will be good, of course, but one in ten might be. It's probably also very easy for you to make the damn thing.
But the only way to ensure you give an idea its best shot is to really stick at it and spend a lot of time and effort on the things that don't traditionally come easily to us programmer-types: marketing, design, sales, customer support.
Disclaimer: I make hardly a penny from stuff I've made in the past (and I've abandoned more ideas than I care to remember) but I'd like to think that my work experience (and the experiences shared by fellow HN members) will help me ensure my next venture is more successful.
You may not like this answer, but the actual hacking and coding is a very small part of launching an independent software company. Your "marketing cofounder" is really in for more like 90% of the work. If all you want to do is code, well, you're in competition with millions of coders worldwide who are willing and able to work more cheaply.
patio11, one of HN's most respected members and best contributors, makes his living on a product that he himself characterizes as "hello world hooked up to a random number generator." His business building was all in figuring out and optimizing AdWords and email responders and the payment pages and logos and colors and so on. Does that truly hold your interest to dive into?