In my case, I graduated and took a job at what I thought was a startup but turned out to be a slow-growing small business. As Marc Andreesen says, "these are great places to go if you don't want your career to go anywhere either." So I quit and founded a startup. It failed, but I learned enough hot skills to prove to my interviewers that I hadn't spent the last year and a half dicking around. Then I got into Google, and once you have Google on your resume, nobody cares that you spent two years working on JSF webapps for a company nobody's ever heard of.
I appreciate my great good fortune to be graduating into the early years of the dotcom boom... But when the NASDAQ imploded I had 3 solid years experience framed on my CV in terms of "I have done X using technologies A, B and C accomplishing result Y". That's what matters a few years out of college and without it it would be very difficult to "catch up" or even to get a foot in the door.
Assuming it never generates enough revenue to get past the "hobby" stage, a few years from now you it will still have value in your resume. Someone who worked on their own project for years is always an attractive candidate for a programming job (at least to us).
Seeking employment at a large company is only one investment strategy and perhaps not the best one right now given the fierce competition caused by the downturn.
Playing devil's advocate here (your suggestion is exactly what I'd do, and it's worked fairly well for me), but...
One reason they don't may be that it's almost never rational to start a new project over sending out a resume. The resume offers the possibility of immediate monetary payoff for relatively little work, and while the chances are low (particularly in this economy), they're certainly higher than the chances that your new project will succeed. The latter has no immediate monetary payoff, requires lots of work, and has risks that are so varied that you don't even have a framework for evaluating them.
It's only when you put a lot of these risks together in sort of a "career portfolio" that it makes sense to take them. Any individual project will probably fail, but together they give you the skills and experience that will a.) get you hired and b.) let you succeed more often with these independent projects.
Alas, in my experience most people don't take such a long-term view of their careers...
Anyway, personal projects are better than nothing but they aren't a substitute for commercial experience, which is about a lot more than just coding. It comes down to not knowing what you don't know. Employed programmers deal all the time with people saying "can we do X" where X is something you'd never considered and might require some research to figure out, and the actual code is simple when you come to write it. People working alone can't help but approach from the angle of "I can do Y, so that's what I'll do".