Why sell your soul and make that decision? Why not do both, and try spending half your time creating trivial fart apps - shouldn't take more than a few hours, right? which you need to get away from your main project that's going to change the way I use my smartphone, etc., etc.
If we're offering advice from the peanut gallery (and I'm no smartphone developer, so I'm way up in the peanut gallery), why not spend that time on a program that can generate trivial apps.
Extract data [legally, now!] from some source (eg wikipedia), and skin it with some boilerplate and some generated UI elements. I think it'd be a more interesting project to implement the more general version, and once it's nailed down it's probably a better source of revenue too.
