Yeah, that would probably work as long as I don't have tough competition that can work full time. So when I release, I have to release pretty complete and bug free, and it'd help if the product was very hard to reproduce without knowing the internals. I can probably do those, since I currently have a year and a half of comp sci research ahead of me (getting a masters).
Any recommendations on especially pertinent fields? I'm planning to look at SOA type stuff - intelligent agents, evolutionary algorithms, systems architecture, etc. I'd like to throw some Godel in there too, but can't really see the applicability yet.
I'm not sure that the "pertinence" of the field is what's important for a start-up. I'd be more concerned about the need for something (unless that's what you meant by "pertinent".)
If you have to look for something to do or ask for recommendations, you'll probably never have the passion needed to see it through.
OTOH, you probably already have what you need. What really bugs you? And others you know? When's the last time you said, "You know, someone really ought to make <some thing>." THAT'S what you should be pursuing.
A deep solution is better in my situation than scratching some random itch, since anyone can do the latter. So, to rephrase my question more exactly: what are the deep underlying problems of the web that need the type of strong theory I can learn at university?
In answer to your last set of questions, something that really needs to be made is an aggregation service for all the web 2.0 stuff out there. People want an integrated, secure, and trust based work/socializing platform, and my current research interests are geared towards solving this kind of problem. Facebook is trying to do this, but I don't think they have a good mechanism yet for integrating apps.
Any recommendations on especially pertinent fields? I'm planning to look at SOA type stuff - intelligent agents, evolutionary algorithms, systems architecture, etc. I'd like to throw some Godel in there too, but can't really see the applicability yet.