To me, the greatest advantage of being a CS professor, is being surrounded by lots of smart, motivated colleagues & students. The greatest disadvantage -- is that much of research appears to be about doing "new" rather than "useful" things, and funding is governed by NSF/DARPA rather than market demands.
YCombinator seems to have this figured out -- put yourself in a position where students who want to launch startups come to you. So, question is: how do you start something like this?