Personally, I do not know many startup founders. But, the basic underlying principle of success (that I've gathered from news.yc) is "build something people want." So startups are there to provide services. Necessarily, founders do care what other people think.
Academic scientists don't necessarily have that constraint.
Perfect example: the search for extra solar planets. Twenty years ago, Geoff Marcy, a prof at a teaching school, SF state, begged for telescope time. Now he's a prof at UC Berkeley, and (I think) PI of a roughly trillion-dollar space mission.
Marcy is a multiple-sigma deviation. His success is rare. Careers in academic science have a very poor risk-adjusted payoff, worst than startups. Would I rather be the guy who pioneered planet finding or one of the google guys? Don't know. :)
That is a pretty huge caveat. If you're lucky enough to be a mathematician, that's easy to say. If you're a molecular biologist... brace yourself for the sticker shock.
And it depends on your goals. If your goal is to enlighten yourself, you don't need to be an academic: you just need a decent library that gets all the Nature journals, a pile of textbooks, a local university full of grad students that like to talk about their work, and time. But if your goal is to do original work, you've got bigger financial problems.
Anyways, in today's world, I see the real need lying in integration, not more and more specialized research. I think this is reflected in the discouraging stories above.
Open a lab equipment catalog. What do you suppose the availability of grant money to most purchasers does to the prices? Also don't forget that the delivery of anything chemistry-related to a private residence is all but banned under American law.
I think it's no coincidence that startup founders tend to be more of the rebellious types.