Everyone who says the investment is worth it if you get a "hard science" degree should be careful. Physics is as hard as science gets--but according to Statistics Canada, un- and underemployment amongst holders of Physics B.Sc.'s is equivalent to that amongst holders of B.A.'s in English.
And I would disagree with the assertion that the university environment is the only, or even the best place to do science. On whose faculty did Darwin serve? Which endowment funded Einstein’s work on Relativity? One could make a convincing case that 'gentlemen naturalists' of a type a mustachian would be eminently suited to become have had at least as much influence on the advancement of science as all the university professors put together.
And why shouldn't they? A professorship was originally supposed to be a teaching position. That's what Professor means: teacher. Even Doctor comes from the same root as educator. (drop the e, and you'll see how that happened.) Even if you're ignoring your teaching function, in academia, you spend more time on grant money than on science, and the science you can do is determined by the politics of the funding agencies.
Why, then, go to college? They're the gatekeepers of a great many professions, 'tis true. If you simply must work in one of them, fine, by all means, attend a University and then attend to the financial burden. If you're here, though, you're probably not going to weld your self-identity to your wage-earning, and you should find a trade that gives you the best ROI instead. (Welding metal is much more satisfying than welding self-identity, anyway.) If you are the intellectual sort, you can still work a trade while you pursue higher learning as a hobby, and once FI, audiodictate yourself to a PhD-level knowledge of any subject you choose to.
No offense, but your view about science is a bit naive. Darwin was a Cambridge graduate, Einstein studied physics at ETH Zurich. It's true that his first influential papers were published while working for the Swiss patent office, but after that he held professorships in the best universities of his time and was of course paid for his research. Anyway times are different now, you'll never get to do a serious research in chemistry, physics, molecular biology and similar areas without a formal education. Scientific research is a lot more expensive these days than in late 19th century and requires a lot of machinery. Fundamental research is mostly done in academic and government funded institutions, applied science mostly in industrial labs (with ample interface between the two). It's not just about educating yourself to a high level, but also about being around many other similarly educated people. Great ideas rarely develop in vacuum!
P.S. I'm not Einstein, but I do scientific research for living (industrial) so I think I know a bit what is going on. :)
At the risk of derailing the thread, I'd like to respond in detail.
One: you no longer need to attend a university to get a the equivalent of formal education; you just need self-discipline. The textbooks, the course-notes, the assignments and exams for full degrees in all the sciences are available on MIT OpenCourswear and other places. I'm not Enstien either, but this view is shared by Gerard 't Hooft, Nobel laureate in Physics, amongst others far above us in the Ivory Tower.
Also, the courses Darwin took at Cambridge was laughably unrelated to the sort of naturalism that made him famous, if you look into the curriculum of the time. This is not true for Einstein, but since much of his best work was done in a patent office on his own time, the fact that he later became a household name and worked at an Ivy League institution is a bit of a red herring, in my opinion.
As for "being around many other similarly educated people.", I never had any of that in grad school-- nobody talked outside of their tiny research groups, because they weren't interested beyond the laser-beam focus imposed on them by their funding committees. When we did get together, it was to forget about our damned theses, not to celebrate them. Furthermore, I'd suggest the internet can take over in this regard, as well: I got a lot more comradeship and assistance from the PhD Comics Phorum (sic, it's a pun) than I ever did anyone local.
"Scientific research is a lot more expensive these days than in late 19th century and requires a lot of machinery. "
This depends on what you're doing. If you're interested in Fusion research, you can be on the cutting edge and build a polywell inertial confinement device for a few thousand dollars, as a number of hobbyists have begun to do.
The fact that the English vernacular now contains the word "biohacker" should tell you something about how much PCR machines and other gene-splicing equipment costs these days.
The fellow who has the (unconfirmed) world record in high-temperature Tc doesn't work at a university or an industrial lab, and appears to be operating out of his basement.
You don't need anything other than a pencil, paper, and a hard surface to bang your head against to get into theoretical physics (though Mathematica does help quite a bit with the unsolvable bits).
The academic space-guard program has started to plug the gaps, but for many years the discovery of new asteroids and comets, especially NEOs, was the parvenu of amateurs.
A dark sky, a decent amateur telescope and a spectrograph will let you make perfectly valid observations of variable stars, and things like astroseismology, etc.
If you're more the Engineering type and want to join the space race, the guys at Copenhagen Suborbitals serve as a good example.
Arguably Paul Stamets is one of the more respected Mycologists out there, and he's just a dirty hippy.
I admit that most of those examples are drawn from physics and astronomy, but that's just because that's where my background lies, but I think I have given enough examples to prove my point, non?
Sure, you're not going to build the LHC in your basement, but that's not the only kind of science left to be done. Just because most work is done by wage slaves in an industrial or academic setting does not mean it has to be.
