Graduate student here. Would love to talk about this stuff.
>There's loads of kids going to graduate school to become servants for mediocre, yet politically-connected scientists.
The first part is true, but the mediocre scientists aren't necessarily politically-connected. More often, they're just busy, so consistently busy doing their own research and supervising their own students that one day some new and improved methodology, instrument, or statistical tool leaves them in the dust. This happens because, frankly, nobody pays them to keep their methodology, instruments, and statistical tools up to date, and it's not like anyone will refuse to publish their papers if they don't keep up-to-date. They get grants because they publish papers, and they publish papers because they keep doing experiments and/or proving theorems -- the fact that they could be learning more and quicker by working some other way isn't built into any incentive structure except their personal hopes and dreams (which are the first thing any young scientist learns to repress).
There is especially no shortage of scientists in the sense that, simply put, there are not enough permanent jobs for the PhDs and post-docs we already train. http://www.phdcomics.com/comics/archive/phd030909s.gif
>Then, once they have spent the X years and $Y in opportunity costs to get a PhD, they have to figure out a career for themselves. This is a bad, not a good, thing. It reduces early-stage career experimentation and encourages low-quality science.
This is quite true. I find the advising structure in graduate school strange with respect to its stated purposes. The definitional requirement of a PhD is that it be a large piece of original research (stereotypically, about three papers' worth) that charts the course for the young scientist's early career. The best PhD dissertations open up whole new subfields, and those are the young scientists who get the good jobs. But how do we allocate labor to produce these "brilliant new ideas"? Largely by assigning the graduate student to work on their adviser's existing research program, which they can slowly transform into something at least partially their own depending on their level of supervision.
(The worst PhD dissertations are basically just longer MSc dissertations: fully supervised research work done in total peonage with little to no original input by the apprentice graduating to become a real scientist.)
>There are loads of theories and ideas in any given field about the way that things may be. What you need, as a minimum, is the ability to stick the course to figure out a way of testing those theories and ideas and do the (extraordinarily boring) legwork to get it done. I know a lot of scientists (physicists and mathematicians) and all of them are stubborn, but not particularly brilliant. 99% of physics and mathematics are made up of people who will do drudge work for years with very little recognition or reward. Very, very occasionally, one of them hits the jackpot. It's not brilliance that defines a professional scientist; it's a willingness to be bored and work weekends for very little compensation.
clap clap clap http://media.tumblr.com/tumblr_m6odkq8jNo1qbolbn.gif
Absolutely, completely, entirely, 100% true. The vast majority of professional science is legwork. Theoreticians can spend plenty of time constructing models or proving theorems to feel brilliant, but ultimately, they mostly just do legwork as well. Experimentalists and engineering researchers have to actually do experiments and actually produce artifacts. Mind, many of us find even the drudgery and legwork plenty fascinating to set-up and do, but almost never so much that we'd object to your taking it away somehow.
Also mind: some scientists are brilliant. They're just not the mode, or even always the shining stars of the field, since professional scientists are mostly selected (and therefore optimized) for tenacity (being bored and working weekends for very little compensation) and consistency (producing 110% the desired results, from elementary school through grant applications).
I would add only one quality: a professional scientist is defined not only by religious devotion to doing loads and loads of legwork for little compensation, but also by an emotional ability to withstand being constantly graded, measured, and stack-ranked against everyone else in your field, in time periods ranging from each semester to every few years. From the beginning of undergrad through tenure and up until you move from doing research to mostly teaching or administrating, you basically never stop having to pass yet another set of exams (each one calibrated to fail some percentage at the bottom, forcing them to find other careers) to continue doing your existing job or to receive a promotion that other skilled professionals would usually get by virtue of experience alone.
While I've heard that many decades ago, funding and positions were plentiful, and so the system was not nearly so burdensome as today, now I am always a little amazed that scientific advancement manages to happen at all, since almost everything a scientist does professionally is connected to keeping his job instead of actually getting and sharing research results.