In comments, JMG3Y asks, "Where should a smart science-oriented high school student with a breadth of interests go to college?" This deserves a much more careful answer, but time is precious, so consider this a rough draft of an answer, which people are welcome to amplify in the comments. (Past installments here and here. At some future date there will be an installment on "How to be a good graduate student.") In reality, colleges and universities are very different from each other, and each should be considered separately. Also in reality, any such institution is huge and multifaceted, and two people can have wildly divergent experiences at the same place. Furthermore, sticking again to reality, this is a question that depends mostly on the individual student, and for which there is no right answer. Being all that as it may, for purposes of exposition let's lump the possibilities into four categories:
Liberal-Arts College (LAC), such as Swarthmore or Amherst.
Specialized Technical School (STS), such as MIT or Caltech.
Elite Private University (EPU), such as Harvard or Stanford.
Large State School (LSS), such as UCLA or Michigan.
These are fuzzy and incomplete categories, of course, but hopefully the ideas will come across clearly enough. At an LAC or STS, you will be forced to learn a lot, like it or not. I'm a big fan of LAC's; the professors are typically talented and dedicated to teaching, and students get invaluable up-close-and-personal time with the faculty. But for people who want to go to grad school, they face something of a disadvantage because the these schools typically won't have graduate programs. That means (1) you can't take any grad classes, and (2) you can't buttonhole grad students about advice for the next step. I went to one, and received a great education, but keenly felt those disadvantages. The STS's are also great (I work at one now). Your fellow students will be interested in similar things, and the coursework will challenge you. There will be plenty of opportunities for research experience, rubbing elbows with grad students and postdocs doing work at the forefront of science. Both MIT and Caltech have a feeling at being at the center of the scientific universe. Of course, they generally won't give you a broader academic experience, if that's what you're after. For me personally, one of the best parts of being an undergraduate was being exposed to ideas in the arts and humanities (and people, both faculty and students, in those areas) that I never would have experienced otherwise. At an EPU or LSS, it's generally much easier to slide by without stretching yourself, if that's your thing; on the other hand, the resources are tremendous, and if you have the initiative to take advantage of them, you can have a great experience. The best thing about an EPU is the other students. So much so, that at a place like Harvard it's generally acknowledged that a large fraction of your education comes from extracurricular activities. You'll meet people, in your field and out, who will be running the world a few years down the line. The professors will be great researchers who may or may not be interested in teaching; there will likely be some opportunities for research and individual contact, but not all that much. An LSS will also have great resources, in terms of faculty and research opportunities. There might be more close contact with professors than at an EPU, but that's quite a generalization. Your fellow students will be more of a mixed bag; some will be geniuses and future world-changers, while many will be there to tread water for four years to get a degree. Of all the choices, the education you get at a large state school will depend the most on your own initiative; the school will almost certainly have more to offer than you possibly have time to take advantage of, but nobody will force you to do any of it. For the particular goal of advancing to grad school, there are certain specialized factors to keep in mind. Having grad students around to ask questions to is certainly helpful. The choice of undergrad advisor is also important, I suppose, but depends much more on individuals than on schools, so I don't know what to say there. It's important to get some research experience, but this can often be done off-campus at other places during the summer (see the NSF Research Experience for Undergraduates and similar programs). Getting good letters of recommendation is certainly helpful -- for that, it's less important where you are, and more important that people there know you well enough to write sensible letters. When it comes to actually applying to grad schools and making choices, it's nice to get advice from people who know what they're talking about; don't be afraid to ask around. Perhaps my own perspective on this kind of question is coming through clearly enough: wherever you go, your educational experience can vary wildly depending on how much you put into it. If you stick to what's required, slide through with just enough work to get whatever GPA you're aiming for, and spend the rest of your time playing video games, you'll manage not to get much out of it no matter where you are. If you seek out new and challenging courses and activities, spend your summers doing research or interesting off-campus activities, and make an effort to talk individually to your best professors and hang out with other students who enjoy ideas, it will be an invaluable experience. If you ask most 40-something professors what they would think of going back to school for four years, to do nothing but take interesting courses and discuss deep ideas with their friends, their eyes would light up with unvarnished pleasure at the prospect. Whatever you're studying, college is a unique opportunity to stretch your mind; make the most of it.