IMO, one principle of our institutions of higher learning seems to be the selection of students with certain 'meta'-learning skills. I'm mostly thinking about the ability to effectively (and promptly) expand some very compressed piece of information (perhaps a sparsely stated / open-ended problem, some obscure jargon or technical acrobatics, or anything that requires a great deal of 'unpacking'). The extent to which so much value is placed on the exercise of competent thinking and not on the thoughts themselves seems to follow from the scarcity of expert attention at the research level. In other words, it's more important that you
could understand
god-knows-what should
Expert A start spewing irreplicable insight at you. So it's not so important you're doing anything of real value 100% of the time. In principle, you can do that on your own time (nevermind that competing in classes doesn't leave all that much time to do so, however).
I'm not really sure what use this style of learning is for people who don't plan to go into academic research, and having such an aptitude for on-the-fly learning is beginning to make less and less sense as education (and research) becomes much less of an oral tradition.
Industry does benefit from the university system as a (very inefficient) way of filtering out the people who, for one reason or another, might screw up at work. (They're also getting the taxpayer to shoulder some of the burden of some of the training that will ultimately earn revenue for the company.) If the students don't plan to go into some obscure area of academic research, I don't see the harm in knocking some common sense into the curriculum and allowing the students to learn some much more practical things that they might actually ever use. Probably a good place to start is to stop
training students in lieu of providing a chance for them to educate themselves.