* mentors
* hands on experience and dives into theory/standard learning
* live "on campus"
I was lucky to have all of these at a standard liberal arts college, albeit not in the exact doses that Khan is talking about. It made my education great.
One of the great successes of the university I attended were several weekly hour-long two-on-one sessions (two students, one supervisor) with either a professor or a grad student, talking about that week's lectures.
The regularity and length of them allowed a relationship to be built up over the course of the semester, and you could be damn sure you'd learn a lot more from that dialogue than you ever would from sitting bleary-eyed in the lecture hall at 9am.
I've heard that's what you get at Oxford. Seems great but as a mere mortal who attended a slightly less prestigious institution, I had to make do with tutorial groups where 18-20 people would cram into a room designed for 12.
You get it at a few places. It has its downsides though: expensive for the university, and a big time suck for the supervisors, but honestly it's a really successful system. I'd go as far to say that if there's any one thing an institution can do to improve the abilities of their students, it would be to adopt a system like that.
I'm not sure how students can be expected to learn from a one-sided teaching system. We're not photocopiers. There's so much value from a meaningful, extended dialogue between a student and a "mentor." Unfortunately, a lot of people have to wait until postgraduate programs to get that, and even then may not.
A lot of US institutions have office hours which could offer much of the same if students knew to take advantage of them in the right way.
Liberal arts education seems to get a bad rap around here (I did take less higher math and science than would have been required at a state university), but the average class size for my college was twenty. Occasionally more, but there's also the counter case, where I had an English class with two other students. One had a voracious appetite for skipping, and the other was consistently absent for mock trial.
I found it substantially raises my perceived importance of finishing the required reading when I think there's a 50/50 chance of being the only one there to be questioned about it. :) I had a huge motivation problem end of high school/early college, and I think if I hadn't had the individual attention of a liberal arts college, I surely would have dropped out.
I've followed a similar background becoming an Engineer. I did the co-operative education while an undergrad and got valuable mentoring and working experience in the marketplace. Now I'm an organizer for a local R users group which meets at a university. We do a lot of hands on and practical theory which has benefited me and our data science community.
