The early Web radically democratized culture, but now in the media and elsewhere you're seeing a flight to quality. The best American colleges should be able to establish a magnetic authoritative presence online.
1. Do not equate "quality" with "incumbents."
2. The MIT's and Stanford's of the world have some advantages. Deep pockets. Great business connections. Smart professors and leading-edge technologists.
3. They also have some disadvantages. I could argue that they do not really know what teaching is. They know how to guide students about what they should learn on their own. That the students then learn this stuff tells you a lot about the selection process of the admissions department, maybe not so much about the teaching process of the faculty. They also will face organizational conflict, so that while part of the institution may want to go full speed ahead, the rest of the institution will be dragging its heels. Why do we think that Thrun left Stanford? And by the way, I signed up for coursera to taste a sample and...nothingburger.
4. I am not afraid of Silicon Valley VC's, either. They have one or two buzzwords (e.g., "gamification") and their typical herd mentality. I think VC money helps when the solution is obvious and you need to be the first company to get to the finish line. I am not convinced that having a VC breathing down your neck is a real help when you need to putter around in your metaphorical garage.
5. "How can we translate college into the online world?" is the wrong question. Education is not going to experience one gigantic conversion from analog to digital. Instead, legacy institutions are going to be pecked to death. One company will tackle a little piece of the problem here, while another company will tackle a little piece of the problem there. I think where we will end up is with education that is disaggregated, rather than replacing the aggregations that we now know as schools with equivalent online aggregations.
6. I am thinking of getting into this education start-up game. Not with A Means A, but with something else entirely. At this point, everybody except my wife is telling me not to bother trying. But that was pretty much true with my last start-up, and it turned out she had better insight.
[update: Joshua Gans got what I think are the key take-aways from efforts to just throw existing courses on line. One is that putting a lecture on line is no great feat, because lectures are not such great teaching tools to start with (See also David Friedman's comment on this post). Another is that student homework serves as a teaching tool and as an assessment tool, but often the latter use gets in the way of the former. My start-up idea is based on these insights.
Instead of starting by asking, "how could I lecture to and asssess 100,000 students?", I start by asking, "How would I teach if I had just 1 student?" If you had one student, you would never give a long lecture, or even talk for 5 minutes without interruption. You would iterate--give explanations, give exercises, give hints, give more exercises, and when you know the student is ready, move on to the next topic. I want to think in education in those terms.]