The real innovators are going to be the people who come to the system from the outside and find a new way of teaching. And it will be much easier for those innovators to gain people's attention if the existing system isn't muddying the water with weak online offerings.
Take law, for example. In CA, NY, and a few other states, would-be lawyers can study via apprenticeship, distance learning, or some combination of the two; although this can be a career handicap compared to attending a decent brick-and-mortar school, it doesn't impede one's ability to sit the bar exam and receive a law license. But in Texas (a large and growing legal market), it does - you can be admitted as an attorney to some other state bar and to federal bar, but if you got your law degree by correspondence then you won't be given a TX law license. Of course this won't last; eventually some attorney with a non-traditional legal education will win some landmark case in a federal appeals court, and barriers like this will be scrapped because nothing succeeds like success, and no state bar wants prevent a big client from hiring the star lawyer of the moment due to some ancient admittance rule. But right now there's much more supply than demand, so employers have little or no incentive to take a risk.
Of course, many jobs do not require same sort of rigorous standards that exist in the professions; you could make an economic argument that all credentialism is a kind of rent-seeking. But it's not going away until there's some alternative measure of skill that adequately predicts performance. Again, it's a bit easier in tech; you can sample the results of someone's work on their web page, and often it either works or it doesn't. It's less obvious how to do that in other fields.
Consider all of the people who go to community college and transfer in to the prestigious school for the last two years, and get the same degree. You don't need to get a degree from Khan academy, you just need to get the knowledge. Then you can transfer to a prestigious school and get the degree that gets your resume noticed.
A nice podcast on this: http://www.econtalk.org/archives/2011/09/winston_on_lawy.htm...
And in turn, a lot of the first-year legal theory is only superficially about the rules, and more about exploring the many different modes of analysis that can be used to interpret those rules. A lawyer friend advised me 18 months ago that the rules are fairly easy, and that it's not unusual for a good paralegal or legal secretary to know the rules better than a lawyer does, but that the skill of an attorney is more about being able to explore the many possible interpretations from a given set of rules, whether to craft a persuasive argument in a dispute or to avert future disputes by anticipating where misunderstandings might arise and pre-empting them by clarifying ambiguities. So the more law I learn, the less I understand, if you see what I mean. At some point this trend will reverse itself...I hope :-)
So I can sort of understand the Texas approach here. Yes, it does involve a certain amount of nest-feathering by lawyers who wish to keep their services positioned at a high price point and use the ABA standards as an exclusionary filter. But on the other hand, you really don't want a lawyer whose only claim on a license is the ability to remember a large number of rules and apply them to straightforward cases, because that sort of lawyer will treat an unfamiliar problem like a legal dartboard and just try to hit as many rule issues as possible in the hope that one of them will score big. That sort of kitchen-sink strategy is good for intimidation or burying someone in paperwork, but it will fail badly against someone with a really strong analytical capability who will be able to weigh the significance of all the different rules in context and guide a court towards a wise decision rather one that is merely superficially correct. The lawyer with an excessively mechanistic, rules-based approach won't do a really good job for the client , and that undermines the standing of the whole profession. It's like the differences between first aid training, the greater skill of a paramedic or nurse, the deeper knowledge of a doctor, and the really expert knowledge of a surgeon or specialist researcher. All kinds of medical expertise have value in different contexts; you don't want to pay a brain surgeon to treat a minor cut, but nor you do want the camp counselor to operate on your subdural hematoma.
You can treat academia like any other industry and you need someone who knows the the system very well to innovate a better means. Plus, mass online education will still have to conform to some form of the highly bureaucratic standards that give real weight to the degree.