With programming and electronics, self-study offers the opportunity to make things, both novel and derivative, and immediately (or near enough) see whether or not the function. Similar to setting up an experiment in a physics lab, but again, who has a home physics lab? Programming can be done on a cheap (< $200) laptop. A large portion of electronics can be accessible with a similar investment in (mostly reusable) parts.
For electronic components, "easy to learn" meant things like designing op amps and logic chips that behaved more like their textbook symbolic representations without weird tricks and pitfalls.
In contrast, you can't change electromagnetics just because you want something that's easier to learn. Nature won't approve. Progress has been made towards explaining it better, for instance, today's notation is simpler than what Maxwell had to grapple with. But such progress comes slowly.
In hindsight, I have this conflict of views sort of. If automation goes the ways in which some fear, then physical institutions will provide a place for people to go instead of rioting out in the streets. To build on Salman Khan's thoughts about online education becoming the Aristotle/Plato model while the physical area acts as a place to build projects and so forth.. this could be one way to go [0]. At the same time, I've read a number of books about the atmosphere of campuses pre-1960s. Personally I feel we protest too much these days. So perhaps a focus on online campuses while limiting fewer people to the physical campuses can allow for the protests to die down and perhaps a return to the pre-1960s era. That said, I have my doubt that people are just going to be content sitting at home in mass learning.. so perhaps Salman Khan's ideal will be what emerges. Then again, it would also be ideal that we don't throw out the old way completely. It would be ideal with some smaller liberal arts colleges could be dead set in their ways in just teaching old liberal arts. Ideally without the computer so to encourage actually reading texts in whole and producing critical thinking. This would be the elite probably due to the cost of that education.
My bias yields to the hope that the majority of all jobs are destroyed by automation and that people having nothing else to do but to get phd's in the subject that they are interested in. Perhaps though the access of education will depend on the level of talent and knowledge; for the masses a khan/coursera style approach, for the mid-level an approach that is a hybrid of khan as Plato model and physical interaction for project development. The upper end, the elite would be in person primarily, small classes in cohesive campuses with lack of protests that disturb the open learning environment; of course moocs would be available to these students but the primary purpose would be more old form to develop better critical thinking skills. The moocs, even if PhD programs are developed will suffer the hive mind problems as one can see in say many social media platforms (reddit, facebook, twitter etc).
[0] A Vanity Fair panel discussion with Salman Khan and Reed Hastings: https://www.youtube.com/watch?v=cnqaenik93A
edit: added youtube link for reference.
I needed the structure and social interaction of the college environment. But oddly enough, it depended on what discipline I was studying. I would not have been able to learn math and physics on my own, and those were my majors. On the other hand, I easily taught myself programming and electronics.
A friend whose daughter got an online degree (while at home with a baby, living in a small rural town with no other options) said that the hardest thing about studying online was the sheer boredom of sitting at a computer all day with no one to interact with.
A particular advantage of going to college for me, was that I learned how to teach by observing my professors in a variety of settings, and imitating their methods. Today, a lot of my work is surprisingly similar to teaching: Presenting ideas to people who are not experts, mentoring younger colleagues, etc.