Great article, Robb. I disagree though with Nicholas about things like “…teaching yourself mathematics like linear algebra.” As one example, consider UT Austin’s LAFF (Linear Algebra - Foundations to Frontiers) presented via edX for free (at least, it was free when I completed it almost two years ago; there may be a modest fee for it now).
I hold a master of science degree in operations research (strong focus on linear algebra) from the institution that founded the discipline in the US. My linear algebra professor was one of the two best teachers I had ever encountered at that point in my life (the other being my university physics prof). But despite the near-unsurpassable quality of my graduate course in linear algebra, I rate LAFF at the same quality in most ways and even better in some ways, because LAFF featured a strong focus on integrating computer programming (in Python, no less) into the linear algebra problem solving method, whereas my traditional graduate course in linear algebra had a milder focus on computer programming. (In defense of the latter though, I completed it only a short time after the first web server appeared on the Internet.)
Robb, I’m a medical student now, and it seems even more bizarre to me that I’ve reached the same conclusion in that discipline as the one you reached in computer science (“I spent $$ on grad school, only to realize that the free online resources were actually better”), but that really has been my experience thus far. Maybe someday I’ll find the time to write an article about my experiences here, but while I do think it is possible to become a practicing computer scientist this way today, I also think that nobody is going to become a practicing physician through the pursuit of strictly online learning resources (at least, not within the next 20 years or so in my opinion). So for now at least, grad school remains a necessity for those of us with that particular goal in mind. But it is rather striking that the quality of the conceptual teaching itself could be (even today) better through free resources than through some traditional, expensive ones like accredited medical schools.
Because of services like Coursera, edX, and so many others like them, I think education is currently undergoing a dramatic revolution like nothing that’s happened since the founding of the universities of Al Quaraouiyine (in 00859) and Bologna (in 01088) or even Plato’s Academy (c. 387 BCE) and Aristotle’s Lyceum (c. 600 BCE). Because the MOOC and similar resources are so new, I think it’s much too early for anyone to be able to understand their long-term impact on civilization, but for me and Robb at least, it’s pretty clear that they are already big-time game-changers.