Aspects of this idea will probably not work out, most especially the math, from the first few answers. The baby that HN has in the bathwater is the infrastructure aspect. The arguments for physically centralized educational institutions as a necessity rather than a luxury good are very weak outside of programs with practical physical components, which I would imagine to be a big chunk of STEM and the fine arts. That sort of physical infrastructure — labs, studios, theaters — is itself much more distributable and interchangeable at the undergraduate level than in the current system. The technology to hold lectures and seminars remotely is fully mature, and even if physical needs remain, those needs can be met by local infrastructure. Pedagogically, development of wider course standards and more sophisticated modes of course credit assignment are both feasible information-architecture level projects, and all these things taken together could transform education from a credential gateway implicitly knitting social networks designed to retain power to a ubiquitous more cloud-like environment focused purely on the learning, scholarship, and research that is needed to actually keep the wheels from falling off society.
A lot of people are less welcoming of OP's style now that the ranks of the very online are less exclusively populated with socially primitive young men jockeying with each other for imagined status, but aside from the ludicrous belief that he might have solved major problems by taking a half-assed crack at them, he's not wrong to be going big on this. Of course everything we come up with is going to be impractical bullshit. At best some posters with real and interesting knowledge will be rustled out of the bushes. But people should be thinking about the gap between what we actually need, what we actually have, and what could be done with the means at our disposal.