I can't tell if this is real or a parody of the SV ed-tech solution-ism of the last decade

Yeah, what could be so hard about buying a 50 billion dollar "big computer" and hiring 10,000 PhDs for an online platform that doesn't exist yet? Ha!

It's real for sure though, this is HN after all. The general idea of free public universities based on an open curriculum is good, but whatever the OP is talking about is not going to get us there.

Either way, the upvotes were unironic.

From sources like this [1], it looks like the cohort of high school graduates is usually between 3 and 4 million. So that means the total college-eligible cohort is 12 - 16 million.

Let's say 25% (a fourth) of them end up attending this magical national public college. That's a 3000000 / 10000 = 300:1 ratio of students to teachers. That's a lot of students per PhD to handle, for a $100k salary with greatly reduced career prospects in academia, since they are not doing any research.

By comparison, the student/faculty ratio at a typical "good" college is about 5:1. Even knocking down the projected enrollment one order of magnitude (so 1/40th of the college-age cohort) gets you a 30:1 ratio which is still 6 times worse.

Feel free to point out a flaw here, but I don't think this plan will work. Basically, you'll end up doing a MOOC site, but those...already exist?

----------------------------------------

[1] https://www.insidehighered.com/news/2020/12/15/what-do-new-p...

> By comparison, the student/faculty ratio at a typical "good" college is about 5:1. Even knocking down the projected enrollment one order of magnitude (so 1/40th of the college-age cohort) gets you a 30:1 ratio which is still 6 times worse.

What qualifies as a good college, for this description? Typical state school, e.g. Oregon State University? What counts in the 5:1 ratio? Because it sure isn't just teachers, that's probably more like 50:1.

Indeed, 300:1 is pretty common in large lecture courses, even ones at relatively high ranked schools in fairly important topics (algorithms, for instance). And that's with the teacher having to give a different lecture to every single section every year, instead of just recording one lecture for everyone for the next several years.

Even more time is saved if we unbundled university education in general, so that the various paths and credentials are their own thing. IE, Intro to Poetry 101 isn't necessary for someone who's pursuing finance credentials.

> Because it sure isn't just teachers, that's probably more like 50:1.

How do you figure that?

Mean class size at a university is probably in the neighborhood of 30 students. Yes, there are some classes that have monster, 200+ class sizes, but this is also balanced out by senior- or grad-level classes that have a dozen students in them.

Students also take several classes, and on average, each instructor only teaches ~1 class a semester (especially consider research professors who may not teach a course every semester). So if students are taking ~5 courses a semester on average, I'd expect to see a ratio somewhere around 6:1, nothing even close to 50:1.

I’ve never had a single class with less than 30 students except for ones where the students dropped out mid-quarter. (Ironically one of my smallest classes was one that was for 65+ students but the instructor was so bad that we had less than 20 by end of quarter)

I went to University of Washington - if it matters.

For the record, OSU lists their student:faculty ratio at 18:1. First year gen ed courses are huge but most upper level classes are a much more manageable size.

https://oregonstate.edu/about

It's been a while, but I don't remember any of my upper division CS classes at OSU being smaller than ~30 students. That's why I wondered if perhaps admins were counted somehow, or maybe grad students, to get to a lower ratio.

To be fair, replacing an expensive diploma mill with a free/public diploma mill isn't that bad an idea. Probably just too "socalist" for North America. Imagine if everyone could have access to higher education without going into massive debt!

Uninformed tech people are better than a lot of experts in many fields. Wouldn't be the first time. Bezos responded to COVID faster than the NIAID.

I agree and have had a similar idea for awhile. Have the best lecturers for given subjects create outstanding virtual courses, develop sets of curricula off of those courses, and have TAs (PhDs in your model) provide local instruction and tutoring to augment the standard coursework.

It seems crazy to me that we have thousands of new faculty members and PhDs begrudgingly spending time developing their own way of teaching material (instead of working on research which is what they actually want to do and are good at) when there are people who are clearly the best in the business at teaching their subject matter whose work we should be using instead.

Based on my experience, we'd get better real world results if we abandoned most universities and promoted community colleges to full undergrad level. My best teachers by far were at a community college. Probably because that's all they did. No begrudgingly 'teaching' students while really wanting to do research.

Not that there isn't a place for research teaching, and students along that path, but like 99% of everyone else would get better results from a different academic culture.

Since this is HN I just want to point out that $50 Billion will buy you 50 new exaflop systems [0]. That’s a lot of awfully big computers to use as web servers.

You could probably deploy most of the courseware using Netlify for a few grand per year. Then you could put that 50 billion in an endowment. I don’t really know how those work, but I’m guessing with a 5% return you’re making 2.5B/year giving you an extra 25k teachers at $100K each. ¯\_(ツ)_/¯

https://en.wikipedia.org/wiki/Fugaku_(supercomputer)

Update: frontier [1] is faster and cheaper but my point stands.

[1] https://en.wikipedia.org/wiki/Frontier_(supercomputer)

> you have created a national public college for pennies.

You've slightly modernized correspondence school.

Just to start with what's missing, the vast majority of students need some degree of individualized attention to learn effectively.

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.

This solution is missing the chance at connections made by attending a school teeming with the useless scions of the rich, who will be handed plum jobs upon graduations. You're not going to have a chance to hook up with something like Skull & Bones sitting there doing remote learning the way you will by coughing up the fees to attend one of the Ivy League schools.

I think this is overrated. I went to a rinky dink community college. It was a very urban school (meaning, no grassy hillside campus, we were in the middle of NYC and students had jobs and families to go) and did not have a campus life, so I didn't make many connections with my other students. I still ended up with a great degree in CS and a great job. For sure if I went to a campus I'd have more opportunities to network and maybe be further in my career, but I achieved success without that.

I guess you can use the same idea to create a national car company, a national cloud service company, a national food company, a national computer store, a national hospital, a national everything company. What a great idea!

an interesting idea…