Also wanted to add that there's nothing to stop vocational schools to also require their students to be well-rounded, and to expect them to take many courses at University as part of their curriculum. As I said in an earlier post, that would in fact provide a great way to further subsidize university, through the offering of these "service courses" to the vocational schools.

That's not a bad way to look at it either. I'm just afraid it would cost more than my solution. I also wonder if everybody would go straight to a vocational school first and then never bother with a non-professional degree, meaning you'd never see the kind of Renaissance Man model you wished for in the first paragraph (and that I would like as well).

The student debt will just get that much higher. Unless the education bubble eventually bursts, the day people start defaulting on their debt repayments.

How would that be worse than the current situation? Right now those who do non-professional degrees spend all that time AND money on something that won't get them a degree. What I'm proposing is to make the state subsidize those degrees so that at least student's aren't in debt once they've completed a program that doesn't lead to a job. If anything, this should mean that we'll see MORE lower and middle classes in university, aided in that by everybody's taxes.

That's why I wish vocational schools had more prestige. There needs to be clear a distinction made between learning skills and getting an education. Neither is a bad thing in itself. I believe that learning skills, at a School (could be anything, ranging from Engineering to Law, Medicine, Journalism, Design, etc), can be viewed as an investment in the future (in terms of getting a job), and as such it is ok for it to rely on tuition fees. But getting an education, at a true University (with Arts, Math, Physics, History, Social Sciences, etc.), should be something that is fully subsidized. It wouldn't cost as much as you think to fund, since not many people would gravitate toward it in the first place. Once it's made clear that a University won't get you a job, you will only have people who go there who don't quite yet know what to do with their lives (until they figure out that to get a job they should go to a School), or people who have truly scholarly interest in the topic at hand.

There would be bridges between the two, of course. Schools would most likely require some courses to be taken at a University (this way, Schools would also partially subsidize Universities).

I agree. It is great time that University libraries take over as publishers, and spend their money hosting and archiving online journals instead of paying these ridiculous fees. The libraries should also function in a federated manner (P2P!), so that searches for a journal or author can be automatically propagated. Again like in P2P, downloaded articles should be replicated in the local university's point of access. This way most popular articles will be even more protected for the long term.

As for the reputation aspect, I'm pretty sure if Stanford or MIT decided to host their own open-access no fee journals that they would easily attract top researchers for their editorial boards and immediately be flooded with submissions. There are already great examples of reputable online journals, see the Journal of AI Research for example.

There's now open-source software that helps manage the workflow of journal publication. The tools are there, the willingness is there. Let's do it!

Thanks for the link, that was a great lecture!

"Sur l'etoile" was a sci-fi comic book he wrote for Citroen in 1983, but more than just a branding operation and a little gift meant for Citroen employees, it was a very beautiful and poetic piece of work.

Wasn't there also supposed to be a Dune movie with his participation and Jodorowsky's?

He was also famous for Lieutenant Blueberry, his western series he signed under his real name, Jean Giraud.

I've liked whatever I've read from Norman Spinrad. He's more interested in political or societal questions that he can explore using Sci-Fi, than the more typical fantasy aspects.

You're still missing out on the course projects, which are probably the most valuable piece. The Stanford students in the AI course get to work on programming projects, and they get evaluated on those by actual humans. Of course those projects are also available to you if you want to work on them on your own, but there's no actual incentive on finishing them, nor any interaction with other students or a TA.

Also, the machine learning course is a "dumbed down" version. The programming assignments are more of an exercise in manipulating matrices than in understanding the deeper mechanics of machine learning. The same prof teaches a much nicer one in person (material is available online though), but it requires math baggage that most people registered to the online course won't have.

I think the database course is the big winner so far. It is very well taught, and is challenging enough to keep you interested.

In any case, kudos to all the instructors for their time and effort.