I'll argue that the most important part of grad school is/was the people. And that is the element YC nailed for me.

I also had to teach myself a bunch of technical stuff to get xobni off the ground. Unlike grad courses this stuff was practical by definition.

For me YC and YC alums more closely fit the profile of an engineering grad program than a MBA program.

Are we talking about masters or PhD version of grad school? I could buy it for a masters-replacement. But an engineering PhD is supposed to be producing new basic technology, not learning existing technology, and ideally producing new technology that requires multi-year research to produce.

That's also possible to do in a more innovation/research-focused startup (biotech does this all the time), but it's not particularly common in tech startups afaik, mostly because "multi-year research 'til product" is not what investors want to hear.

Granted, the results are not always amazing in grad-school either, but quite a bit of basic research in PLs, systems, algorithms, etc. does come out of PhD theses. Could you really do something like Okasaki's "Purely Functional Data Structures" within the YC program?

Graham himself is actually doing an outside-of-academia version of PLs research with Arc. But most of the YC companies don't seem to be embarking on Arc-type projects.

I think equally as important (but not more important) as the people you're studying with, are the people who are advising you. After all, I was always told to pick a grad school by the quality of the advisors you'd get to work with, not by the school's overall reputation. A really good grad school advisor teaches you not just the theory, but the "practical side" of theory - introducing you to their peers, explaining how grants get written and funded, walking you through paper writing and otherwise showing you how the sausage of research is actually made.

Of course, YC has plenty to offer in the advising department - that said, I think the distinction is worth noting. Why? Because I think there will be more quality advisors shaken out of the woodwork by YC-like programs - particularly in different cities where perfectly talented mentors have built their fortunes and families and have no desire to move from. Just like not everyone has to go to Harvard to become a great grad student, people won't have to go to YC to have great start-up mentoring. And that's a good thing; our economy needs more than one "startup grad school". I'm excited not just about YC, but about the trend YC has kicked off - the number of great people attracted to other YC-like programs across the country is going to skyrocket, sooner rather than later. This will help the economy grow overall - and help shield the tech industry when the bubble that is "expensive academic credentialing as job qualification" pops.