[908B64B197]

Maybe DIY degrees using MOOC make sense to replace "amphitheater style" universities but I really don't see how it could be any better than the education at most serious institutions.

Courses in sciences are often split between lectures, that might be "amphitheater style" for introductory courses (think 8.01 or most introduction to programming), lab work and recitations that are typically done in much smaller groups with a T.A to work on problem sets. MOOCs have no obvious alternative to the last two. In my experience it's relatively easy to work through a course by skipping lectures and reading from the textbook than to skip recitations and lab work.

That and group projects, that are often a requirement to graduate, makes MOOC-only a tough sell for me.

[ghaff]

Lectures in big rooms can be trivially replaced. And, arguably, are improved upon by even a relatively modest digital effort. You get time shifting, acoustics and video can be better, you can rewind, and you can get the best lecturer to do it. 8.01 or 18.01 or 6.001 (Intro to physics/calculus/algorithms) doesn't really change YoY.

You could do that on VHS tape if you wanted to.

The hard part is problem sets, recitations, grading, peers, etc. And MOOCs do very little there. Automated grading is better with programming than it is wirth other things. But it's still just looking at the result.

[908B64B197]

There's always MOSS to spot suspicious entries. [0]

[0] https://theory.stanford.edu/~aiken/moss/

[ghaff]

I'm not even really worried about cheating in this context (although that matters when you get to certification--a lot). I'm thinking more of: It gets the right result but it does so in a really crappy way. (And, yes, in some circumstances you can measure CPU time but you still mostly in RIGHT/WRONG grading.)

[908B64B197]

It's an other reason I wouldn't trust a fully DIY degree.