One thing I like about this is making the actual difficulty deltas between colleges more felt/legible/concrete (by anyone who takes the exams). What I might do in your system at my IQ level (which is pretty high outside of EA but pretty mediocre inside EA) is knock out a degree at an easy university to get warmed up then study for years for a degree at a hard school^[1].

In real life, I can download or audit courses from whatever university I want, but I don't know what the grading curve is, so when 5/6 exercises are too hard I don't know if that's because I'm dumb or if 1/6 is B+ level performance. This is a way that the current system underserves a credential-indifferent autodidact. It's really hard to know how difficult a course is supposed to be when you're isolated from the local conditions that make up the grading curve!

Another thing I like about your system is tutoring markets separated from assessment companies. Why is it that we bundle gatekeeping/assessment with preparation? Unbundling might help maintain objective standards, get rid of problems that look like "the professor feels too much affection for the student to fail them".

This is all completely separate for why your proposal is a hard social problem / a complete nonstarter, which is that I don't think the system is "broken" right now. There's an idea you might pick up if you read the smarter leftists, which is that credentialism especially at elite levels preserves privilege and status as a first class use case. This is not completely false today, not least because the further you go back in time in western universities the truer it is.

my prior, 15 years ago, looked like "stanford has a boating scholarship, so obviously selectivity is a wealth/status thing and not reflective of scholarship or rigor", so the fact that I now believe "more selective colleges have harder coursework" means I've seen a lot of evidence. It pains me, believe me, but reality doesn't care :) ↩︎