Jonathan Blow

@Jonathan_Blow If you got a D in the class, you didn't do homework, and you don't recommend university classes, then how did you learn compiler stuff well enough to end up developing one today that looks like it is going to be better than the existing compilers?

Because most of the stuff taught in those classes is irrelevant to making a good compiler. A large part of it is just general good software engineering, which most modern-day professors don't really know about, because they don't have time to program enough.

To be fair, the graduate-level class that I took was much better than the undergrad class; and, both provided at least some useful background knowledge. But still, I would not recommend people to use such classes as their primary reference in how to make a compiler.

But also, that was a graduate class from the 90s, before things went Full OOP at universities. A graduate class today might well do way more harm than good.

Everything seems to reach a peak of excellence and then decline into mediocrity. It certainly feels like we are past peak software and way past peak academia.

What other examples do you have in mind? Throughout history most people think they have passed peaks.

Physics Ability to Build Things Ability to Coordinate on Large Projects Desire to Do Hard Things

To pick the building example, I'd say there have been multiple highpoints for building throughout history: Roman temples, medieval cathedrales, New York Skyline. I wouldn't say we're at a peak now, but the model "peak followed by monotonic decrease" doesn't apply