I remember a Linear II test where we had to do Gram-Schmidt on a few large matrices and the prof was a stickler for showing steps. I'm not sure if writing was the limiting factor but it was definitely a major factor. Quantum mechanics is also one of those where there can be a lot of intermediate steps if you don't have things like group theory under your belt (and you usually don't if you're in Griffiths).

I think I'd be careful about generalizing your experience, nor mine. If my time in academia has taught me anything is that there is pretty high variance. Not just between schools, but even in a single department. I'm sure everyone that's gone to uni at one point made a decision between "hard professor that I'll learn a lot from but get a bad grade" vs "easier professor which I'll get a good grade." The unicorn where you get both is just more rare. Let's be honest, most people will choose the latter, since the reality is that your grade probably matters more than the actual knowledge. IMO this is a failure of the system. Clear example of Goodhart's Law. But I also don't have a solution to present as measuring knowledge is simply just a difficult task. I'm sure you've all met people who are very smart and didn't do well in school as well as the inverse. The metric used to be "good enough" for "most people" but things have gotten so competitive that optimizing the metric is all that people can see.