I would push back on a couple of your points
@Dr. Barber.
I wouldn't say that the exam got
easier with the switch from long-form "show your work" questions to multiple-choice scantron, but the focus definitely changed a lot.
With the old long form question format it was about thoroughness, good assumptions, and good design decisions.
With the scantron format it's really about speed and time management.
I only took the scantron format PE exam, but having worked some of the old long-form type problems, I personally found them to be much easier. Maybe because that's what I do everyday at work, but I think that's true of many engineers that work in design and analysis. Obviously each person is different, but making good design assumptions, and taking my time to solve complex a problem is much easier for me than cranking out A LOT of "simpler" problems super fast.
Also you mentioned that the problems are no more complex than undergrad work. That has always been the case of the PE exam and that's generally true in industry (unless you have specialized in something very complex). Almost all engineering design is algrebra-based, maybe a little bit of "calculus 1" or the simplest of differential equations. But definitely nothing more complex than that. I think that my current work is at the more complex end of engineering analysis in industry (structural analysis of visco-elastic materials), and still I've never had to use the derivations and advanced math that I learned in grad school. Same was true when I was in aerospace and in ship-building. The research and theoretical derivations that we learn in grad school just isn't how day to day engineering is done. Now I'm not saying that it's not valuable. Getting my MS in MechEng has definitely made me a better engineer. But that type of work has never been the focus of the PE exam.