This general idea shows up all over the place though. If you do 3D scans on thousands of mammal skulls, you'll find that a few PCs account for the vast majority of the variance. If you do frequency domain analysis of various physiological signals...same thing. Ditto for many, many other natural phenomena in the world. Interesting (maybe not surprising?) to see it in artificial phenomena as well

It's almost an artifact of PCA. You'll find "important" principal components everywhere you look. It takes real effort to construct a dataset where you don't. That doesn't mean though, for instance, that throwing away the less important principal components of an image is the best way to compress an image.