I did read the paper (now). Their question is of “learnability”, which relates to aggregability/compressibility and dimensionality. This stuff is all in the sa

I did read the paper (now). Their question is of “learnability”, which relates to aggregability/compressibility and dimensionality. This stuff is all in the same ballpark. Having read their paper, I do appreciate the formulation and approach. I just don’t find the result “surprising” in the slightest.