To the extent that blocks are assembled and validated not by one node (which is all that would be necessary), but by (tens? hundreds? of) thousands, there is massive duplication of the actual work by a factor of 1e4 or 1e5 or more.

But, you are right, even that enormous inefficiency is peanuts compared to the Proof of Waste, which ramps it up to unimaginable proportions. (A single BTC transaction uses up as much electricity as a 2 person household in Germany in a year, and produces 2 iPhones worth of electronic waste.)