I definitely agree that PoW is very hard to understand because it is not a computer thing, it's about the relationship between large number theory, thermodynamics and statistics/probabilities.

I describe it to myself as a clock, I've written my thoughts down here, you might find it interesting: https://grisha.org/blog/2018/01/23/explaining-proof-of-work/