The bug in Curve was in vyper's compiler not Curve itself.
I agree with this essay, and your point.
No average user, nor average expert can spot this type of exploit. My own background is in compiler work.
Hence the author, Evangelos Georgiadis, advocates for formal verification methods, but even these methods are not perfect. The most intriguing point for me was the author's concept of conceptual corruption, referencing work by Markus G Kuhn.
It leaves off with an easter egg quote of former compiler expert Donald Knuth.
This 3 pager is quite fitting for current events.
