I think the main reason for Wirth's law stinging as much as it does, isn't so much because of ignorance of algorithmic complexity. Rather, it's because most programmers (and all the languages, libraries, frameworks and various assorted tools and materials that they use) still operate on a mental model of how computer architecture works that is literally decades out of date.

Mainstream infrastructure hasn't done a stellar job in seamlessly abstracting new hardware features (though compiler construction is still decent there in many regards). The one that has, isn't usually mainstream.