>> In other words, I simply don't believe that interviewers are, as a group, so stupid that they regularly ask obscure Excel questions when they are plainly not relevant.

I was interviewed just a couple of weeks ago. The interview itself was pretty straightforward, but they asked me to do a few test questions, which were very much like those described in the article.

"Here's some really obscure C code with undefined behaviour, what's the problem" "Well, that you wrote it that way, who the hell would write that?" "Sure, sure, but can you see the problem" "Well yeah, you've got someone who delights in writing obfuscated C in weird ways, that's your problem. You could avoid many classes of error here by writing things out in simpler steps."

They were all seriously contrived. (--edit-- and equally as semi-relevant to the role, which is to do with microservices and scalability)