[veets]

I always find this argument of "Google gotcha" questions weak. I agree 100% that tricky questions around dynamic programming, linked lists, etc. are a terrible choice. However, I think a lot of these questions do a good job of measuring how well the candidate understands basic data structures. It amazes me how many "seasoned engineers" do not understand when to use things like hash sets/maps (just look at the monthly blog posts about how you shouldn't use `in` in Python on lists because it is O(n)!! It blows my mind this is blog-post worthy).

Also, I would far rather have a company not hire me because I failed a programming assessment than have them hire me and then fire me a week later because they were finally able to assess my skills.

[jen20]

The issue fundamentally is that most people do not use this kind of knowledge on a daily basis - even those of us doing low level work, and when it is needed, it is a mere web search away - or if necessary I can reach for Knuth.

I think you misunderstand my notion of a "google gotcha" - I'm not proposing that interviews do not into rigorous technical detail. I'm proposing that interviews that test knowledge of easily-looked-up trivia rather than an understanding of engineering pragmatism and trade-offs are fundamentally flawed and yield markedly worse results than those that do not _regardless_ of company.