> I suspect this is because a large block of programmers were asked it and continued to ask it, unaware of how deeply tied to the historical context the question actually was.

Bad hires are expensive and false negatives are lost opportunity. The industry is constantly looking for ways to improve its hiring. No old practice is off limits.

Google no longer asks brainteasers [1]. Rumor has it Facebook discourages dynamic programming now [2].

Yet somehow the author assumes that nobody ever thought to question the Linked List because tee-hee unlike him we're all naive lemmings.

No, we do iterate on all our questions and remove the ones that don't correlate well with results. We just continue to find linked lists to be a decent predictor for now, at least no worse than other DS.

I do agree that all-or-nothing questions like cycle detection are poor, but those kinds can be found in all data structures.

[1] https://qz.com/96206/google-admits-those-infamous-braintease...

[2] https://www.teamblind.com/article/DP-questions-at-Facebook-i...