I also suspect that everyone is asking the wrong questions. From the anti algorithm questions for interviews crowd, there is scant acknowledgement that it can't be a blanket policy, as undoubtedly some positions are strongly algorithm-heavy roles. From the pro- crowd, there is scant acknowledgement that by the time you are trying to find new high tech hires through a process that bases an evaluation upon actual interactions that are measured in hours or at best days, you may have already failed. The industry is fitting an interviewing process last majorly overhauled during the industrial age that screens for basic three-R's and attendance skills onto a post-industrial landscape where those skills are barely above cosmic background radiation noise, and demonstrated achievers are as often groomed into hirings through networking over years and decades. Perhaps part of the response to better outcomes to "this high tech interviewing process that takes place in hours or days" isn't "find better questions" or "find a better procedure that still fits in hours or days", but "re-think our premises"? Some kind of PE might fit into that re-think, but the brokenness is so bad that Google has quantified the inability of our conventional processes to yield significantly-better-than-random outcomes, so it might be time to start questioning the entire premise that we can even hire based upon tech interviews in the first place.