One of my favorite things about how companies interview is how they assume their own competence in actually assessing people at a level that is woefully out of

One of my favorite things about how companies interview is how they assume their own competence in actually assessing people at a level that is woefully out of line with reality. For example. "We ask puzzle questions because it helps us figure out how a candidate problem solves." No it doesn't, and not because puzzles are a bad way to assess problem solving, but because you as the interviewer have basically no idea what you're doing when performing the assessment. Typically the interviewer has no clear idea what profile of person they actually need to hire for a given type of job function. Even in the rare case they do have that necessary understanding, they almost certainly have zero training in how to create or identify puzzles that actually align with the kinds of problem solving that match the profile they're looking for. The interviewer will have zero training on what to observe about how the person solves problems and what the signals actually mean. They'll instead almost certainly pick up all the wrong signals and miss all the important ones and interpret the ones they picked up in the wrong way. Usually people tell me they ask certain kinds of puzzle or programming questions because they want to see how the person solves problems, collaborates, communicates, etc. As a quick litmus test I always respond, "When you ask these puzzle questions to your candidates do you already know the answers?" 100% of the time, "Yes, of course." To which I say, "If you already know the answer, you're unable to remove your bias from the way of coming to the solution, and you're learning absolutely nothing about how they communicate or collaborate. Next time ask them a puzzle you don't know the answer to and work on the solution together. Then you'll have half a chance at learning something valuable about the candidate."