Quote:
Originally Posted by Victor
well, I am probably in that magical zone that he describes where the interviewer will allow for direction and molding so I would happily accept it.
If you don't have a strong opinion on what interviewers should do, not sure what it is that you want people to have to learn. There's no objective body of knowledge on which questions are appropriate for what type of jobs - while we know which types questions are illegal and which types of questions are inappropriate, we don't know if networking questions are better than OS questions or dynamic programming questions are better graph theory questions.
Quote:
and maybe the egotistical guy that you are describing is not the best person to conduct the interviews or to have such power over the hiring decision.
But who decides who's the egotistical one? Who polices the police? Either you empower individual interviewers or have an even more egotistical person who thinks everyone should ask X type of questions enforce a policy made around his own biases. The point of a decentralized process and multiple interviewers is that an average of lots of people's biases is better than one person's bias.
Quote:
these guys are looking for every advantage possible. they are heavily researching and testing new ideas for small edges. it seems completely crazy that they will allow for things like the "anti-loop" to occur.
This is a problem for job candidates - not a problem for Google. Google isn't trying to hire everyone who is qualified. Interview anti-loop is also a theoretical concept that merely has a loose existence proof based on the level of selectivity required, not something empirically shown to be relevant. It's also a rhetorical device for Googlers to make their friends who were rejected by Google feel better.
Quote:
would it really be that hard to structure the interview cycle so that candidates are not asked such disparate questions that they are highly unlikely to know both?
You want to ask disparate questions such that most candidates are unlikely to know both - that's how you select the best ones. The problem isn't that you don't want to ask disparate questions but that for any given candidate, it's possible to find questions that this person will fail on. It's a statistical artifact - you can't know everything or be good at everything, so it's possible you can end up with a bunch of interviewers who happen to hit your weak points. Another thing is that Yegge is overstating this to flatter his audience that surely includes lots of people who are rejected from Google.