[brandonb]

I used to be on the hiring committee at Google, and sometimes we'd ask for extra interviews if the person was very strong, but feedback was mixed in a particular area. For example, if a candidate bombed one algorithms question but performed strongly on an algorithms question from a different interviewer, then we might want a third data point.

That happened for <10% of people -- it was really the exception case.

So I think b) and c) could be that process.

a) sounds like they were just disorganized -- which happens a lot with early stage startups. Takes founders time to learn how to hire well and I don't think any accelerators or VCs are providing interview training to their startups.

[lingua_franca]

I find the interview process at Google/Facebook extremely disturbing for senior devs like author of Homebrew. Nobody can remember how to manipulate binary trees/linked lists/(un)directed graphs after working on real world projects for several years. Why can't you just interview like the rest of the industry? I understand some deadwood might get in this way; just simply fire'em after several months on the job. Netflix is an excellent example in this regard.