According to the article, "most people think they are heretical, but are actually mainstream" -- that includes the interviewer!
Any candidate smart enough to have heretical beliefs is probably also smart enough to know that the beliefs are heretical because they are unpopular and voicing those opinions to someone who, by the odds, most likely isn't heretical, but thinks they are, can only be a negative for your hiring prospects.
It's just another case of "do you interview well". An interviewer should try to control for nervousness, unless they need someone truly outgoing/confident. And asking someone to reveal something very personal is the opposite of that.
An interview should measure competence for the job first, and additional traits second. I'd much rather have a technical question, followed by a discussion on alternative solutions and why they might not be a good fit. I want to see how they reason about a problem, and I feel like that might give me insight into what Peter Thiel is after.
Your argument is noy exactly true. If the value of finding a similarly open-minded individual outweighs the low probability* of finding it, then it makes sense.
Put another way if given a utility function with terms (utility of open minded individual) (probability of OMI) > 1 then it is rational.
All that being the case, there's still a certain irony to this chapter in Thiel's book.
Thiel thinks that thinking different on god or the existence of god is a terrible answer, even though it often means someone has a heretical belief vs. perhaps their parents and potentially the bulk of society.
His justifications for why this is a bad answer are flimsy, and seemingly the only reason he thinks this is because he agrees with mainstream thinking.
This proves the danger of answering the question--the spirit it is received, even in book form from the very person advocating it, is judgment of the answer and an underlying preference for mainstream thinking.