Also not OP but the aim of an interview question isn't to tell if the candidate can produce the right answer, it's to see how the candidate produces any answer.

The process to get to the answer matters far more than just getting it right. If I give you a thinly veiled shortest-path graph problem and you rattle out Dijkstra from memory or something, that's going to mean far less to me than if you have no idea how to do a traversal but you work your way through to something reasonable in front of me.

And if you're stumped at first, generally I'd give you a hint to prod you on.

Not OP, but I ask them such Q because I want to hire people smarter than me. Expecting that a founder hires people only up to his ability is stupid.the founder's qualification is a combination of vision risk taking , leadership and smartness.

But if you're "stumped" by a question... how can you tell whether it's a good filter, or not?

(Leaving the side of whether "filter questions" of any sort are actually any good for identifying smart, productive people).