Don’t. Give. Up. Ever.
What they’re testing is your ability—and your willingness—to tackle problems that seem completely baffling. They are looking for you to keep working toward a better and better answer. If you’re ever about to give up remember this: The interview process is designed to challenge you, and even the best candidates feel “stumped.” The question is this: What do you do next?
Just when we thought the fad had finally crested on this patently useless, and more to the point, gratuitously toxic style of interviewing... here's another puff piece to keep the cargo cult going for another 12 months or so. As if the interviewers (themselves most likely just a couple of year out of college) could even hope to simulate the pressures of genuine, real-life, high-pressure technical or operational dilemmas through the device of a made-up puzzle problem or two.
Especially since, about 97 percent of the time, the quality and caliber of grit that brings the most value in business environments isn't of the "solve this made up problem, which I know the answer to (but am hiding from you so I can, you know, watch you think on your feet) -- or hit the road, Jack" variety. But rather of the "no one has any idea where this problem came from yet, let alone whether it's even really solvable" sort. And for which the most effective response, in the vast majority of cases, is decidedly not to go down the next immediately visible rabbit hole and fight the problem to the death... but rather to find a way to avoid going down said rabbit holes in the first place. (Typically by revisiting the stated requirements, and finding what the true customer need is, for example).
And if people think they can simulate these kinds of dilemmas -- let alone "test for" a candidate's ability to tackle them, in an interview context... frankly, they're kidding themselves. Which, being as the true "purpose" of these questions seems to be to buttress the egos of those asking... is apparently perfectly fine, in some quarters.