Abstract concepts in IT is not enough to understand real life problems (or the domain of the problem). By making those connections one is, on average, better prepared to solve those problems. In other words, one may know how to write a program but one might not seem to find a problem to solve it. Giving people these brain teasers helps assess _somewhat_ how well equipped one is to deal with certain problems (and how fast they are to solve them). And yes even completely or seemingly bogus ones.

Sure, it might seem random and bogus to ask a Google employee to count the number of balls in an airplane, but suppose Google is developing a self-flying airplane and they need to account for the transport of balls within it for some reason... maybe it's not so bogus anymore. Interestingly, something similar was attempted by Waymo (or rather its precursor) to deal with. How would a car deal with a certain completely random and absurd situation? They have to account the possibility or risk damaging their endeavour.

We never know when those crazy outliers may happen but when they do we'd rather be prepared than being caught with our pants on our hands.

> Giving people these brain teasers helps assess _somewhat_ how well equipped one is to deal with certain problems

No, the whole point of this post's source is that this is not the case, with even Google's former VP of People Operations calling it a "complete waste of time".

> but suppose Google is developing a self-flying airplane and they need to account for the transport of balls within it for some reason

If you had to bet you'll never, ever put money on Google building airplanes to transport balls because we both know this is a completely ridiculous and nigh impossible scenario. You might as well ask candidates Santa Claus trivia. And even if you were right this question would still be useless because nobody in their right mind would let software engineers design airplanes.

I remain unconvinced. I still think brain teasers help understand one's decisions. A tangent to this subject is Google's famous post-mortems. An incident X occurred - what did the team do? What processes were in place to minimize risk, etc? Was there anything that the team didn't think about? (Unfortunately most of the time these things are taught by experience and in production.) An so on. It might seem bogus but sometimes it's not. Reliability in certain domains is very important because sometimes human lives are at risk.

  nobody in their right mind would let software engineers design airplanes.

I never said anything of the sort. Software engineers do have to know certain aspects of its domain problem otherwise the software controlling the hardware will be fundamentally broken.

When I was kid I thought maths were pointless - because I couldn't see how that would be applied to real tangible things - until it suddenly clicked how it all worked (this happened because I started reading more about it and actually finding interest in it). Brain teasers help one think even if the scenario at first seems completely random and bogus, and yes even "trick" questions are valid.

This new standard of thinking, about these problems as a waste of time, is probably related to something else; probably politics and the feeling that those said brain teasers are considered by some as "elitist" (or white as they call it now). If that is the case then how about explaining things to people in other ways? I'm a big proponent of educating the masses. (Something like what the Khan Academy or 3Blue1Brown are doing.)

Another reason might be because people feel anxious about job interviews already and having stranger shove in a brain teaser makes the whole experience even more stressful. If that's the case then maybe this person is in the wrong field or applying to the wrong company/job. Or maybe that person needs to solve other life's problems before applying to any job that's stressful.