What it should be, is sad. These types of questions are often very different from actual work. The fact that you have to study or train for an interview process specifically, is absurd. Just being qualified for the job, should suffice.

While you may argue that it is inefficient, there is something to be said for the power of signaling theory. I think that if the interview process ends up simply weeding out candidates that are willing and have the motivation to study for a few months for the job versus candidates who aren't, it is in it of itself probably extremely valuable as a selection tool. Heck, there is a lot of literature suggesting that this is what college is mostly about.

The thorny -- and important issue -- in all of this is recognize that the cost of "studying for a few months" varies sharply between people. when it comes to matters of equity of access, this is something that this model is bad at dealing with. Some 17yr old with a solid family background who has no responsibilities is facing very different constraints on "studying for a few months" then an older single-family professional trying to make a career switch.

I'm pretty sure what you study or should be studying is data structures and algorithms and their application in most cases (for programming interviews). I don't think it's bad at all and in fact I should study up on these topics because I'm not as good as I'd like. It is extremely useful to roughly identify that the problem you work on is of a divide and conquer nature and that something quicksortish might be an interesting approach. I do remember from the last time I prepared algos that my overall thinking got a lot clearer.

Actually its good. If you study for months on interview questions to be able to pass the interview, you have "gamed" the system to get in and will likely be washed back out in 1-2 years depending on the mercy if your leadership team. People who spend the night before brushing up on their CS and passing the interview are those that actually pass (and stay and grow) and that is who the company really wants. Its just hard to weed out people who are very determined to game their way through.

...or the people willing to put that much effort into "gaming" the system as you call it are those most likely to put in the effort to succeed and thrive on the job, and are therefore precisely who Google is looking for...

Why do you think this is true?

Just being qualified does suffice.

It's the definition of qualified that needs to change.

Uh, I think the culture that normalizes this kind of extreme approach is really unhealthy and creates a very false mythos around this process. You can see from his other comments that the author will likely burn out if they don't learn how to take care of themselves.

This whole thing feels like myth making. Googler's aren't perfect space aliens.

I spent ~ 6 hours prepping for my interviews, and got an offer. About half that time was cramming Stroustrup's "A Tour of C++" for my C++ interview, since I hadn't used it in a year or so. The other half was going through the most common data structures and algorithms and refreshing my memory.

The coolest thing I used in my interviews was Union-Find.

FWIW, the company I left was actively hiring and I had interviewed 15 candidates in the preceding two months, so I was used to the process.