The main issue that I see is that interviewers get extra picky with all the minor details and they fail candidates due to minor issues that would get solved in a code review or just with a bit more time
And while fizzbuzz using strings only is cool, it seems it didn't add anything to the predictive power of the interview
You see a "minor issue that would get solved in a code review".
I see a production outage.
You cannot rely on others catching and fixing your bugs. It happens, but it's all about probabilities. You want to reduce the probability that something bad will happen. That means not just relying on more senior people to catch problems in your code, it also means relying on you to be careful.
No, I see people complaining about a test not looking like what they expect, or minor nitpicks like variable names or not using the interviewer's favourite design pattern
Of course the same people who nitpick are the ones writing O(n^2) code and actually causing a major bug