I think one thing the author fails to realize is how much worse it is to hire a terrible engineer than not hire a good engineer. I've seen a terrible hire cause more than $1M of damage in the year it took for him to get fired.

While I think that these sorts of people are rather rare within the population of engineers at large, they appear much more often in the interview process as they are more likely to need a job, and probably have to interview many more times before they find their way into a company.

That said, I do agree with the author that extremely deep technical interviews are also unreasonable. Variants of FizzBuzz are generally enough to weed out the terrible engineers and it's pretty hard to tell the good from the great until you've actually worked together.