Normally, I don't like nice and tidy coding quizzes. But I'd actually kind of enjoy finding the first 10 digit prime in Euler's constant. Of course it isn't a

Normally, I don't like nice and tidy coding quizzes. But I'd actually kind of enjoy finding the first 10 digit prime in Euler's constant. Of course it isn't an actual interview question, it's just a question you need to answer to get to the page where you apply for a job. One thing I'm curious about… suppose someone does an a-ok job with the euler's constant question. How predictive is that of success on the job? How much more predictive would it be to ask that person a lot of data structures stuff and so forth? I guess what I'm getting at is, how useful are the intense, deep technical grilling we often go through for interviews. let's say you don't like them, but take a more moderate position that there will need to be some coding, and it can't be too trivial and easy (like fizzbuzz). So you ask a question that's a little tougher, and perhaps not anticipated, like this euler's constant problem. If someone gets a "passing" grade on this (ie., there could be better efficiencies, but all in all, they present an approach that works), how much more predictability do you get out of spending an additional 8 hours on sql outer joins, binary trees, cycles in linked lists, and so forth?