Interesting points you bring up and Google effectively came to the same conclusion when they were smaller (low hundreds instead of tens of thousands of employee

Interesting points you bring up and Google effectively came to the same conclusion when they were smaller (low hundreds instead of tens of thousands of employees). Everything at Google is data driven, including the hiring, but things have changed drastically. Back in the day, Larry and Sergey had a forumla that used the SATs and your graduating GPA. The theory was that your SAT was a measure of your potential, and your GPA was the ability to deliver on that potential. As you can imagine that ultimately ended in failure, but kind of an interesting idea none-the-less. Since then they've wound up taking a much more Googly approach to hiring, by collecting tons of data from their employees. They effectively gave everyone a survey that attempted to figure out what defined the Google culture and success there-in. By hiring people consistent with this data, the theory is that they can better keep the culture and continue to hire successful engineers. Lots of interesting tidbits came from the study, but the most interesting was that a masters degree will (statistically) always outperform a b.s, but a phd is closer to a 50/50 for outperforming a masters (they attribute this to being too specialized. I forget the phd-to-b.s. probability). All this is from memory and is a couple of years old, so take it as you may.