A lot of the early FAANG hiring was to prevent staff from going to (or founding) a competitor. So build a pipeline to identify the people who are likely to buil

A lot of the early FAANG hiring was to prevent staff from going to (or founding) a competitor. So build a pipeline to identify the people who are likely to build a competitive threat to e.g. Search and hire them before they get a chance to do that. Pay them enough over the years that they get lazy and start thinking about retirement by 35. It's a bonus if they ship anything useful, and you have in theory increased the odds of that happening by hiring so many of the talented young programmers that graduate every year.