Yeah, just to elaborate on this (because I think the point of this comment has been missed a bit elsewhere on the sub-thread):
- TripleByte imposes a cost (having to go through their process) in exchange for getting to signal programming ability, a prerequisite to getting in the door (getting an interview)
- People with strong credentials don't need to incur that cost, so they probably won't do so.
This means TripleByte's pool probably doesn't have any Stanford CS grads who are looking to leave their Google job. But these are also the people who command the highest salaries. Let's say it chops off the top x-percentile of the market, for some reasonable definition of x (10? 15?)
Then, TripleByte's screening process probably also chops off the bottom y-percentile, because those are people who can't actually pass the screening.
Once you restrict the range like that, and also make paper credentials less relevant because there's an alternate signal available, of course the rates are going to be compressed.
- TripleByte imposes a cost (having to go through their process) in exchange for getting to signal programming ability, a prerequisite to getting in the door (getting an interview)
- People with strong credentials don't need to incur that cost, so they probably won't do so.
This means TripleByte's pool probably doesn't have any Stanford CS grads who are looking to leave their Google job. But these are also the people who command the highest salaries. Let's say it chops off the top x-percentile of the market, for some reasonable definition of x (10? 15?)
Then, TripleByte's screening process probably also chops off the bottom y-percentile, because those are people who can't actually pass the screening.
Once you restrict the range like that, and also make paper credentials less relevant because there's an alternate signal available, of course the rates are going to be compressed.