> It's becoming a bit like transfer news in all of our favorite sports leagues. Except engineers don't have a union nor contracts like athletes, so: * When la

> It's becoming a bit like transfer news in all of our favorite sports leagues. Except engineers don't have a union nor contracts like athletes, so: * When layoffs / mass firings happen, engineers don't get guaranteed money * The comp of the top 1% doesn't pull up the bottom 99% (not nearly as fast). Much of the bump in SWE salaries in the past 5 years came from the uncovering of the Apple+other co's no-poaches. The $1-5m retainers Google was offering in 2010-2012 to keep G from jumping to FB look now like peanuts. * Engineers at competing companies have no say over each other's comp. It takes actual offers for salaries to rise instead of Eng being able to pool the salary data and stats and determine who's Big Head and who all are the 10,000 underpaid SWEs. Moreover, the glamour of the "transfers" is pretty tightly contained within the author's bubble. A lot of people in AI-adjacent fields still don't even know what Pytorch / Tensorflow are. Important historical context the author leaves out: Hinton accepting $35-40m for his lab to do deep learning at Google is what set most of the initial benchmark. Many stories out there where Hinton broke his NDA, here is one for example: https://www.gq-magazine.co.uk/culture/article/cade-metz-genius-makers-extract https://www.gq-magazine.co.uk/culture/article/cade-metz-geni... It's important context because there's so much arbitrage of private info in AI, you can't take posts like the author's at face value.