In the corporate IT model, the business defines requirements and then seeks implementation resources. The goal is on-time delivery and passable correctness at minimum cost. That might be fulfilled through in-house IT employees, a COTS product, an onshore or offshore consultancy, or some hybrid. In some cases there is no substitute for expensive in-house onshore talent, but the business is always incentivized to seek one.

FAANG engineering shops are not like this at all. Tech teams on their own initiative or with very general charters conceive, create, and release products. If product “requirements” are even written down, there’s a good chance the lead engineers on the project are coauthors. Engineers beyond the entry level are expected to be curious, creative, self-directed… basically bounce around a general problem space creating value wherever and however they can, as long as stakeholders sign off prior to any disruptive changes and there’s a measurable impact to cite for year-end perf.

The value created can be mind-boggling - and so can the compensation packages, at no real loss to management. Any team’s backlog of opportunities is deep, and all of them are much more valuable than any amount of wages that could conceivably be spent chasing them. You don’t just want the job done to spec at minimum cost - you want the most capable and innovative people you can get to be at the whiteboard thinking of jobs it would be interesting to do, because the carrying cost if they’re just okay is insignificant, and the upside if they pull something really brilliant is unlimited.