[worker767424]

> a lot of engineering work of dubious to negative value

Part of this is that FAANG is often willing to spend $20M to try to move the needle 1%, or $100M for the next big thing, even if there's only a small chance it works.

[closeparen]

I don’t mind working on an optimization or a market hypothesis, but there is also a lot of churn for organizational politics and promotion gaming reasons, which sucks to get stuck doing.