A fully rational company should be thinking "what do I have to pay to get the work I need doing done?" If that work can be done remotely then that optimisation works out as:
- Who can I hire in low cost locations and pay less for, without sacrificing quality (which I'm using VERY broadly to represent ANY differences in the work, team interactions, etc.).
- If I can't hire enough of those, fast enough, how much am I willing to pay in high-cost locations.
In the short term, when the company is trying to hire fast, talent is scare globally, and the company isn't connected to all global talent pools (so struggle to hire there) a sensible optimisation is salaries adjusted by location.
In the long term, as the company has time to build its remote brand and (if) its hiring requirements stabilise, a sensible optimisation will be work leaving high-cost centres like SF.
Don't feel too sorry for people in low-cost centres who work gets redistributed to at a lower cost - in the long term work from the high-cost centres is shifting to them, unless there are quality reasons for it to stay in the high-cost areas.