Barring an AGI that can take care of knowledge work, companies will continue to pay a premium to developers because they are often times the core value creators of the business. Even in traditional areas like finance, quants with CS PhDs are displacing Harvard MBA’s trading on fundamentals b/c their returns blow the latter out the water.
Both are paid a premium, but FAANG is paid a massive premium, and there isn't much precedent for collecting that massive premium continuously for 30 years.
EDIT: Actually in that spreadsheet I would argue that the $250K compensation is too low, the 10% return on investment is too high, and the 30 year timeframe is not sustainable.
Also I kinda question the assertion that these companies being active and important in thirty years would suggest something untoward. In most other industries, the "Blue Chip" companies are pretty durable. JPM's lineage goes back to 1871, and that fact does not prevent its current employees from being well-compensated.
Lol, I heard the same arguments in the 1990s about webmasters, which at that time were also commanding large premiums over the market median. I also remember when any engineer who touched a linux kernel could make 3x "normal developer" wages. Most FAANG engineers aren't working on anything too special; the biggest competition will be off the shelf frameworks/libraries/application which can do what previously required custom work.
We are also in a period of easy investment money - the biggest threat to FAANG companies is the market demanding a return on their investment - P/E ratios are at historically unsustainable levels. Either "this time is different", or this will all end very badly for a lot of people, just like the first dot-com boom.
There are good years, and then there are bad years...