> if the software is served over the Web or to any platform with automatic updates, E2EE is fundimentallu broken; and, first-party end-to-end encryption is snake oil.
This statement, and this way or extremist thinking is the classic no true soctsman fallacy. Taken to the extreme it goes back to trusting trust, and not everyone is a cryptographer who wants to verify the encryption is strong and implimented correctly.
Privacy is a sliding scale and people have different threat models.
The statement is also so broad that applying it as a specific issue with E2EE seems foolish - after all the same could be said for any website. You're trusting that Amazon doesn't maliciously update their website to start sending passwords or credit card numbers in plaintext. Hopefully this example demonstrates why encryption in transit is useful.
Just as encryption in transit has tradeoffs (mainly caching), so does E2EE. The benifit of E2EE is data at rest - the server owner cannot suddenly decide to decrypt your data for their means. As you've outlined, for strong privacy this requires tradeoffs (local calculation, more complex systems). As you've also outlined E2EE is also not a magic bullet (trust the client), but that doesn't suddenly discount all of the benifits.