Also, I think the outcry would be larger if they did it from scratch compared to if they did it as an extension to some existing, known capability. If that's the case, they'd have less to lose in doing such a thing if the base system is already in place.
No, the GP was correct. As soon as any closed-source software implements automatic software updates, you've always one malicious update away from the system betraying you. Having "a system in place" for doing potentially evil things is unnecessary. Interim steps of any kind are unnecessary.
What Apple has done this week doesn't bring the iPhone closer or further to your hypothetical dystopia than it already was. Or Chrome, or Windows, or Android, etc. They update themselves. Every update your devices have done in the past decade could have betrayed you.
Anything that automatically updates is always one step away.
You're talking about "installing" a change, and talking about more about their capability to change what happens with your data.
I'm talking about 1) the effort required to _write_ the change and -- more importantly -- 2) the potential backlash being different as to whether it's a modification of an existing functionality vs an entirely new type of functionality. This second point is a major one, because it would be seen as much worse if it looks to the public like they've gone out of their way to do something wrong, and would be much more damaging to their reputation. IMO anyway.
https://pingthread.com/thread/1424873629003702273
> For the conspiracy to work, it'd need Apple, NCMEC and DOJ working together to pull it off voluntarily and it to never leak. If that's your threat model, OK, but that's a huge conspiracy with enormous risk to all participants
Apple could intercept hashes that are sent and compare to their own database.
Someone in the NCMEC could add non CSAM hashes to their database.