4 ms·
The "slightly paranoid and cynical" hypothesis I have is that the reason the whole set of docs looked like something pushed out at 2AM, and that the "Clarificat
by Syonyk 5y ago
The "slightly paranoid and cynical" hypothesis I have is that the reason the whole set of docs looked like something pushed out at 2AM, and that the "Clarification" documents looked like they were written by people operating on a week of no sleep, is because they were. Apple wasn't planning to really bother documenting the "feature," and someone found something sufficiently objectionable to leak it. Maybe someone wanted to turn it on early or something, or the hash database had some "non-CP data" in it, or... I don't know, and probably will never know.
But they decided that releasing what they had was better than letting the leaks run wild - just, it turns out, the actual system was really that bad. The "clarifications" were a wild run of "Well, OK, how can we answer this objection? And that one? Oh, crap, this one too!" - which is about how the later documents read. "We just came up with this idea that might address your objection!"
I'd love to hear the inside story of it at some point, but accept I likely won't.
However, it's been enough to drive me off Apple products. I can only hope enough other people have done so as well to hurt their profits, though it's doubtful. :(
- jodrellblank 5y agoBut why? The way Microsoft, Google, Facebook do it is run the CSAM scans on their server side on everything you upload. Apple could have done that. They could have done it without announcing anything. They could also have put it on-device as a background service which scanned every photo you took or saved, or every website you visited. Instead they put a non-trivial amount of effort into designing a much more complex system that doesn't require them to be able to see everything you upload, one that has input from multiple countries to guard against a single organization adding meme images to it, one that has an adjustable threshhold for some consideration against false positives, one that doesn't scan everything on your device even though it could do, etc. The actual system was not "really that bad" as you say, considering what it was trying to do, it was a reasonably good attempt to do that. (Not flawless). Apple already send notification of every program you run on macOS back to Apple HQ and sync your browser bookmarks and passwords through them, Google know everything you search for and which results you click and can track you through sites using Google Ads, Microsoft try to get you to send all your browsing traffic to Microsoft and report on executables picked up by Windows Defender. Compared to all these things, Apple's scanner design was tackling a more difficult topic in a more privacy protecting way. I wouldn't be much surprised if Apple push their scanner to macOS as well as iOS, still only for images uploaded to iCloud and not more than that. I would be quite surprised if Microsoft pushed a similar thing Windows 11 and only restricted it to scanning OneDrive uploads; I doubt that the money saved in running the scans on-devices instead of in-datacenter would alone be worth the engineering effort of moving such a system from datacenter to device as Microsoft would have to do, compared to Apple apparently starting from scratch. That is, if it appeared in Windows 11 I'd expect the reason to be being a whole device scanner. I wouldn't be dreadfully surprised if such a system appeared in the Chrome source code and thence to Chromebooks, Chromium and Edge. The next 10 years is probably going to include various pushes from governments and companies into more invasive tech, getting in with a design that is somewhat privacy protecting and setting that as an expectation is not a terrible idea - especially if you assume they don't want to do nothing, or are not allowed to continue doing nothing.