@sprawlbug yeah, I guess I can take some solace in the fact that apple is at least *pretending* that it provides some kind of security, instead of going full joe rogan "[monster truck voice] we're really sticking it to those EU freedom-hating pencil-necks with this technically compliant but totally useless feature"
Oct 13, 2026, 7:01 PM@aparrish @sprawlbug I've not worked on this stuff at apple but I have worked adjacent to this stuff at a previous job and:
the whole point is to be able to say stuff like "you are only told of the presence of a specific person, but you can not just constantly record the video/audio footage".
I'm happy to talk about what kinds of useful privacy stuff it DOES enable if people are interested!
Oct 13, 2026, 7:32 PM@erikarn @aparrish @sprawlbug I had figured this was just to work around wiretapping laws. I'd be curious to hear more, though!
Oct 13, 2026, 9:15 PM@varx @aparrish @sprawlbug If you want to implement things like "wake word" on microphone, or "is someone in the room" on camera (think smart speaker, video conferencing portal) then in the past you'd have to stream the audio/video into the SoC, run normal software on it, and make a decision.
Later on you could offload that to some always on part of the SoC.
In both cases, going from that to "always recording and uploading into the cloud" is a quick jump.
Oct 13, 2026, 9:17 PM@varx @aparrish @sprawlbug Similarly, if you wanted to do biometric fingerprinting of someone's voice or photo for authentication? Same deal. Having it end up at the SoC for some random bit of software to run over means (a) someone could just write software to always read/record it, and (b) malware could do the same thing.
Oct 13, 2026, 9:18 PM@varx @aparrish @sprawlbug With something like this, you can't - you upload firmware/model into the secure enclave, and the rest of the SoC doesn't /see/ the audio/video. You give it something like a set of vectors representing the faces -> names you want to match, and then not even the OS gets to open the camera/microphone. All it gets is output like "person X is here", "person Y said a word", etc.
It means you can do stuff like absolutely sandbox away camera/mic access unless REQUIRED.
Oct 13, 2026, 9:19 PM@varx @aparrish @sprawlbug Now, this would be better if it were done by someone whos trusted to not abuse it for advertisement/people tracking reasons, I agree.
But it DOES mean you can treat security much more seriously. You can have the SoC completely off and unable to access camera/microphone until you've authenticated yourself, and authenticated you're allowed to access the camera/microphone. Malware can't just hijack the camera/mic drivers and get access to background audio/video.
Oct 13, 2026, 9:22 PM@varx @aparrish @sprawlbug So in short - it changes how you interact with the hardware. Before hand your OS would need access to the camera/mic to implement random features, and malware / bad apps could just open those in the background however which way they could to gain unfettered access to stuff they shouldn't.
Now you can add things like wake word, photo auth, audio auth, etc by pushing biometric info to the secure enclave and getting auth results out, without granting access to hardware.
@erikarn @varx @sprawlbug that does make sense, especially the malware angle, but i still think it's a little disingenuous on apple's part to advertise it specifically as a "privacy" feature—I think when most people are thinking about privacy, they mean "my activity is meaningfully private from advertisers, and even the vendor" not "bad actors can't access raw video/audio" (as a consumer, i would be pleased to learn that a device offers the former, and would just assume it promises the latter)