I mean, it already kind of does (on iPhones at least, don't know about iPads) — ever notice that the screen will automatically come on when you pick up the device and tilt it to a certain angle, but only if you're looking at it?

Mind you, that's a trigger from sleep. But you could just as well have one that's a trigger "from stillness" — i.e. when the device hasn't moved in a while, and then it does.

I believe a few months ago HN comments were riffing on a hypothetical "persistent background authentication" system Apple could be developing that would justify their disinterest in bringing back TouchID for iPhones. It would basically work by using every sensor available to try to constantly determine whether the same person that previously unlocked the device, was in continuous physical possession of the device. As long as that was true, the device would remain unlocked (rather than there being any kind of lock-during-sleep timer.) But as soon as the device was passed to someone else, or taken out and left on a table, all authentication would be discarded. (It wouldn't necessarily "lock" in the sense of taking you to the lock screen as soon as you place your phone on a table — you might want to show someone a cat photo, after all — but it'd at least temporarily assume a low-integrity kiosk mode for that app, becoming "locked underneath", in the same way that the photos app viewed from the camera app accessed from the lock screen is low-integrity + "locked underneath.") This would mean that any authentication that is required could take a lot longer / be a lot more thorough — because the most common kind of auth, the "incremental re-authentication" when you take your phone out of your pocket for a second — would no longer exist.

I feel like this kind of thing is totally possible — even plausible/practical — given Apple's fondness for developing low-power sensor-tracking ASICs for the Apple Watch et al.

> If I were apple and I cared about this, I'd release a simpler version first and wait for feedback.

Apple doesn't really do this; they seem to try to "get UX right the first time", in the sense that they'll never really make a UX iteratively better, only ever completely throw a UX away and then create an entirely new UX that's a-bit-more-than-iteratively better, with an entirely different name and branding (E.g. Exposé → Mission Control.) — presumably so that users don't think it's the old thing, try to use it like the old thing, and fail.

But, mind you — the Apple TV's tvOS already has "user profiles" and "user switching" in exactly the "V1" way you're describing! (https://support.apple.com/en-ca/guide/tv/atvb59ec8e2e/tvos) tvOS doesn't do any kind of automatic switching; it only allows for explicit switching.

So, conveniently, the code for most of this logic is already there in the iOS codebase. It'd just need to be adapted to the iPad's UX; and some annotations added to apps to mark them as "allowed in low-integrity mode" or not, where apps that aren't "allowed in low-integrity mode" apps — i.e. "requires high-integrity" apps — should never be allowed to be run/installed on an iPad set up for sharing (just like "requires high-integrity" apps aren't allowed to be published at all for tvOS to begin with.)