3 ms·

> It can’t do that “protecting your privacy” thing Besides 'it's Meta' ; what is it doing with my privacy? I mean actually proven things, not 'probably it is'.

> It can’t do that “protecting your privacy” thing Besides 'it's Meta' ; what is it doing with my privacy? I mean actually proven things, not 'probably it is'. (I am not saying it isn't, but I haven't heard anything in this regard, so it would be interesting to know)

PaulHoule 2y ago
As a dev I see Meta made many decisions to respect privacy that constrain the kind of app I make although I've heard these will be somewhat relaxed. I'd like to place a picture with a QR code in it, have somebody scan the code, then have the option of jumping into a world. Apps can't access the cameras so you can't write a QR scanner. The Quest has a decent web browser but you can't access the cameras and make a web based QR scanner. Without access to the cameras apps cannot at all understand the environment and enable you to interact with it. AR apps now have a special module that identifies a physical volume inside your space on a session by session but that's a pale shadow of the SLAM tracking of the Microsoft Hololens and Apple Vision that let you stick a "hologram" into the corner of your office and have it stay there. Quest 3 devs need more access to make more interesting apps.
- wlesieutre 2y ago
  Vision Pro doesn't let you access the cameras either, being able to stick an app in the corner of your office is handled by the OS. I'd rather not hand out camera access, the problem that needs fixing is the "session by session" part where Meta's OS doesn't maintain permanent app-volumes.
  - godelski 2y ago
    Isn't this more an issue with how the data from the camera is used? Or rather, where it can go? Maybe I'm a bit naive here, but isn't the issue that we don't want camera data going to third parties (or really... any party). So is there not a way that this can be sandboxed? Camera data must stay on device in an enclave. Honestly, this seems what's wanted for a lot of type of data. Is this just not possible to actually implement in the OS? Because I'd imagine not being able to process vision data really limits what you could do. I think of this kinda like the data collection with many phone apps. I understand that it needs access to the networks to make connection to wifi or bluetooth, but why is this bundled together with an app's ability to record and send that data back to the developer? There has to be a better way to handle all this.

PaulHoule 2y ago

As a dev I see Meta made many decisions to respect privacy that constrain the kind of app I make although I've heard these will be somewhat relaxed. I'd like to place a picture with a QR code in it, have somebody scan the code, then have the option of jumping into a world. Apps can't access the cameras so you can't write a QR scanner. The Quest has a decent web browser but you can't access the cameras and make a web based QR scanner. Without access to the cameras apps cannot at all understand the environment and enable you to interact with it. AR apps now have a special module that identifies a physical volume inside your space on a session by session but that's a pale shadow of the SLAM tracking of the Microsoft Hololens and Apple Vision that let you stick a "hologram" into the corner of your office and have it stay there. Quest 3 devs need more access to make more interesting apps.

wlesieutre 2y ago
Vision Pro doesn't let you access the cameras either, being able to stick an app in the corner of your office is handled by the OS. I'd rather not hand out camera access, the problem that needs fixing is the "session by session" part where Meta's OS doesn't maintain permanent app-volumes.
- godelski 2y ago
  Isn't this more an issue with how the data from the camera is used? Or rather, where it can go? Maybe I'm a bit naive here, but isn't the issue that we don't want camera data going to third parties (or really... any party). So is there not a way that this can be sandboxed? Camera data must stay on device in an enclave. Honestly, this seems what's wanted for a lot of type of data. Is this just not possible to actually implement in the OS? Because I'd imagine not being able to process vision data really limits what you could do. I think of this kinda like the data collection with many phone apps. I understand that it needs access to the networks to make connection to wifi or bluetooth, but why is this bundled together with an app's ability to record and send that data back to the developer? There has to be a better way to handle all this.