This is just FUD.
How?
It can reliably hide your location whenever you need it , which is easily verifiable, unlike for any other phone, by looking at the published schematics. If you are serious about tracking, a blind trust in Google’s hardware (a tracking company!) is not reasonable in my opinion, unless you explicitly trust it for some reason. But the latter is far from “most if not all threat models”.
This is not how location tracking works, and the schematics doesn’t disprove it in any capacity.
There are several ways location tracking can happen:
- Via the IMEI/IMSI : You cannot do anything about this except turning on airplane mode.
- Via network location: Android has strict control over this (via the nearby devices permission). So long as you deny the nearby devices permission and use a VPN to hide your actual IP, an app cannot find your coarse location this way. You can turn on Wifi and still do not expose your location.
On the other hand, practically every package app installed directly on PureOS can just check for the access points around you, check their MAC address around you and you are screwed. See how iwlist works.
To make matters worse, Flatpak apps can see your network interfaces and associated IP addresses (including your real IP) if they have the network socket access. See: Install What IP on Linux | Flathub.
On another note regarding Flatpak apps (unrelated to location tracking), there are various sandbox breakouts with Flatpak in non-obvious ways like the dbus access, so good luck spending time figuring out which dbus is dangerous. Oh, and how could I forget, the fingerprinting surface via /sys available to them is massive. Want proof? Open a a shell in any Flatpak app and check /sys and have a field day.
Here is another point as to how “microphone killswitch” can’t provide you with any privacy when you need to talk on a Librem (and the point of having a phone is that you can talk, right? Right?)
- The vast majority of apps on Linux right now uses PulseAudio, including Signal, Matrix, your browser, games, and so on.
- PulseAudio has no concept of permission for audio in and out.
- Flatpak can only control access to the PulseAudio socket, it cannot control audio in and out of that socket.
- You need to grant access to the PulseAudio to apps which you wanna have audio playback.
- Everytime you turn off that killswitch to talk (like having a private conversation in Signal), every app with access to the PulseAudio socket - be it a Flatpak app with
socket=pulseaudio
or an unsandboxed app - can record your conversation.
On the other hand, neither Android nor iOS have this problem, because they actually can control the microphone permission per-app.
How can I verify the Google’s sandboxing? Is there schematics available? Are all drivers and firmware FLOSS? With kill switches these all are not even necessary, and I can be sure it works as intended.
Since you like the fallacy that open source = trustable so much, let me point out the following:
- The app sandbox is done the AOSP source code. Yeah, if you wanna read that, go ahead.
- Are all drivers and firmware FLOSS? No, and neither are the firmware on a Librem.