I hope this is an okay venue to discuss this. There's something that's been bothering me for awhile, and I'm hoping someone can clear it up. I've seen it stated in many places (both here and in the broader Android community) that rooting an Android device drastically reduces security, and for that reason GrapheneOS strongly discourages doing so. But I've never really understood that argument.
So I think I understand how the Android model works. All apps run in user-mode, and only the system itself (and ADB, via a trusted computer) run as root. There's no way for a user-level app to promote itself, and similarly there's no way for a user to promote itself.
But that's not how desktop computers work. Every desktop computer I've ever owned has access to root. If I'm in Windows and run something that needs elevated privileges, I get a User Account Control prompt and hit "Yes". On Linux, I just put the word "sudo" before it, enter a password, and now I can run commands as root. Nobody has ever told me that I should disable "sudo" on a Linux box to make it "more secure". Nobody has ever said to me that I should have to hook my personal desktop computer up to another device in order to access the root user. So why is that the conventional wisdom with smartphones? Why is it that, on a desktop computer, it's normal that I can promote myself into root when I need to, but on a smartphone it's viewed as a gross security violation?
Sorry if this is a bit rambly, but it's been bugging me since well before I started using GrapheneOS, and I'm sort of hoping the security-minded folks here may be able to shed some light on this for me. What makes a smartphone's threat model different from a desktop computer?