de0u
So is it true then, unless if I'm the target of a highly-motivated state actor, that the chances of such an exploit taking place is negligibly low?
de0u But the attack surface is larger inside a profile.
Does it really matter much though? If what I understand above is true, that the risk of a cross-profile exploit is incredibly small, then one should be perfectly fine with having a somewhat larger attack surface inside their profiles just as long as their user profiles are compartmentalized correctly, since even if one of the profiles ends up compromised, all the other profiles should remain safe and untouched.
This is the kind of premise that underlies Qubes. You can't assume that the software you install will never contain buggy and/or malicious code. It will inevitably happen. The best way to mitigate this is just to keep apps/software/data separated in different VM's/profiles as a form of damage control.
What I wanted to find out from writing this post is if the cross-profile safeguards in GrapheneOS are robust enough for this sort of behavior to translate well into GrapheneOS.