I mean, it's about your threat model isn't it? The way I think of it, there's basically three ways for someone to break into my computer: remote code execution

I mean, it's about your threat model isn't it? The way I think of it, there's basically three ways for someone to break into my computer: remote code execution with no external preparation (as might be caused by ssh exposed to the internet with a common username and password), installing malicious programs, or opening malicious files that exploit buffer overflows or whatever. The first way might be protected by a sandbox or might not: if it is ssh that's pwned, presumably in a sandboxed environment you'd have given ssh all the permissions anyway, right? The second way is far less common a vulnerability on a desktop Linux system than on other OSes, including Android, because 99% of packages people are going to want to use on their desktop are going through the distro maintainers, which isn't a perfect process but seems to be pretty good. Contrast that to Android where malware gets through the Play Store all the time. iPhone seems to be about as good as Linux in this regard though. I'm not totally sure that sandboxing would help in this regard either. If someone exploited my PDF reader it'd probably have access to all my other PDFs wouldn't they? However if someone exploited my video player they wouldn't, and in either case they probably wouldn't have access to my browser, so yeah sandboxing would help there. But also included in the threat model is what you want to protect. Obviously bank passwords, tax information, etc, but I also want to protect my privacy in general, and right now there's no OS that's as privacy preserving as Linux. Furthermore, I don't trust sandboxing implementations in proprietary OSes from protecting my privacy either: they might say they protect my data against malicious apps trying to steal my stuff, but then if the OS itself tracks my location, search history, files, etc, sandboxing doesn't do much good, does it?