Hacker News Reader

Comment by j4hdufd8 | Hacker News Reader

> On Linux never required

Oh okay! Would you mind running a really cool program I have here for you?

If you stick to your distribution repositories, flathub/snapstore and never run `curl | bash` from untrusted sources, you would be fine.

And if I have to run your bespoke program, I will use a sandbox or VM whether an antivirus is running or not.

And if you do your AV program wouldn’t catch it anyway.

I rember many years ago trying to run netcat [1] on Windows for some tests. AV: Blocked as "hacking tool".

I asked a friend who happened to have Visual Studio installed to build it for me from source. AV: No problemo!

[1] https://nmap.org/ncat/

Did you try renaming it to “notnetcathonest.exe”?

Perhaps a better question to ask: why, in 2024, is it still possible that running a random app infects an operating system?

Don't we have cpu's with a whole set of hardware features to limit what <insert executable here> can or can't do? Don't we have OSes with fine grained permissions, VM's, capabilities, etc, etc? Weren't these things figured out like, in the late 70's?

> Don't we have OSes with fine grained permissions

No, we really don’t. We have OS’s with the tools to give us fine grained permissions, but at the end of the day on a desktop OS, any program you run can do any damn thing to any other damn thing owned by the same user. Look to phones for security boundaries that make sense in a post-1970s world, but not any shipping desktop OS.

Doesn't Qubes OS solve this problem? It is linux only though.

Average users tend not to like operating inside heavily sandboxed environments that make it difficult to move data between apps. See: iPhone.

Because security isn't perfect and attackers need only one mistake. There were some early computers built with security in mind, yet their cost and performance were impractical for most uses.

There is also the problem of everyone having different needs, so permission profiles must vary widely. OS builders cannot assume anything about the limits of their use.

Finally users can be the weak link. They may not understand the risks of what the screen is asking them when prompted.

Reply on news.ycombinator.com