alt Hacker News > just giving the LLM the ssh keys and having at it. It’s pretty good.
Whereas, it you let the AI go free on a server, you mostly have no idea what it did and where, so it's quite hard to audit and even harder to try and fix things.
So, unless relying on a declarative paradigm like NixOS does, I'm not ready to give AI the keys to my castles
Replies
yoz-y • today at 7:57 AM
Admittedly this was not my castle. More of a $1.5/mo shed that I got just to try this.
➕ show 1 reply
While admittedly this can be a bit risky, there's a lot less possible blast radius or thing that can go catastrophically wrong if you give the LLM the ssh keys and sudo access on a KVM VM (or equivalent) that's been created as a blank slate exclusively as its playground. I have a template debian 13 KVM VM that can have its ssh host keys and hostname wiped, other unique per-node things, and duplicated with a basic bash script in a couple of minutes to deploy for home lab testing.
I'm fine with giving a "smart" self hosted LLM like qwen 3.8-flash-next Q8 access to it to do things in opencode.
Don't give it the keys to any castle but give it access to one specific containerized thing...