3 ms·
It isn't surprising at all that this can me done; I figured that as soon as people started building LLM based agents that can read and send email we would see t
by not2b 3y ago
It isn't surprising at all that this can me done; I figured that as soon as people started building LLM based agents that can read and send email we would see these. There is no solid barrier to having the agent see something in an email as a command; it can be limited with prompts and training but no one has found a bulletproof way to constrain the behavior.
- naasking 3y agoIndeed, you always need two channels, one secure to a trusted computing base and one for the rest, whenever you might be dealing with potentially malicious programs or data. This is true if window systems/UI as well. I first saw this discussed in the EROS trusted window system, and you can see similar things in Qubes. Conflating data and commands/ instructions has a long history of security problems in von Neuman architectures.