this post was submitted on 24 Nov 2024
6 points (100.0% liked)

Hacker News

309 readers
840 users here now

RSS Feed of HackerNews

founded 3 months ago
MODERATORS
[email protected]
[email protected]
6
Robot Jailbreak: Researchers Trick Bots into Dangerous Tasks (spectrum.ieee.org)
submitted 1 month ago by [email protected] to c/[email protected]
1 comments fedilink hide all child comments
 

Comments

you are viewing a single comment's thread
view the rest of the comments
[–] SpaceNoodle 4 points 1 month ago

And this is why you don't just plug your LLM directly into motor controls. You need an executive unit that acts in a well-defined manner based on sensor inputs, with failsafes.