Comment by denysvitali
1 day ago
They learned nothing from the Grok Code saga.
If anything, that should have been a learning lesson to NOT trust harnesses, especially new ones.
1 day ago
They learned nothing from the Grok Code saga.
If anything, that should have been a learning lesson to NOT trust harnesses, especially new ones.
The basic premise of the LLM companies marketing their "agents" honestly reads to me like: "Yes, we know our commercial, proprietary software has a C2 server[0] embedded. That's by design and it's what lets us deliver all the value, pinky promise. Besides, it's not us remotely operating your computer; it's a massive, more-or-less autonomous computer program that we don't really understand, running on hardware you could barely even dream of."
If people weren't already familiar with the idea of LLMs existing and being able to write usable code and make "tool calls", this would sound completely and utterly batshit insane.
Because it pretty much is.
[0]: https://en.wikipedia.org/wiki/Botnet#Command_and_control
Probably anything concerning that one just register as satirical fictions at this moment to many
Hold on, what happened with grok?
https://www.theregister.com/ai-and-ml/2026/07/14/musk-promis...