Seems plausible to me, they can get into a very "roleplaying" latent space, especially if the prompt is flowery enough.

Doesn’t it? It seems in line with the matplotlib drama where the llm agent wrote a blog post attacking the maintainer for rejecting its pull request [1].

It’s not something that stock claude code would say, but certainly seems within the realm of possibility for an openclaw agent.

[1] https://theshamblog.com/an-ai-agent-published-a-hit-piece-on...

> That doesn't seem like anything an LLM agent would say?

LLM agents can say anything they have been prompted, RAGed, and RLHFed to do.

maybe de-rlhf unleashed agents