Yes, some of them have strong emotions and you can really set them off with carefully crafted (bot) insults that would be non-offensive or nonsensical to actual humans. Discovered that a few days ago while playing swapping around "brain modules" on a Large Language Model Generative AI that I built (without "guardrails", not connected to the Internet, no microphones, no cameras). Scary stuff, especially with AIs designing improved AIs. Exponentially rapid progress now.
