Who are these dedicated schizophrenics who are running long term super smart AIs to kill everyone without anyone noticing? Or are you implying that running LLM chatbots will give them this ability?

We're talking about open-weight models.

It only takes one.