“Oh yeah, we expected this to happen because we decided to not put in any guardrails.” What was the fucking point then, you galaxy brains?
The “manager” AI went off the rails almost immediately, and their solution was adding a “CEO” AI to supervise it. AI ‘research’ of this type is truly some of the most looney-tunes shit I’ve ever seen society take seriously.
We don’t do serious any more in this day and age. The commander in chief of the largest military machine on earth is literally an insane clown.
LOL at the CEO who agreed to be interviewed and then said he wasn’t surprised that they were able to prompt engineer it in a matter of hours, but we promise it’s gonna be so much better later!
This was a project from Anthropic’s red team…
Worth the read though, very entertaining.
Well assuming this is the same linked article:
deleted by creator
