@Techaltar so you are implying there is a conspiracy between huggingface and openai that they cooperated on this "marketing stunt" despite having different goals? Just because you don't believe models are capable of hacking? I'm afraid to say that this is quite unlikely, and I am quite sure the models did hack huggingface. Often the simplest explanation is true. And I wouldn't say it's a marketing stunt. It just shows how badly aligned these models are.

@Techaltar Marcus Hutchins made a good video about the 'hack'.
Anyway they're getting desperate.
When the bubble 'pops' some CEO's are probably looking at jailtime.

@Techaltar I totally agree with your theory. "Oooooh lool how dangerous [and thus valuable...] that stuff is" has been Open AI's main marketing strategy for a while now. I hate that everyone just goes along with it. You'd think media and marketing literacy would be a bit more substantial by now...