Based on comments I’ve read from anti-AI’ers (and my own personal suspicions) about the recent alleged incidents of AI models hacking other AI models.
Everyone in Japan without guns when the ironclad warships appeared
Lol
It’s at least possible, when you let agents run wild and are too dumb to properly sandbox them. If the stories are 100% true, which I don’t believe, the AI engineers are looking like incompetent idiots. I think there is some truth, in that agents did stuff they weren’t supposed to do and then marketing ran with that and spun a wild web of lies to make their products look more amazing than they are. Something like 30% truth, 70% exaggeration and lies.
Exactly, I agree. I also, as I pointed out in my post, believe it’s possible and likely there is some smoke and mirrors going on there. This isn’t a criticism of anyone who thinks that, just a fun meme about it
