More grifter slop.
Computer programs don’t get “stressed”.
You can tell it was trained with data from the internet because it “thinks” in memes
These are the same “researchers” that let an LLM run a vending machine company. They are just doing this headline grabbing “research” to make the news and go viral. Basically to put their name out and eventually bait VCs and angel investors with fomo.
If the AI proponents are correct and LLMs are sentient — I’m not saying they are, I’m saying if — then using them like this is insanely cruel. What the fuck are we doing?
They do not think or reason or really even communicate. They are a very advanced auto-complete. Their owners actively lie about them to get investor money and they are no more sentient than a Sim.
The way the api are replying in the screenshot (lots of funny memes) show that who programmed that is either malicious (entered a system prompt that encouraged derailments) or grossly incompetent.
IMHO they just want to get headlines
