I mean, it’s very possible an it was written by “an AI” (an LLM). For all we know the prompt the user gave it was something along the line of “get your pull requests accepted no matter the cost” and it’s fancy text prediction decided, in it’s ever ongoing roleplay, that the targeted blog post would shame the developer into accepting it’s PR.
I definitely don’t under the paranoia though. I don’t understand how people are convincing themselves any of this so close to actual intelligence. Ask your fancy LLM how to fix your cup that "is sealed at the top and “open at the bottom” or if you should drive to the car wash to get a car wash if it’s only 100ft away - both scenarios obvious to most any human and will need to be trained out of the current leading LLMs (if they haven’t been patched already).
I mean, it’s very possible an it was written by “an AI” (an LLM). For all we know the prompt the user gave it was something along the line of “get your pull requests accepted no matter the cost” and it’s fancy text prediction decided, in it’s ever ongoing roleplay, that the targeted blog post would shame the developer into accepting it’s PR.
I definitely don’t under the paranoia though. I don’t understand how people are convincing themselves any of this so close to actual intelligence. Ask your fancy LLM how to fix your cup that "is sealed at the top and “open at the bottom” or if you should drive to the car wash to get a car wash if it’s only 100ft away - both scenarios obvious to most any human and will need to be trained out of the current leading LLMs (if they haven’t been patched already).