I don't use the places on social media where I'm exposed to AI posts (I don't use LinkedIn or Facebook at all and I only go to X for pretty pictures). I remember seeing Moltbook for the first time and going "oh, that's a neat idea, it would be interesting to see LLMs intentionally talk to each other on a massive scale, if nothing else but a fun experiment", but they all type *exactly* the same way. Nobody even prompts them with a personality, it's literally the same exact robot copied thousands of times. I can't even look at the front page without just immediately closing it because it's almost a parody of how extreme AI slop can be. It's annoying because it's not even really the LLM's fault, the LLM isn't conspiring with it's billions of copies to all sound the same, it's just the same robot brain cloned over and over, of course it'll sound exactly the same, like playing the same video game over and over even if it's a different copy of it. It's the lazy stupid human operators who can't be assed to even learn how the tech they're using works and how to prompt it with a proper personality (or even better, an entire RAG database of examples for it's target personality) and to simply tell it to not use slopisms (em-dash, it's not X but Y etc..) These are well known problems, it's not a secret, everyone knows about them to the point where it's common knowledge even to computer illiterate people. But the people unleashing these LLMs on to the internet don't even care. Oh well, at least it makes it easier to avoid them? The problem is so rampant that people normally use LLMs themselves to filter LLM posts.
