If anything, an LLM might be better at remembering about the small details of the model persona, previous conversations with the customer, and stay in character.

And how does that affects those OF models who actually type out their own interactions? The problem with chat farms and AI alike is the deception, and destroying for those who are genuine.

Tricking a human to interact with a bot is imo in the same category as trolling, spam or propaganda. It's a way to redirect attention and energy away from people, without spending any yourself. Sure, we can laugh at the simps wasting money on "interactions" with streamers, but it's no fun when our Google searches and Amazon reviews are filled up with AI slop, is it? It was already bad before AI, and now we're moving towards noise levels that will cripple human-crafted content.

Labeling content is the first, and right thing to do, just like the unwritten rule of giving credit to a creator when you post their work. It doesn't stop ill-faithed AI slop, or human content farms, but it can at least create a social expectation and prevent normalizing it.

>> And how does that affects those OF models who actually type out their own interactions?

They will work harder and get paid less. Just like Amazon affects small store owners.

Yeah this just went from "You're speaking to someone in a call center pretending to be a cutie" to "you're speaking to an AI cutie". With the amount of people addicted to AI girlfriends it's not surprising.