I’m getting some replies that look like LLMs wrote them

is this a thing here?

New users having to discern between bots and other real users is an interesting wrinkle to the cold start problem

It’s the worst right now for text-based content networks, but will extend to all new content networks 

Probably need to curate heavily while the network is in its teething phase and edges are still forming

sometimes it's also because people are not english native and using chat gpt to improve what they want to say. when I'm tired, I'm doing this and sometimes, yes, the result is very LLM style 🫠 difficult then to un-improve my text to make it more natural when the message is clearer like this