But then at this point we can take a username, take a user's posts on one site, train a LLM with these posts and ask the LLM to write comments in the style of that user on another forum/subject.
How do you even determine anymore if something is really written by someone?
Websites are already for a huge part written by bots/LLMs and we all know to take them with a huge grain of salt.
How long until we consider users posts aren't to be trusted anymore either?
It already started (impersonating usernames) for sure.
So what is this even tracking?
Heck, at this point it's nearly a guarantee we already have bots trained on outputs of other bots.
I wonder what the implication of all this is going to be.
How do you even determine anymore if something is really written by someone?
Websites are already for a huge part written by bots/LLMs and we all know to take them with a huge grain of salt.
How long until we consider users posts aren't to be trusted anymore either?
It already started (impersonating usernames) for sure.
So what is this even tracking?
Heck, at this point it's nearly a guarantee we already have bots trained on outputs of other bots.
I wonder what the implication of all this is going to be.