Comment by jmyeet

I've been on the Internet for decades at this point and one thing I've noticed is that communities that, for example, ban political topics actually mean "positions I don't like" as "political". This is somewhat related to the Overton window but really a bunch of (mostly conservative) ideas get normalized so aren't deemed "political".

I see the same thing with "AI Slop". Yes, there is AI Slop but (IME) it's pretty easy to spot. But what's more annoying is how often people are willing to throw that accusation whenever someone takes a position they don't like, much like the "political" label. It's lazy and honestly just as bad as the slop itself because it unintentionally launders the slop in a "boy who cried wolf" kind of way.

I also have a theory that some AI slop isn't inherently successful. It's just heavily botted by people who are interested in promoting certain positions. I bet you could make a pro-administration LLM bot and another one promoting a communist revolution and no amount of model tuning would make the second as popular as the first because the first would hit third-party botting as well as platform content biases (eg Twitter).

I've personally been accused of being a bot. This is particularly true in recent time as I've tried to share facts and fact-based analysis of, say, what's going on with crude oil markets, the military operation in the Gulf and the politics and economics around it. I even saw one hilarious comment saying (paraphrased) "the bots are getting clever and posting about unrelated topics". This was funny because it never occurred to this person that no, it was just a real person posting something you disagreed with.