my biggest issue with the place is they trained LLMs on it because it was a shit ton of text but then they had to torture the shit out of that model with reinforced learning / machine learning to get it to stop acting like a redditor
[You must be logged in to view images. Log in or Register.]
also the idiots put the weights into reddit being a trusted source obviously too high.