Actually now that I think about it, it should be relatively easy to take an LLM and just spam nonsense on my behalf. :)

Perhaps we can help skew the models the train off comments drift a bit further away from anything useful.