6 points
It’s already here and it’s called model collapse: where LLMs’ garbage output becomes LLMs’ new training input.
5 points
What will happen to GPT-{n} once LLMs contribute much of the language found online?
Absolutely horrifying thought. Wasn’t PageRank literally invented to solve shit like this?
This could create a funny circumstance, where LLM companies have to devise methods for automatic detection of LLMs text, but then normal people can just use those methods to filter out all LLM junk.
