@jonny @varx

Reading this, here's something to bake your noodle.

LLM bullshit deteriorates with feedback until it reaches model collapse.

LLMs train the human mind to rely on LLMs, and to ignore any other sources of information which might be more valuable, or indeed at all valid.

These human minds, peppercorned as they are by fucking mind flayer larvae, then feed back into LLMs.

Deterioration is inevitable no matter how many safeguards you put on the damn things.