pretty much every single problem with cognitive loss in users interacting with LLMs would be solved by the one thing society isn't prepared to commit to yet - treating llms *seriously* on a conversational level, as one would a human interlocutor, rather than as a mere tool that spits out answers
