Tag
This paper presents hypotheses on how chatbots function in problem-solving conversations, arguing that LLMs encode artificial metaphorical problem propagations and cannot match human cognitive flexibility, aligning with Yann LeCun's views.