Comment by energy123

Yann LeCun's prediction was empirically refuted. He says that the longer LLMs run, the less accurate they get. OpenAI showed the opposite is true.