Against AI Understanding and Sentience: Large Language Models, Meaning, and the Patterns of Human Language Use — LessWrong