x
Understanding the diffusion of large language models — LessWrong