I know you probably already know, but for the love of all that is holy... don't just blindly trust the output of large language models.