Steven Byrnes comments on Google’s new 540 billion parameter language model