Is There a Use Case for Language Models?

I think this issue of time is super important in thinking about what these models, what these systems can do. I think they read faster than humans, they don't get tired, they're reading thousands of words kind of in parallel,. Even when these systems aren't more knowledgeable or sort of better at some task than humans are, they can still have this huge impact by just being faster. This lets you get a reasonably nuanced understanding of a news article a lot faster than having a human actually read it. That seems like a potentially compelling kind of use case.