Similar to the 1000 interns question from the article, I've heard:

"if AGI was available via an API tomorrow, what would you build?"

And it's difficult to come up with a great answer. Certainly the summarization, classification, translation, named entity type tasks are a great start (in fact it's the main thing I use LLMs for). But surely there's a more impactful direction you could steer the tech. Even if the resource was just 1000 slightly above average interns.