Tyler Cowen: preference

Transcript context

…Upcoming Conversations with Tyler guest Patrick McKenzie asks, “What has changed the most in your production function since LLMs became widely available? And what will change the most over the next five years, assuming continued improvements?” What I do when I prepare for a guest who hasn’t written a lot — I keep on interrogating the LLM about background and context, and that works very well. I mentioned that already with Lazarus Lake. In general, when I’m reading books, unless I know the area quite well, there are all sorts of references where I just don’t know really what they’re talking about. I just type a query into my GPT-4. Usually, I’ll experiment with the other services too, and it will tell me a lot. I find that is more useful than trying to read more books, is taking one book and the things I don’t know, asking questions. That would be the main difference. Over time, it’s hard to predict. What I’ve already started doing is, I do more personal appearances and give more talks because I think simply writing isn’t good enough anymore. You don’t have to believe the next GPT will write a better book than person X, but I do think many more people will be playing around with LLMs rather than reading books, so we need to diversify more out of only doing books. Has ChatGPT condensed the number of tabs you typically have open? Or has it expanded it? Or has it just substituted maybe one-for-one with what you might Google for?…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.