THE MANY WAYS THAT DIGITAL MINDS CAN KNOW

Summary from the TLDR newsletter

Large language models do both of the things that their promoters and detractors say they do, at the same time, on the same prompts, and while it is difficult to tell which they're doing, both of them are useful.