When LLMs become better than humans at the following: 1. Knowing what you don't know 2. Knowing who is likely to know 3. Asking the question in a way that th

When LLMs become better than humans at the following: 1. Knowing what you don't know 2. Knowing who is likely to know 3. Asking the question in a way that the other human, with their limited attention and context-window, is able to give a meaningful answer