Using LLMs to enhance lifetime learning

I heard an interesting conversation between AI scientist Chandra Sripada and Sean Carroll (the physicist philosopher, not the evolutionary biologist) in which both seemed to agree that for the polymath, or even just the layperson with wide ranging interests, LLMs like Claude and ChatGPT can be absolutely fantastic tutors... they can look up, digest, and re-present the state of knowledge on almost any subject, are infinitely patient, and can actually teach things like mathematical procedures and protocols more effectively than classroom learning. (Mindscape podcast, available widely). I have explored this use myself and it accounts for my largely optimistic attitude towards LLMs. (Which does not necessarily carry over to other forms of AI, where we as a civilization will have to make, and make correctly, a whole series of critical choices in the near future if we want the absorption of AI into our general culture a positive thing, and not a nightmare). 

 

______________________________

MUSICA SUMMUM DOLOREM TRANSFIGURAT 

ET SUMMUM GAUDIUM SALTAT

______________________________