Rading most comments, people don't seem to be very impressed by it, I ain't either tbh - its not something mind-blowing (I have been doing this with Astra myself just by writing better prompts to make explainer interactive interfaces), but I do think its a good first step towards a better way of consuming information/answers than just reading text-vomits. I wonder if there is a way to train an LLM native to this kind of thing?