I get it, but also you probably could w the right intermediating layer to allow the LLM to store & retrieve the original answer as context for each subsequent question? the problem w chat ui is we don't let chat have any secrets 🫶 https://x.com/goodside/status/1723747357278249020
More notes on nearby ideas
Reply · Oct 5, 2024@Return2Mimetic communicating with, but particularly wondering about language for LLM to LLM to LLM round tripping for chain of thought or whateverMachine learningTweet · Mar 4, 2025do your very long context llm chats tend to get...Quote tweet · Mar 23, 2023me n which LLM? 😻 https://x.com/_Xenofy_/status/1639001945200529409Reply · Nov 27, 2023@lumpenspace @neirenoir yea true and i'm conflating w my unrelated complaint; and neither are actually proof you *can't* ... is just annoying bec…