Self learning architecture for AI
retained text
I am trying to make self learning architecture the idea is that learning is a very noisy process llms already know a lot so when it sees a new document instead of memorising whole thing learn just what it does not know for each document that is new to the llm we can make a document card which will have learning and each documents learnings will be linked to all other learnings so at query time the llm know how it can travel between documents and every step taken would be logged so whenever anything is wrong its original path can be traced back and can be improved upon so in future the same mistake do not happen any feedback or questions are welcome
document fetched 2026-10-21 00:19 · canonical https://reddit.com/r/AI_Agents/comments/1xajjxt/self_learning_architecture_for_ai