This is super cool AND cleverly branded It’s basically “self-authoring RAG at runtime” where the LLM treats long prompts as a data processing problem doesn’t nearly slap as hard as Recursive Language Models!! In this information env, poasting words matter (for research too)