Reading up on what it would take to deploy an LLM that auto fine-tunes on user sessions and we’re still a long way from “consumer ready” with AI agent infrastructure, aren’t we?