Make Vellum less glitchy

This should be a lot better these days, especially as of 0.10.5! If you keep seeing glitchiness issues, please report fresh

yes there is some "lagginess" - and I don't mean LLM response (I'm on the managed cloud instance). E.g. after pressing "Send" in the type box, there's always a noticeable delay until the message actually appears in the chatbot (very noticeable compared to all the other cloud LLMs I've used, claude.ai, chatgpt.com, etc)