Jupyternaut now chat with a generate output style, I want to know why not use streaming output style？

[johnsGuo]

Problem

Jupyter-AI's Jupyternaut is currently using a generated output style for chatting. I would like to know why not using a stream output style? Because I think streaming output provides a better visual interaction experience.

Proposed Solution

May provider a switch for user to select？

Additional context

When I use jupyter ai to ask relatively complex questions, the LLM model sometimes outputs a bit too much content, and the current output style gives me a feeling of lag

[JasonWeill]

@johnsGuo Thank you for opening this issue! Let's keep the conversation going in #228, which covers this feature request.