I think we can do without people pasting llm generated output into the body of an LLM comment, if people need to provide it for some example, they can do it via a one of the services like https://pastes.io and an outbound link with appropriate description.