Due to the limited width that the assistant's responses are shown in there is much dead space left on screen and when the assistant/ai/llm outputs code blocks, they are often wrapped in a way that makes them more difficult to follow than would otherwise needed.
Attached above is an admittedly extreme and oversimplified example to demonstrate what exactly I mean.
Im no UI/UX designer, so I can't really suggest an exact layout, but a seemingly obvious way to remedy this would be somehow use some of the dead space and extend the width of assistant outputs. This would allow code outputted to be wrapped less often and thus more readable more of the time
