The markdown feature was a great addition that really improved the quality of the responses on R1 and in Rabbit Hole. But I feel like the markdown rendering still falls a bit short. Here are some gripes that I have:
- in rabbit hole it still does not display all kind of markdown properly, like code blocks. This is a big let down as a primary use case for LLMs (especially for me) is to have them help you code.
- Markdown, in the rabbit hole, only renders when the journal entry is for beta rabbit. If rabbit hole can render markdown then it would make sense to just have it render it for every entry, and not selectively just for beta rabbit ones.
- Even things that should display properly on R1 (like bullet points and bold text) does not always display properly on R1 for seemingly no reason. If you ask R1 “beta rabbit, show me some different markdown examples” you can see that most of the time all that markdown does not render properly. Even for simple things like headings it does not render properly on R1 when they should.
If there is any improvement to be made here, I think starting with code blocks and proper text block support in general would be great. What does everybody think? I posted some images below of what I’m talking about:
This block here just didn’t display properly
Here we see a disparity between rabbit hole and R1
