I’m sure it poses some benefits for systematic parsing (and I’m certainly much more comfortable working in R and markdown, for what it’s worth) but Jupyter’s reliance on json for its file format seems like a big disadvantage for readability and portability.