Markdown being the most understandable format lines up with my anecdotal experience. But I also was only using OpenAI models at the time. I do think there are a lot of unexplored methods/table transformation tools that we haven't even thought of yet.
I worked at a startup that heavily relied RAG, and this article definitely articulates most of the same issues we ran into. I do think RAG still has its place but it is definitely becoming a harder to justify case.