So, It leverages on HTTP range requests to try fetching only what's needed, but to me it seems that it's something that should be efficiently achievable without having to have a unified document.

In my filesystem I could arrange things to be easy to fetch

ls ~/Org/Social/2025/10/

and then fetch the documents I didn't have.

I feel that all this logic is just to workaround problems that may arise when fetching multiple small files, mostly tied to overhead and artificial data ordering when fetching a set of tiny files. Hasn't HTTP3 fixed most of them already?