I can't help but think there's something inherently wrong with #logseq. There's a heavy abstraction layer on top of markdown and as a result, your files are barely usable without the app. Why not use something like sqlite at that point? They made an awful compromise and ended up with something that is neither markdown nor a real DB.
On the other hand, #obsidian is clunkier, but I can open my folder in a text editor and not drown in irrelevant metadata and generated ids.
FYI this is how a page from #Logseq documentation (it's a Logseq graph itself) looks like when rendered by GitHub:
https://github.com/logseq/docs/blob/master/pages/Knowledge%20Graph.md
It looks readable enough to me; also Logseq being FOSS means I will always be able to run it and even modify its source code for the years to come; the company just leads the development but the tool in itself and your ability to read your note are not tied to the company at all.