The folders-versus-tags argument usually splits along one line. Folders are neat and predictable, good for things with clear categories: legal files, reference material, a project with defined parts. Tags are more flexible but feel like admin, and a lot of people try them, decide they're a chore, and go back to folders.
Both sides are arguing about a case that doesn't cover everything.
The case neither handles well
Some work you can categorize on the way in. You know a document is a contract, so it goes in the contracts folder, done.
Other work you can't. You're reading a paper on the science of hearing and it gives you an idea for a guitar design. Where does that note go? You didn't know it would connect to anything when you started, and you won't know its full set of connections until later. Filing it anywhere means deciding its meaning before you have worked it out.
This describes most research, most writing, and most thinking that is still in progress. The structure is something you discover, not something you set.
What actually helps
For predictable material, folders are fine. Use them.
For the exploratory kind, the thing that helps isn't a better tagging scheme. It's being able to gather everything on a topic later without having remembered to file it under that topic when you saved it. You want to decide the structure after the fact, when you can see how the pieces actually relate.
How Graphia handles it
Graphia doesn't ask you to file anything. Notes and sources sit on a canvas. When you're working on a topic, you can pull together everything related to it by asking, and the AI brings the connected cards back, including ones you'd forgotten were relevant. The relationships between cards are inferred when you want them, not classified when you save.
If your work has a clear, stable structure, a folder tool will serve you well. If you can't predict where your research is going to go, the useful move is to stop trying to.