If you have any insights about data organization, can you answer me this question? I have the intuition that in a world of fuzzy finders, rich data, burgeoning AI tooling, and hash algorithms, just putting everything in one bucket seems to be the best method of organization for digital goods. Am I wrong?

The reason we don't use this method in real life is because... well, in meatspace, we just call that a pile of trash. But in computer space, binary data is already linear and homogeneous, and we already have the tools to reinterpret that homogeneity into its intended conceptual shape. So why do we use folders? Is that still a useful abstraction? Directories make sense, but only insofar as name extensions, or built-in tags for things. (I don't know enough about filesystems to know if that's how directories are already implemented.) Whenever I've put the effort to section data into arbitrary subcontainers, eventually my method becomes an obsolete encumbrance, no matter how forward-thinking I try to be when implementing it.

A good fuzzy finder or contextual search or sorting algo has always proven to be the best tool to handle organizational complexity, in my experience. Is this wrong? It feels I'd rather have Felix's Magic Bag or a Bag of Holding than a bureaucracy of conjectures holding my stuff hostage.