I would posit that it's easier to understand things like objects, arrays and API endpoints if you have that basic familiarity with folders.

In many cases hierarchies are just ceremony or boilerplate that slow down a user. Say a note taking application. It can store the actual notes in a database that can sync between machines and the web. Using search to find notes, especially if it supports some sort of inline tagging like hashtags, is faster for the user than trying to come up with some folder hierarchy. Hierarchies also are inconvenient when documents straddle multiple categories. Grouping by tags let a note/document exist in multiple categories without reorganizing everything.

Thirty years ago the only "view" into storage was directories and files. File systems were relatively sparse in terms of metadata support. Directory hierarchies were the only tool for organization. Information Retrieval systems were all about adding an over the top database and document indexing system on shitty sparse file systems.

Today a lot of those Information Retrieval concepts are rolled into online storage systems (Google Docs etc). So hierarchies aren't the primary organization unit since they're not the only mechanism the backing store has for organization.