Thank God they went with file name extensions so we didn't have to preface every source .txt file with header content to instruct the editor about what kind of content it would have.
I see images, audio, or video files distributed in zips far too often. You're getting maybe a percent of compression if you're lucky; just distribute the raw files or use a non-compressed bundle format like tar.
Not sure what the original point was but curiously I happened to use file on a an Apple .numbers file recently and found that it was a .zip file in disguise with zero compression.
So maybe the point was that it’s used often as a container format more often than it’s used for compression? Just my (unrelated) general computer work would also suggest this.
You're not missing much. A few modern file types are zips with expected folder structures, especially MSOffice files. But this is nowhere near universally true.
You can open a file in your text editor of choice and if you see it start with PK (for Phil Katz the creator of the format and the original PKZIP/PKUNZIP programs) then it's probably a zip.
There are basically two types of files. Text files and binary files.
Most information are stored in text files so humans can easily understand it, and it's easier to find errors, review, parse. But text storage takes more space than binary files. And many complicated softwares normally need multiple text files or data files, many of them just store them together as a zip file so that it's easier to handle. Examples are .docx,.pptx, etc files in MS Office, try unzipping them and see what they contain. Zipping also has advantages of reducing file sizes.
I think 'most' is hyperbole for dramatic effect / increased engagement. "more files than you might think are actually following the zip file structure" isn't as punchy.