json columns pretty much obviated the need for ORMs. It used to be that you'd sometimes have a deep nested thing you really only ever query all at once rather than in pieces, so you'd use an ORM to automate that, but now you can just shove it into json. And then use regular SQL for the relations you actually care about.

Regardless of the number of rows, it doesn't really matter, there are useful cases for where you might be consuming json directly, so instead of parsing it out into a schema for your database, why not just keep it raw and utilize the tools of the database.

It's a feature, not a replacement.

It's been around for quite while, but DB people hate to explain where they got an idea. For all I know Vertica got it from somewhere else; I think postgres got jsonb around the same time.

Now there’s a name I haven’t heard in 10 years. (I’m only tenuously connected to the kinds of teams that use/would have used that, so it doesn’t mean much.)