Can you briefly explain the advantage of tearing apart the JSON document, rather than just storing it as JSONB and using the various JSONB functions/operators/indexes?
He did explain that - it can yield significant space savings where there is a common "type of document" because the field names do not need to be stored in each document instance, the way they would have to be in JSONB. Also, it means you can query these tables with normal Postgres query tools and they will actually make sense.