It's hard to figure out what he's talking about , when he says the "whole social security database". Like in which tables are they duplicated? Does it mean the entire row is duplicated or just the SSN, it might make sense to be duplicated depending on the schema. Is it an append only db, so there might be updated columns on the same ssn and you need to filter by the latest update timestamp? Who knows.

But also, saying that there's a "social security database" and then following that up by the govt "doesn't use SQL" so.. the db is actually just a spreadsheet? A .txt file? The SSNs are just written down in someone's notebook? Lol