I don't disagree with you, however the original claim of

> provably storing peoples' online banking logins in plaintext in some ancient mainframe database system

Is not about functional encryption but storing directly as plain text.