The big advantages for me are autocomplete, documentation, and simple bug-finding (you passed an Optional[str] to something expecting a str; that’s almost surely a bug). Being able to hover my cursor over methods and get their arguments and doc strings is invaluable. But that’s only possible if the type of the caller is known.
This is too true, too often is the real type signature str. And with all the inconsistency we have in the ecosystem and even standard library it saves you so much esoteric crap.
