Hi! I'm missing something very basic here. The core idea of Presto seems to be to scan data from dumber systems -- Hive, non-relational stores, whatever -- and

Hi! I'm missing something very basic here. The core idea of Presto seems to be to scan data from dumber systems -- Hive, non-relational stores, whatever -- and do SQL processing on it. So isn't its speed bounded by the speed of the underlying systems? I know you're working on various solutions to that, but what parts are actually in production today?