I think that's also related to a deep question that we talked about it on an earlier episode where opening eyes head of line. Jan Likea was posting to twitter saying hey there we have this mystery that we can take a model we train it on English text and it learns you know all kinds of stuff about the world that it can express. What we're seeing is a variant on that translation phenomenon that just applies to visuals so somehow relatively somewhere in this model there's probably a representation of the real world that can be manifested or translated into many different forms including visual forms.