I would like to just have a storage engine that can be very aggressive at deduplicating stuff. If some data is redundant, why am I storing it twice?

That's already pretty common, but the goal isn't storing less data for its own sake.