We know enough about what's out there to be able to make predictions, after all.

As far as I understand (and I could be wrong), you believe that it's possible to construct two different models of "what's out there", both of which will yield good predictions, but which will be incommensurate. If this is true, how can you then say that we "know enough" about what's out there ? Sure, we may have a model, but chances are that there's another model out there which yields predictions that are just as accurate, and yet has nothing whatsoever to do with the first model; thus, we're no closer to understanding what's actually real than we were before. That's not "knowledge", as I understand it, but perhaps you meant something else ?