We only know that information can't escape a black hole by applying GR inside the boundary of the black hole, and by doing so, we find out there's a boundary. I

We only know that information can't escape a black hole by applying GR inside the boundary of the black hole, and by doing so, we find out there's a boundary. If you can't apply GR inside the boundary, you can't generate the proposition that there is a boundary in the first place. It's the logical equivalent of sawing off the side of the board you're standing on, Wile E. Coyote style; it's a load-bearing part of the theory. I think you'd need to find some other way of generating the conclusions of GR (a different board to stand on) if you want to eject the interior of black holes from consideration. EDIT: Oh, and there are people trying to do that, I think, but it's not that easy.