> Obviously, I'm extrapolating a long way out, and maybe at some point we'll run against an unexpected wall.

GP has set the 'low bar' of providing a material that survives a series of nuclear blasts whilst generating useful thrust. I'm not qualified to judge whether or not that requires new physics but it seems to me that if we had such a material that we'd be using it for all kinds of applications. Instead, we rely on the physical properties of the materials we already know in configurations that do not lend themselves to the kind of use that you describe.

That's the difference between science and science fiction, it is easy to write something along those lines and go 'wouldn't it be nice if we had X?'. But if 'X' requires new physics then you've just crossed over into fantasy land and then further discussion is pointless until you show the material or a path to get to the material.

See also: space elevators, ringworlds, dyson spheres etc. Ideas are easy. Implementation is hard.