That's the point. They are not the rules. QM and GR are approximations. The real rules are unknown to us. It just happens that "in the small", QM seems to be a

That's the point. They are not the rules. QM and GR are approximations. The real rules are unknown to us. It just happens that "in the small", QM seems to be a good approximation that is consistent with experimental results. And "in the large" the same holds for GR. Neither of them works for everything though. That just means we need to find a better approximation. That's not insane. Bits are not an approximation of software "in the small". They are the real building block. We know that because we made them. First made them, then observed how they behave. But QM is a theory created by first observing. Physics is a natural science trying to understand what we observe. Math and CS are not. The objects they care about are conceived by us and observed second.