Consider 6 hypotheses:
(1) String theory is the only mathematically plausible way of unifying quantum field theory and general relativity theory.
(2) String theory (with the infinite nature hypothesis) predicts supersymmetry.
(3) String theory (with the infinite nature hypothesis) implies that the equivalence principle is valid for ordinary matter and dark matter.
(4) Einstein found the correct mathematical formulation for the equivalence principle.
(5) String theory (with the infinite nature hypothesis) implies that, after quantum averaging, Einstein's field equations are 100% correct.
(6) String theory (with the infinite nature hypothesis) predicts Einstein's field equations.
Are the 6 preceding hypotheses empirically valid? Consider some ideas of J. D. Bekenstein:
"Literally taken the MOND recipe for acceleration violates the conservation of momentum (and of energy and of angular momentum) ..." (page 2)
"... What principles should the relativistic embodiment of the MOND paradigm adhere to? ...
° Action principle ...
° Equivalence principle ...
° Positivity of energy ...
° Relativistic invariance ...
* Causality ...
° Departure from Newtonian gravity ..." (pages 4-5)
"Tensor-vector-scalar modified gravity: from small scale to cosmology" by Jacob D. Bekenstein, submitted in 2004 & revised in 2005, arXiv.org
I suggest that empirically successful MOND does indeed imply that gravitational energy is not conserved and that string theory with the finite nature hypothesis is empirically valid. (If dark-matter-compensation-constant were equal to zero, then Wolfram's cosmological automaton would not have a uniform timing mechanism.)