Yes, indeed, in geometrodynamics the gravity field is simply another dynamical entity on a global non-dynamical Minkowski spacetime. Therefore, all dynamical "force" (boson) fields are simply local gauge compensating fields from various global symmetries of the dynamical action of all the lepton-quark spinor fermion fields. Geometrodynamics is simply the local gauge theory of the universal space-time symmetry Poincare group (possibly conformal group as well that is spontaneously broken). The equivalence principle is simply the universality of the Poincare group "kinematical" choice.

All spin 1 boson field theories should have only dimensionless couplings and should be renormalizable in t'Hooft's sense. Gravity is such a theory at the "Dirac square root" (metaphorically speaking) tetrad-spin connection level. The non-renormalizable spin 2 argument is simply asking the wrong question - it's a chimera.
The spin 2 description is not at all fundamental and is completely analogous to the old Fermi 4-fermion model prior to the SU2 W-Boson model with parity violation and the Higgs-Goldstone spontaneous vacuum symmetry breaking.

Why have all the Pundits missed this obvious idea? Does the Emperor really have no clothes?

In addition the emergence of the classical tetrad-spin connection c-number fields is similar to the "Aristotelian dynamic"

v ~ Gradient of the action

in de Broglie dynamics and also in superfluid helium

with the quantum gravity noise described as I say above.