1. When you place an accelerometer into free fall, it will measure exactly zero at all times, irrespective of how it falls. Therefore, no forces act on it. Ergo, gravity isn’t a force of any kind.
2. You can place (electrically neutral) test particles behind appropriate shielding that blocks out all (or at the very least most of) the ambient radiation hitting it. This has demonstrably no effect at all on how they behave gravitationally. Ergo, gravity has nothing to do with ambient radiation pressure differentials.
3. Newtonian radiation fields (ie vector fields) of binary sources are dipole in nature, and thus their polarisation states differ by 90 degrees. Real-world gravitational radiation on the other hand is quadrupole radiation, with polarisation states inclined by 45 degrees. Ergo, gravity isn’t a Newtonian force field.
4. Ambient radiation (being mostly photons) cannot and does not couple to rotational angular momentum of a near-by central body. In the real world though, angular momentum does very much have specific gravitational effects. Ergo, gravity has nothing to do with ambient radiation.
5. Real-world gravity distorts shape and volume of freely falling test bodies, which ambient radiation differentials cannot do. Ergo, gravity is not radiation differentials.
6. Ambient radiation does not dilate the relative rates between clocks, as real-world gravity does. For example, electrically neutral unstable elementary particles such as the Z-boson do not couple to the photon field, and yet their lifetimes are dilated in the same way as any other clock under the influence of real-world gravity. Thus, gravity isn’t just radiation pressure.
And many more…
This is nonsense - “radiation” in this context is just electromagnetic fields, which can be measured to extraordinarily high precision. Using SQUIDs, you can (e.g.) measure magnetic flux densities on the order of 10^-18T. So we have a really good idea about ambient radiation environments. And by shielding it - which is simple enough to do - you can check whether this has any effect on gravity, which it evidently doesn’t.
Furthermore, we can directly measure the gravitational effect of small-ish everyday objects on each other just fine, even using simple table-top setups like the Cavendish apparatus (we did this as a project when I was in high school). You can buy DIY kits for this and try it out yourself at home, in fact I would very much encourage you to do so. You can also place that same setup into a vacuum chamber surrounded by appropriate shielding, and find that the gravitational effects do not change at all.
In fact, even just wrapping 2-3 layers of standard kitchen aluminium foil tightly all around the Cavendish apparatus in a way that leaves no gaps and no conducting paths will reduce the internal ambient EM flux densities by roughly a factor of 10,000 (~80dB of shielding or so), and thus any gravitational effects should reduce accordingly. You can (at least in principle) try this at home, and what you will find is that that is not what happens - there will be no changes at all in gravitational attraction. Thus, the notion that gravity is just ambient radiation pressure is a very bad model, since it bears no resemblance whatsoever to what we actually observe in the real world.
Spacetime is a model - we take the description of some real-world scenario (like e.g. two bodies beginning a free-fall toward one another), put it into the model, do the maths, and out comes a prediction of how this system will evolve. We can then check this prediction by comparing it with what actually happens in the real world, and it will either turn out to be correct, or not. This is very much in accord with the scientific method - if the predictions are correct, the model is good; if not, it needs to be amended or discarded. A model in physics does not in general make any claims about ontological truths; it claims only that its predictions accord with real-world observations, and thus that it is valid in that specific sense. In the case of GR, the claim is that the mathematical entity of a semi-Riemannian manifold with curvature that is constraint by a specific relationship between metric and energy-momentum distribution shares the same dynamics as those observed for real-world test particles under the influence of gravity - there is a specific mapping between these two, but not necessarily an ontological identity. There might well be an identity, but that isn’t the a priori claim here, and in any case largely a philosophical question.
You on the other hand don’t have anything at all - it seems you don’t even have a mathematical framework that allows specific numerical predictions to be made, and thus you don’t have anything that can be subjected to the scientific method in the first place. So you aren’t doing science, you are just wildly guessing - and it appears you are doing this solely based on your personal dislike of GR, which, in my humble opinion, is a really bad reason. We want to move forward with better models, not regress back to LeSage et al, which we already know cannot work for fundamental reasons.