Maybe I'm just reinventing the wheel and it's time to read Gamov again. Maybe this isn't that different from Einstein's understanding. (I've heard that the ultimate wisdom can be reduced to "maybe".) I think at least the popular rendition doesn't go far enough. Let's consider the shape of the universe and the real reason for gravity.
A bubble (ignoring the thickness of the film) is a two-dimensional surface with a "positive" (spherical) curvature in the third dimension. Suppose there's a particle of iron suspended in the bubble material, and a magnet at the center of the bubble. The pull will distort the bubble toward the third-dimensional center, forming a funnel-shaped indentation. (Does it have a general name?) (Surface tension gives a fluid film elasticity, much like the elastic fabric of a rubber balloon.) You may have seen rigid versions of such a funnel in novelty stores or science museums, in which a ball or coin rolled into it orbits the central hole. They aren't accurate analogies.
Let's bring back "ether", a fluid through which matter, and, in this version, light, moves without friction or any interaction. Think, for the moment, of it forming a bubble with a three-dimensional surface, and forth-spacial-dimension curvature, in which all the matter of the universe is suspended, and from which it can never separate.
[This is the first part I haven't seen mentioned:]
>Gravity isn't simply the pull between two masses in the universe as we know it. There is a force at the 4D center of the bubble which attracts all mass in the 3D universe, and thus distorts the shape of the bubble in 4D. Where there is a concentration of mass, such as a planet, star or black hole, the surface of the bubble is pulled toward the center, to form the funnel shape mentioned above.<
Each smaller cluster of mass forms its own funnel distortion along the slope of funnels created by larger masses in the 3D surface. The popular version of Einstein's theories shows a smaller mass simply sliding (Not rolling because there's no friction with the surface. Rolling would give it a different path.) "down" the forth-dimensional slope of the funnel toward the larger mass (usually orbiting it in the process). [And maybe new from here on:] >I say any two masses tend to move together because merging their two funnels into one reduces the stretch of the surface.
It's hypothesized that matter/energy falling into a black hole might emerge in some other volume of the universe (I don't think anyone, except in the Disney movie, suggests it can keep its original structure). This makes sense in my model also, at least for now, because a black hole would be so deep that it would reach the center and matter could pass through and might emerge from other black holes.
(If you want to get into religion, call the force at the center of the bubble "God", since it "sees" all and we're attracted to it but can't generally reach it from the physical universe. Then the question arises of what sort of "gas" keeps the bubble inflated. Call it the Devil, since it keeps the universe separated, and separate from God. Is the Devil getting stronger, or God getting weaker, to cause the universe to expand?)
But the analogy of starting with a bubble is not accurate. Suppose instead we started with a flat film on a flat frame of infinite dimensions, with just enough gas pressure "under" it to keep it from sagging. This is about as far as our analogy can take us, but it's still not enough. In cosmological terms, the expansion of the universe suggests a 4D universe surface with a 5D "saddle shaped" funnel curvature, rather than spherical or flat, as follows.
Any area on any such a funnel, or the area between two funnels, takes a saddle-shape in the next-level-dimension, convex "up and down" along the surface, concave horizontally (assuming you're looking at it from inside the funnel). No area on the surface can be said to be different in kind from any other, except in its mass content. By the way, on a spherical surface, any circumference is LESS than 2*pi*(surface) radius. Consider a circle on the earth's surface, inscribed between the north pole and the equator. On a saddle surface, C>2*pi*R, circumference is MORE than 2*pi of the radius along the surface.
In a spherical-curved universe, if the "central" attractive force moved closer to one side of the bubble, the distance to the surface would be less than indicated by the local curvature. The mass of the universe would "collapse" toward the point where the bubble was nearest to the attractive force. If the universe was flat, and the attractive force was at an infinite distance, it would be stable, not expanding and not contracting. If the radius was longer than warranted by the curvature, necessarily a saddle shaped surface, the mass of the universe should expand (and the bubble itself might expand).
At this point, I get a bit confused, and I'm not even sure I know why. It seems there must be a separate attractive force at right angles to the (relaxed) surface, at each dimensional level. I'm used to having only one force that attracts all matter without respect to polarity or orientation. Maybe an analogy would be a fine rubber band suspended in the surface of a bubble, with less surface tension inside than outside, to keep the rubber band "inflated". This would be a 1D surface with circular 2D curvature, contained on a 2D surface with spherical 3D curvature, compared to a 3D surface on a 4D surface, on a 5D surface. (Are there infinite dimensions?)
Maybe part of the mystery is the "gas" that keeps the universe "inflated". Is it maybe light-pressure between the stars, or forth-dimensional centrifugal force, or different surface tensions? For that matter, what's the attractive force now?
How far could one go in making a reasonable model, using a large magnet, a sheet of elastic rubber on a round frame, and "soft" iron balls as "celestial masses"? Suppose the magnetic force was applied by many little magnets, spread over a wider area, to represent a force at infinite distance. Suppose the area below the sheet was enclosed and one could vary the air pressure, changing the curvature of the "universe". How about one major stretched funnel, the bottom of which is anchored in place? How would this affect orbits? With strong enough magnets, it would be more a novelty to have the whole assembly up-side-down. (On rubber film, the gyroscopic effect of rolling balls would somewhat distort orbits, as would friction of particles in a real-life fluid.)