It's not a new theory, it's a new valid solution for the Einstein Field Equations.
http://archive.ncsa.illinois.edu/Cyb...n.field.sm.gif
That's the simple, symmetric, basic as shit form, matter tells spacetime how to curve, spacetime tells matter how to move.
If you want to model an actual spacetime, you need more dimensions, and thus less symmetry is possible.
That is the EFE's for 2 dimensions, going up to 3 or 4 winds up with thousands of terms, and is something which is ridiculously hard to find a solution for.
You can set programs running on supercomputers that scan for valid solutions, like trying to solve algebra problems by plugging numbers til one fits. If x-5=0, the computer goes x = 1? x = 2? x = 3? x = 4? x = 5, stop!
Now imagine going through thousands of values per line, thousands of lines per equation, etc.
If this is an exact solution, it is fascinating, but it isn't QUITE like Kaylia is saying about GR needing to be totally thrown away or some shit, GR is still a workable approximation to reality to some degree, just as Newtonian Mechanics is for low energy levels.