You write:
The true, fundamental zeta function is not the linear sum. It is the curved sum: (...something...)
But people are interested in the properties of Riemann's zeta function, not yours.
It does not really matter which one is the "true, fundamental" one.
People have linked Riemann's zeta function to many other properties.
These relations are rigorously established, but only for Riemann's zeta function,
not for your "curved" one, even if it were "better" or "the right one" in some sense.
Also, your "curvature aware addition"
a(+)b := arcsin( k · (a+b)) with k ~ 0.5599...
gives for example 0 (+) 1 = 0.594...
Is this intended? Doesn't this change / invalidate the meaning of 0?
This "addition" has no neutral element (unless you would change to k = 1),
so it shouldn't be called addition, which is a term reserved for abelian group operations,
according to universally agreed terminology.
- Maximilian