merging two proofs together

[mimoo]

@mrmr1993 mentioned that idea to me, if we have two circuits that we want to prove, we could in theory prove them within the same kimchi proof by committing, and evaluating to two different set of polynomials at the same time throughout the kimchi protocol.

At the end, they can be aggregated together in the IPA.

One upside of this is that we can verify two proofs for the price of one recursively (one circuit verifier) and the merged proof uses the max domain of both proofs.

Intuitively, this should work for N proofs, as long as the verifier circuit is aware of this in advance.
It's not clear how much of a pain it would be to implement this, and where it would be most useful.