Recursive STARKs: in addition to proving the correctness of transaction execution, we also can prove (off-chain), the correct execution of a STARK verifier that verified the proof. The on-chain verifier running time is now shorter by orders of magnitude.
h/t
@OhadBarta
@HaymanLiron
@OhadBarta
What is the theoretical limit on this recursion? If I just verify the verification of the verification of the verification of... my log(log(log...n)) becomes effectively O(1). You must be losing something there / seems to contradict uncompressability?