Can you build a solver from a verifier?
I see that if you start with an NP-verifier TM the answer is yes, you can build a solver TM. How about for circuits? Can you go from a circuit that implements a verifier for instances of size n that is of size polynomial in n to a circuit that implements a solver for instances of size n?