Proving AdS/CFT

Have you ever tried to directly prove that type IIB string theory on a strongly curved AdS5 x S5 gives you a weakly coupled gauge theory, as required by the AdS/CFT correspondence? I've been looking into this question for some time, but there is a new idea around. If you have something relevant to say, it may be appreciated.

Sorry for the reduced traffic. The things I have to do now are mostly wrong topics for blogging.