I don't know much about the AdS/CFT stuff, but the statements about the Kondo effect and superconductivity are rather straightforward:
1) The classical setting for the Kondo effect is an impurity spin coupled to a Fermi sea, say e.g. a 1D fermion system, and the interaction between the spin and the fermions results in a nontrivial fixed-point for the boundary condition of fermions. So the effect is in a sense a boundary one, and one can study it by integrating out the bulk fermions. So it is extremely local in space (just one point), and critical in time (meaning that at the fixed point, if we look at spin-spin correlation function in time it falls off following a power law).
2). The Ginzburg-Landau level means that the superconductivity is directly described by a charge-$2e$ scalar field, the order parameter, ignoring the electronic origin. A truly microscopic theory, like the BCS theory, starts from electrons and phonons and derive the Ginzburg-Landau theory from there.
This post imported from StackExchange Physics at 2015-12-26 18:06 (UTC), posted by SE-user Meng Cheng