### Abstract

In recent years we have seen signs of a new technological revolution in information processing, a revolution caused by a paradigm shift to information processing using the laws of quantum physics. There have been significance developments in all optical quantum information processing (QIP) following the recent discovery by Knill, Laflamme and Milburn that passive linear optics, photo-detectors, and single photon sources can be used to create massive reversible nonlinearities. Such nonlinearities are an essential requirement for optical quantum computation and many communication applications. These nonlinearities allow efficient gate operations to be performed. However, such operations are relatively inefficient (they have a probability of success significantly less than 50 percent) and hence are not scalable themselves. This is primarily due to the current state of the art in single photon sources and detectors. Before {\it true} optical universal quantum computation and information processing can be achieved, the efficiency of such detectors must be significantly improved. This is likely to require a drastic change in the approach to detection technology. We discuss a new and novel approach to the problem of creating a photon number resolving detector using the giant Kerr nonlinearities available in electromagnetically induced transparency. Our scheme can implement a photon number quantum non-demolition measurement with high efficiency (>99%) using only a few hundred atoms, and can distinguish 0, 1 and 2 photons. We discuss various applications of this detector and indicate how it can be used to significantly improve the success probability for the linear optical gates.