I'm a Physics grad student who just happens to be doing my Master's project on negative index materials (or commonly known as NIM). I'm not an expert in this subject but our reseach group actually discussed this same subject last week. The point here is that the individual photons are not moving faster than light. In fact (what I was told by my professors and others) is that the pulse going in is NOT the same pulse going out. It's the front end 'tail' of the pulse which 'piles up' to become a new pulse which is seen coming out the other end in the shape of the original pulse. The incoming 'peak' of the pulse collapses (actually a portion of it gets reflected which for some reason doesn't ever show up in these simulations of the phenomenon) so only a portion of it exists after going in (I see this in my 1D FDTD simulation all the time). There is actually alot of distortion of the pulse at the interfaces (and inbetween) to the point that it's hard to say what is the original pulse and what isn't. In fact, if you just send light in with no 'peak' you will still detect a 'peak' coming out.