The key issue is how little of the radio waves are reflected.
Indeed, I couldn't agree more, and this is a property which rarely seems to be mentioned in meta-material discussions. Supposing that such a material passed 95% of the energy undisturbed and only reflected 5%. I think that this would rightly be regarded as an excellent technical achievement, and after all some glass isn't that good, but it would seem to be of limited military value. As we know from the radar equation's R^4 term, this will only reduce detection range by a little more than 50%, and you can easily buy commercial radar absorbing materials which will do much better than this.
Of course, your absorbing/transmitting solution also has to meet many other criteria. It needs to work over a wide range of frequencies, polarisations and angles; be rugged enough to survive on the outside of a military vehicle without needing maintenance every five minutes; not be so heavy that your aeroplane can't take off and most of all, be affordable (I appreciate that military affordability is not the same as normal affordability, but never the less budgets are much tighter these days). Of course, progress is probably being made in all these area's but my feeling is that there is still a long way to go.
My other big problem with meta-materials is seeing how they can possibly be applied in real life. As I see it, they are basically being used to make a sort of cunning lens which diffracts the wavefront around the object to be concealed. Sadly, the geometry of the situation seems to result in devices which have a high degree of symmetry (for example, cylinders). I can imagine how this might be extended to a sphere (great if we need stealthy cannon balls) or maybe even a prolate spheroid (which perhaps we could fashion into some sort of missile), but to deal with the complex shape of an aircraft in such a way that the different parts of the wavefront passing around different parts of the wings and fuselage all add up correctly in phase over a useful range of angles seems to me to be an extremely difficult problem.
Still, I am not going to sit here and say that it can't be done. The people working on these problems are no doubt very brainy and may come up with a practical solution, but presumably the paint and shape solution is also being advanced and it is my feeling that this older simpler technology is going to stay ahead for a considerable period of time.