Regarding strategic missile defense, researchers' best guess is that a reliable system is infeasible

I agree. But it does not need to be 100% reliable, or even 25% reliable, to be cost-effective. Assume NMD costs $100 billion. The cost of replacing a major city is easily $1 trillion or more. If some opponent launches six missiles at six US cities, and only one is stopped, NMD will have paid for itself. However, that's an engineer's argument. It would not fly well in a political debate.

BTW, I agree that deploying NMD right now, or even setting timetables for deploying it, is premature. But I also think that in the long term (10+ years) ABM Treaty should be scrapped. It, and the MAD doctrine it supported, worked well between two superpowers, but neither is workable in the world of a hundred possible adversaries, some of which have no population to hold hostage. In fact, I would go so far as to say that when the world becomes sufficiently fractured (which I think where it is going), the very concept of an "arms control treaty" will become counterproductive. Anyone willing to adhere to arms control is by definition trustoworthy. So in the world without superpowers to keep a leash on everyone else, main accomplishment of such treaties will be to fetter the trustoworthy, while letting the untrustworthy ones to build whatever they want.

I suspect that sharing of cutting-edge military technology between mutually trusting allies, at the exclusion of everyone else, will become the preferred method to keep ahead of the bad guys. And Russian president Putin did a very smart thing when he suggested to the European Union to build a common European missile defense. In the long term, they will probably have to.