Mutually Assured Destruction (MAD): If you attack me, I’ll retaliate. To do so, I need to survive a first strike, and have the will to launch an attack that will likely reduce global population by half, or 90% or something like that.
I may be cautious, and trust that enough of my ability to retaliate will survive to erase you and all you love (“launch on impact”). Or, I may be rash, and “launch on warning,” which is an excellent way to (a) preserve the ability to retaliate; and (b) dramatically increase the prospects of an “oops, just a computer glitch. SORRY!”
Surviving your first attack is greatly enhanced by having part of my ICBMs under the ocean. Yes, yes, that makes communications a bit more difficult, but I can always leave word that if I don’t say stop, then the Sub Commander is to assume she is ordered to launch, on risk of court martial. Again, maybe a little inconvenient if communications are disrupted by satellites being knocked out or some other jamming.
All of which assumes that the Russians love their children as much as we do (1960s reference). But … what about the North Koreans?
Anti-Ballistic Missile shields! If I can build enough ABMs to prevent any – not even one! – of your missiles getting through, they I will feel so much safer. You, on the other hand, will not feel safe because now I can launch without fear of MAD. Of course, you’re probably going to do the same, and that gets expensive for both of us. And, we can both keep increasing the tonnage we toss at each other, until we think (not know, but think) we can beat your shield.
Unless, of course, you come up with suitcase bombs, truck nukes, or drone swarms that make it all moot.
Can we talk about this, instead of building bombs?
Please?