[ExI] North Korea's super EMP Bomb

[Anders Sandberg]

Regardless of the actual existence and putative properties of the DPRK 
emp-bomb it is a good example of a deterrence multiplier. I have been 
keeping a small eye on big e-weapons over the past decade, and it seems 
to me that back in 2000 the US was working quite happily on them. Then 
interest seems to have faded, probably since it became clear that 1) 
they are pretty useless in the underdeveloped theaters of war the US is 
active in, and 2) excellent for attacks on hightech societies like the 
US. Not a good idea to develop weapons most useful only to one's main 
enemies.

The real problem is that we have no real idea about how bad an EMP would 
be against a developed society because it wouldn't have a simple effect. 
Different pieces of electronics react very differently due to local 
shielding and their function. A big EMP is likely to be bad because it 
will hit at least one crucial part of infrastructure like power, and 
then that failure will produce a series of other bad effects. But the 
real unknown is the systemic effects. It could be that inocous pieces of 
tech in apparently not essential domains are extra vulnerable, and their 
failure cause much more significant disruptions than expected. A sudden 
failure of (say) AEI railcar identification tags or car electronics 
would mess up transport infrastructure a lot - especially if the failure 
is not deterministic in time and space. The combined effect of 
correlated failures across a lot of domains also could have systemic 
effects: computers becoming unreliable would negatively affect much of 
society and require expensive repair/replacement - at the same time as 
the repair/replacement infrastructure is also made vulnerable.

Most weapons work by maximizing entropy one way or another. A known 
target can be hit with a surgical attack that destroys a key 
functionality. The less information you have about he target, the more 
you need to just swamp things in entropy.  Strategic weapons do most of 
their work not by blowing up physically but by existing as a known 
deterrent. The good thing about the cold war nuclear arms race was that 
it was fairly well defined and game theoretically sound (the actual 
impementation of course turned out to be horrifically full of mistakes 
and sloppiness): the entropy *on the strategic level* introduced by the 
existence of a ballistic missile is low, probably much lower than what 
conventional weapons and troops in all their complexity does. However, 
weapons with ill-defined or badly understood effects introduce strategic 
entropy. We tend to avoid uncertainty, especially when it comes to 
losses, so having a high-entropy deterrent likely produces more 
deterrence than an equivalent low-entropy deterrent - at the price of 
making the game much more risky and uncertain. That is fine if you have 
little to lose, which is the DPRK situation.

-- 
Anders Sandberg,
Future of Humanity Institute 
James Martin 21st Century School 
Philosophy Faculty 
Oxford University