Hacker News new | past | comments | ask | show | jobs | submit login
Inside the Apocalyptic Soviet Doomsday Machine (It actually exists) (wired.com)
48 points by tsestrich on Sept 22, 2009 | hide | past | favorite | 38 comments



This article is a massive bait and switch. The 'doomsday machine' turns out to be... a protocol for delegating launch decisions to some dude deep in a bunker if certain conditions (phone to Moscow goes dead, ground shakes a lot, someone remembered to turn the system on) are met.

Even the British version of this is more unsettling: http://en.wikipedia.org/wiki/Letters_of_last_resort


Actually the British version is less unsettling, it's resting the command on four Captains of the Royal Navy. First they have to confirm that the UK has been destroyed (this means no government or no established military chain of command) then they have to establish the prime minister has been killed. After this they execute their last order, which is whatever is held on the letter, which could potentially be anything. It would likely be targets, but it could equally be orders to contact an ally for appropriate targets prior to launching a blind attack.


I'm reminded of an Arthur C. Clarke story, "The Last Command", where the letter orders the men to... surrender to the enemy; since deterrence has already failed.


Sure, some person has to make the decision, but what happens next is total apocalypse. Works as advertised, IMHO.

(I agree though that the whole 4-page article could be condested to one paragraph)


Which makes it no different from any number of past and present Cold War control loops.


It's a "state machine": a device that can be in one of a set number of stable conditions depending on its previous condition and on the present values of its inputs. Machines don't need to be built out of springs and gears. RFC1149, I'm looking in your direction.


It utterly amazes me that civilians are surprised that military people around the world actually plan for the worst eventualities, even if the likelihood of occurrence is low.

You can plan when everyone is home and safe in their beds, you can't when they shit is actively hitting the fan.


One thing that upsets a lot of my friends (with no military background) is the difference between the military and civilian uses of the word "plan". To the military, plan literally means, make a plan. To a civilian to plan implies an intent to do something. So when the media reports that the military is drawing up plans to do X, they freak out.


Why do you think the "worst eventualities" are so catastrophic, if not because of such planning?


Are you suggesting that planning for bad things causes them to happen? Don't you carry life insurance?


Thats not what abefortas is saying. Its different in the case of larger and larger build ups. Here, each side is causing

1) The possiblity of doing more damage.

2) Which leads the other side to do the same. Going back to 1.

The problem is that building up your arsenal is always better from a strict game theory sense (assuming, of course, the other side is rational).


it seems to me that the possibility this device does not in fact exist deserves more investigation than is given by the article.

for instance, if you assume for a second that this device doesn't actually exist, then the current russian response makes perfect sense... it would be in their best interest to neither confirm nor deny existence of of such a machine. i.e. the benefits of hesitation on the side of a potential enemy that believes the machine may exist far outweigh the dangers of accidentally ending the world (since there is no danger at all).


From the article:

    The Russians still won't discuss it, and Americans 
    at the highest levels—including former top officials 
    at the State Department and White House—say they've
    never heard of it.
From "Dr. Strangelove":

    "Yes, but the whole point of the doomsday machine 
    is lost if you keep it a secret! Why didn't you tell 
    the world?"


Please read the entire article. It explains that the main function of the system was not intended to be deterrence towards the enemy, but deterrence towards the own side's hawks, who might be less eager to launch a nuclear response based on incomplete data when they know that there is an automated response designed to work when all else fails.


I still think it's a little odd that they wouldn't want to simply achieve both goals at once by telling people about it. That would still mean that they would be less likely to preemptively pull the trigger themselves, and at the same time would deter the US from attacking them.

Not to mention, if I (as the US) knew that this system was in place, I would probably be a little less terrified that the USSR was going to randomly decide to attack (since this is clearly meant to be a purely reactive system). This would subsequently lead ME to be less likely to randomly attack, etc, etc.


you are ignoring the fact that if the awareness of the doomsday machine rose above suspicion, then Americans would take steps to attempt to locate and preempt the machine


perhaps, but this is talking back during the cold war (maybe today would be different). I would think that we'd be too afraid to try something like that during the cold war in fear of starting a legitimate war.

But again, maybe that would be different today, and we might not-so-politely ask them to disarm the system


> I would think that we'd be too afraid to try something like that during the cold war in fear of starting a legitimate war.

It certainly didn't sound that way from the speeches of B-rate actor you had in charge of your military at the time.


* I would think that we'd be too afraid to try something like that during the cold war in fear of starting a legitimate war.*

Yes, but the system could be located in advance and disabled as part of a first strike. Even the vague details in the article would give an attacker a lot of ideas on how the system could be disabled -- apart from bombing the bunkers (which is difficult) you could spoof the sensors, or just convince the computer that it's still in contact with Moscow.

And even if the Americans can't find a way to disable the system, the fact that the Russians don't know that the Americans can't disable the system means that the Russians can't trust it any longer and so are back to being more likely to make a first strike.

The optimal solution is probably just to hint to the Americans that you might have such a system. Then again, you also have to assume that the Americans are clever enough to have figured out that you might have such a system anyway, so you probably don't even need to hint about it.

Gosh, Cold War game theory is fun. I wish I could travel back to the 60s and get a job with RAND.


They could also, if needed, point out they have a "doomsday machine" ready to go if they were pressed into it by the US.


The article's explaination is not convincing so the question still stands.


How exactly is "not convincing", compared to the alternative?


That Dr. Strangelove quote is in the article as well, page 3.


"Permissive Action Links. ... So in 1962, Robert McNamara ordered every nuclear weapon locked with numerical codes. Effect: None. Irritated by the restriction, Strategic Air Command set all the codes to strings of zeros. The Defense Department didn't learn of the subterfuge until 1977."

The most egregious default password in history.


Does anyone else feel that one of the most interesting parts of the entire article was introduced in the last sentence?

"After all, he says, Dead Hand is continuously being upgraded."

This seems contrary to the spirit of most of the rest of the article. The fact that is still active is presented as it should be - a surprising fact - but constantly being updated shows a commitment several steps beyond that.


I agree, "Russians Continue to Upgrade Doomsday Device" would have been an even better attention grabber for the article title.



This is kind of a disappointing 'doomsday machine', it's basically an over complicated and/or machine.

I was hoping for at least a stockpile of nuclear weapons in a converted cobalt mine (http://en.wikipedia.org/wiki/Cobalt_bomb). Get a few hundred multi-megaton nukes wired up together and if the country gets nuked it blows, kind of like a dead-country's switch. If you generate enough radiation you can essentially guarantee you wipe out your enemy by wiping out the majority of life on Earth.

Essentially any large-scale nuclear war can potentially render the world uninhabitable. The amount of soot put into the upper atmosphere by the detonation and widespread forest fires would potentially eradicate all ozone, meaning heaps of UVA, UVB and even UVC. The concern is that we only receive 1.3% of the suns UV's thanks to the Ozone layer and can still receive major burns. Consider that it takes ~30-60 minutes on a bright day to get a burn (at least it does for me), after a major nuclear war it could possibly take less than a minute of direct exposure to get a sunburn and it could possibly take 100 or more years for the depleted ozone layer to restore itself.


> If you generate enough radiation you can essentially guarantee you wipe out your enemy by wiping out the majority of life on Earth.

The point of Perimeter in confrontation with the USA was to punish the USA, not to wipe the life off Earth.


Well they're a bunch of pansies, couldn't build a real doomsday device(!)


> 24-old junior officer fresh out of military academy. And if that person decided to press the button ... If/then. If/then. If/then. If/then.

The most deadly group of if/then statements in existence.

"A strange game. The only winning move is not to play." -Joshua, WarGames (1983)


Oddly enough, I'm actually a little relieved that it does eventually come down to a person's decision. I guess if the decision is to end the world, a 24 year-old junior officer is as qualified to make the decision as anyone else.

I'm definitely more than a little concerned that they're still maintaining this thing though.... creepy


As the article says, when all of the information said junior officer has points to the fact that his country has been destroyed, he is just about as likely to turn the key as an if/else statement.


The fact that we'd tit-for-tat several hundred million lives is a saddest reality of the cold war. It's frightening how destructive our revenge emotion can be.


So a possibility just occurred to me. Build the whole device, but disable the key. You can't launch!

It would serve all sorts of deterrent purposes, but in the end, no revenge.

Of course, you'd need a few people to keep secret, but you could probably keep this to 2 or 3 people.


And where are the "else" branches, eh? Is this the final confirmation that multiple method exit points are evil?


At least they didn't use switch/case. Can you imagine forgetting to include a break between case statements?


There's a joke about Duff's (Doomsday) Device somewhere in there.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: