"The process was rigorous, thorough and fully governed by a checklist that was, to our knowledge, without defect. The room for human error was minimal."
So, this would be easy to automate. I often wonder whether part of the reason both the USSR and the USA put people in this process was that the hope was that those humans would not go through with real orders.
If the idea was that it (also) would help in detecting cases where the electronic parts of the system fail, that is clear indication that those in charge thought the humans in would sometimes deviate from the script.
Having humans inbetween the automated steps reduces potential errors introduced by bugs or compromised software. If the last step in Cryptolocker was "Now have the user manually enter in terminal commands to encrypt each individual file", the virus wouldn't work at all. Requiring two human beings to agree with each other that an alert is legitimate and for them to both need to confirm coded messages, target the missiles and to simultaneously turn four keys at the same time - it's a pretty big meat-based firewall for malicious or buggy code.
Not that what runs the silos is really that easy to mess with. They still do software updates with eight inch floppy disks.
In the age of the internet, having safe-guards from hacks and redundancies in place to ensure that messages are not misread seems like standard engineering IMO. AFAIK, every nuke that was ever launched was supposed to be launched, and we haven't had a single misfire, or lack of launch, that I'm aware of.
I do not see how adding redundancy to such a tiny part of the command chain would help much. You would have to make all parts of the system redundant. For example, were the messages sent in duplicate via two separate channels?
People do all these extra checks that computers don't do though. Like, if it's a sunny Tuesday afternoon and all is well, when suddenly launch codes arrive, the humans are probably going to say "ummm... I don't remember nuclear war being on the agenda today. Lets get the commander to call the president." That may be in complete violation of the protocol, but it could be the right thing to do.
Humans are more apt to see that something is amiss than computers are.
Yup. This is why these guys get classified intelligence briefings before going on duty (in this article the ones that had been five minutes long and are now hour-long powerpointfests).
I'm always wondering what PR is behind articles of journalists visiting nuclear silos. This one looks like a first-person account, so was a military allowed to write about his job? How comes he writes so well? He could have been a journalist... Also, was the piece read, checked and corrected by a commandant before being published? If so, which parts were left out? Why did they leave the part where he admits not dressing with the uniform, installing a hammoc, playing video games and so on? Last question, he describes his role as reading a console, decrypting and triggering a button at request, and at the same time saying human error is the most probable cause for a nuclear winter: Why do we need a person to transcribe a machine message to another missile-launching machine? Couldn't both machines be wired up together? Sure it would allow hacking the launch of the missiles, but at the same time, if he receives the correct message, he's going to send them anyway...
The good judgment of Stanislav Petrov shows that this isn't guaranteed.
Automation that can harm people always requires a human fail-safe. Leaving judgment to a mechanism only implies either hubris that assumes all possible situations have been accounted for, or enough narcissism to not care about the (possibly deadly) errors.
It's more that we don't have strong enough decision making computers/algorithms to rival a human's capabilities in cases where the action (in this case not having lots of weapons but apparently only one being launched) isn't exactly a binary decision. A computer could theoretically have advocated that this wasn't an actual nuclear launch, but the only way that would happen is if a programmer had been clever enough to have written an algorithm to account for the situation when writing the transcription machine.
A human can currently do a much better job when the outcome is that millions of people die , especially if its a sensor error. That's not to say that humans are infallible, just that we do not currently have the level of AI required to replace one.
In Dr. Strangelove, the absurd behavior of buffoonish humans contributes to the doomsday scenario. Idiots blindly following the orders of a madman on one side, and psychos constructing a completely automated self-destruct mechanism attached to a retaliatory weapon on the other side.
Both are bad, but people playing the game at all, were the real problem.
With great reluctance, Eisenhower agreed
to let American officers use their nuclear
weapons, in an emergency, if there were no
time or no means to contact the President.
Air Force pilots were allowed to fire their
nuclear anti-aircraft rockets to shoot down
Soviet bombers heading toward the United
States. And about half a dozen high-level
American commanders were allowed to use far
more powerful nuclear weapons, without
contacting the White House first, when
their forces were under attack and “the
urgency of time and circumstances clearly
does not permit a specific decision by the
President, or other person empowered to act
in his stead.”
Unbeknownst to both Kubrick and George, a
top official at the Department of Defense
had already sent a copy of “Red Alert” to
every member of the Pentagon’s Scientific
Advisory Committee for Ballistic Missiles.
At the Pentagon, the book was taken
seriously as a cautionary tale about what
might go wrong.
A decade after the release of “Strangelove,”
the Soviet Union began work on the
Perimeter system - a network of sensors and
computers that could allow junior military
officials to launch missiles without
oversight from the Soviet leadership.
Perhaps nobody at the Kremlin had seen the
film. Completed in 1985, the system was
known as the Dead Hand. Once it was
activated, Perimeter would order the launch
of long-range missiles at the United States
if it detected nuclear detonations on
Soviet soil and Soviet leaders couldn’t be
reached.
Also some of the quotes from US military leaders at the time sound like something from Stranglove:
"Restraint? Why are you so concerned with saving their lives? The whole idea is to kill the bastards. At the end of the war if there are two Americans and one Russian left alive, we win!"
> Accounts differ as to the degree of automation of Dead Hand... more recent sources indicate the system was semi-automatic. In a 2007 article, Ron Rosenbaum quotes Blair as saying that Dead Hand is "designed to ensure semi-automatic retaliation to a decapitating strike." Rosenbaum writes, "Of course, there's a world of difference between a 'semi-automatic' doomsday device and the totally automatic—beyond human control—doomsday device."
Why does it surprise you that someone in the military can write so well? There's a great history of brilliant writers and poets coming from the military.
Many people join the military before doing something else with their lives so you get a tremendous mix of talents and passions in units. I found that a surprising number of people turn out to be excellent sketchers and painters when working in the military, for example.
Yep. Some of the best writing most of the world never sees happens in internal memos in large organizations. Leaders need to be able to communicate clearly, memorably, etc. Any decent manager typically learns to write well in order to get shit done.
Likewise, one of the most poignant commentaries on life and death I've ever read came out of the military. I wish I could find it again, it was on HN within the last year or so I believe.
"John Noonan is a policy advisor and defense writer. He served as a Captain in the United States Air Force, assigned to the 321st Missile Squadron in Cheyenne, Wyoming."
I wonder if this will ever change now. As technology goes forward, we're creating a world in which an individual can gain more and more destructive power, and in which when something goes terribly wrong, it can wider and wider consequences. We're at global level now, and fast approaching extinction-level.
Two areas to highlight the point:
- High-energy anything. There's a saying that any propulsion system that could allow us to travel between planets in a reasonable (for human) timespan is also a weapon of mass destruction. And it's not just space travel - history of technology in general is about getting, storing and making use of bigger and bigger quantities of energy.
- Self-replicating system. The best-known example of these is also known as "biology". Something we're only beginning to play with on a serious level. Mishaps, or worse - deliberate malicious actors by crazy individuals - could have planet-wide consequences as our ability to manipulate biological nanomachines grows.
Sleep tight :).
(Also, alt-text from https://xkcd.com/728/ - "Maybe we're all gonna die, but we're gonna die in really cool ways.")
Of course a lot of the risk comes from the trend for technology to operate at larger and larger scale to achieve efficiency. We could see a trend in the opposite direction where individual households and communities can be more self sufficient. For example, efficient small scale energy production could reduce the risk of nationl level grid disruption. That same biotech industry could enable local food production without needing to depend on complex infrastructure for evey meal.
I agree those may be desirable directions for development, but they don't change the overall dynamics - we expand our capabilities by dealing with more and more powerful, and thus dangerous, stuff. As individual demand for electricity increases with each appliance we incorporate in our lives, the household battery/generator of 2050 could very well be powerful enough to level a city block if weaponized. Even more likely, the same biotech industry that can enable local food production will enable creation of new, deadly pathogens.
Humans, unfortunately, are very optimistic when it comes to workplace and household safety, while at the same time extremely resourceful when they need to make weapons out of mundane objects.
On the other hand, if a lot of countries had apocalyptic weapons (a phase that naturally precedes such weapons becoming available to individuals), it could all but eliminate war. Countries behave surprisingly like rational agents in international relations. This period would be the window of opportunity for one lucky winner to bootstrap an AGI.
The US and USSR were moments away from war several times. As Pakistan and India show, being nuclear armed does tend to end conventional wars. But as more countries become nuclear armed, the higher the chances one of them will actually use them.
When Saudi Arabia and Iran are both nuclear armed, how long before one of them strikes first, even if it is at or through a proxy?
The Iraq Iran war was filled with incalculable human tragedy and horror, of which both sides were perfectly willing to accept. Both the people who chained children together and forced them to charge machine gun nests and the people who mowed down children chained together from their machine gun nests.
If we ever achieve some form of reasonable, in-system spaceflight, then we are really going to have to start worrying about every crazy person with a spaceship bumping a smallish space rock into an impact orbit with earth. Or even just unfortunate accidents. But it wouldn't take much of an asteroid to burn down through the atmosphere and create a Tunguska-level impact.
The most recent Expanse novel is an interesting take on how apocalyptic such an event could be http://amzn.to/2aD2AFl
Sure, but intrasystem travel will only be economical if we can mine and get fuel, energy, and mass from non-Earth sources.
If we get to that point, scaling up just consists of waiting a few decades of exponential growth to kick in. It might not even involve humans, but just letting automated robots continue to build more robots.
If the technology is there, moving mountains instead of ships is just a small jump.
Exactly. We need to find ways to manage powerful technology socially, or we are going to have serious problems in the near future. Unfortunately, I'm not sure[1] our species can adapt fast enough.
Thanks for linking to your old comment, it's pretty insightful. I agree that we seem to be accumulating technological advances faster than we're maturing as a civilization to handle them.
I don't have an ad-blocker and didn't even notice there were ads until I read your comment. Why bother with the effort of denying them a little revenue when their ads aren't even intrusive at all?
Air Force silo jockies, tho I do suppose the nomenclature is more commonly used to refer to submariners. All i know about submariners is that the movie Down Periscope is more accurate than satiracle.
there is an argument by some that the rise of suicide bombing as a weapon of war meaningfully changed the parameters of mutually assured destruction, because there are people entirely comfortable/supportive of their own destruction for a cause.
This outlook seems historically ignorant. In the midst of the very first use of nuclear weapons there were hundreds of suicide attacks in the form of kamikazee suicide runs by aircraft, ships, and soldiers.
One of the likely reasons no terrorist group has tried too hard to obtain a nuclear weapon is they know whoever it is they fight for will be targeted for in kind retaliation. Aka MAD.
Why do we still pretend nuclear weapons are necessary when the only military power that still reserves the right to use them in a first strike is also the only power to have forces deployed all over the world?
Russia has dropped No First Use policy in 1993. Besides relying on a written statement from a potential adversary is silly. The capability is always there.
So, this would be easy to automate. I often wonder whether part of the reason both the USSR and the USA put people in this process was that the hope was that those humans would not go through with real orders.
If the idea was that it (also) would help in detecting cases where the electronic parts of the system fail, that is clear indication that those in charge thought the humans in would sometimes deviate from the script.