I. The Bargain
Imagine a fantasy world inhabited by both humans and ancient, unspeakable evils. The ancient evils ravage hundreds of kingdoms every generation, taking countless millions of lives year after year. Eventually, the humans learn to protect themselves, lessening their casualties, driving the ancient horrors back. Some are even vanquished entirely. But it is slow, and uneven, and a single mistake can undo decades of grueling progress. Millions still perish, but fewer each year, and bit by bit humanity inches towards victory.
One day, a band of human renegades attacks the kingdom, slaughtering thousands of innocents. The kingdom retaliates and hunts the renegades across the world for years, dedicating a substantial fraction of their resources to wiping the enemy out. But the leader eludes them – even years later, when the renegades have been all but completely crushed, the once-leader still taunts them from the shadows, and the citizens cry out for vengeance.
One of the Ancient Horrors, on the brink of total annihilation, hears of this, and senses an opportunity. From corrupted shadow, a broken whisper hisses in the ears of the kingdom’s leaders: “I offer you this exchange, mortals of the kingdom: I will give you one chance to annihilate your sworn enemy, the one whose blood you so crave – and all I ask in exchange is that your clerics lower their defenses in the human lands where I yet remain.”
The leaders confer among themselves.
“Er, I’m not sure this is a good idea,” says one. “This horror has killed millions, far more than the mere criminal we hunt, and we are so close to wiping them out entirely. Best case, we lose years of progress and hundreds of victims that we could have saved. Worst case, we’ll risk a full resurgence and millions of deaths if the horror comes back into its full power. Do I even have to say this? It’s obviously not worth allying ourselves with an ancient unspeakable evil whose voice is the essence of death just to kill one lousy human, no matter how awful they are. Right? We all agree, that’s just obviously stupid, right?”
II. Revealed Preferences
The US spends $16.6 billion on counterterrorism efforts every year. This only counts federal spending, within the intelligence agencies, explicitly marked “counter-intelligence”. It does not count, for example, increased costs from waiting in TSA lines and missing flights, increased costs, injuries, and fatalities from people driving instead of flying, wars, or any military spending. The actual number is probably much higher. Terrorism, in turn, kills fewer than 10 people in the US most years. Once every…actually, just once, there’s a 9/11, which kills 3,000 people. But maybe this is misleading (it’s probably not) – maybe all that counter-terrorism money is incredibly effective at countering terrorism, and stops one 9/11 every single year. That comes out to about a bit over five million dollars per life saved.
Motor vehicle accidents have killed more than 30,000 people in the US each and every year since 1945 – usually way more. The entire budget of the National Highway Traffic Administration is $815 million. That’s for specifying, inspecting, and approving every car part and every stage of manufacturing, shipment, delivery, and maintenance. Let’s say that all of those things – all the crash tests, all the inspections, the entire license plate system, reduce the number of traffic casualties by 1/10 of what they would otherwise be, and that there are no other positive effects from this agency, and we’ll use the lowest recorded number of fatalities since WWII as the baseline. That comes out to about $250 thousand per life saved.
I’ve tried to bias these estimates towards counterterrorism effectiveness as much as possible, but I still ended up spending 20x as much per life saved on terrorism rather than traffic safety. It seems like we treat being killed by terrorism as at least 20x worse than being killed in a traffic accident.
Why? What makes us so much more determined to fight terrorism than traffic accidents?
III. Anthropomorphism
Enemies are fun.
Fun might not be quite the right word – but enemies are certainly more interesting.
We have a cultural rule that stories almost always have to have an antagonist to root against. Oh, that guy gained spider powers and can now swing around the city stopping crime and improving the world? I guess that’s pretty interesting, but it’s not really worth seeing unless you also have an insane scientist with giant metal octopus arms to fight against. We’re wired to find enemies to be fun. Protagonists fighting against something that’s broken isn’t nearly as engrossing as protagonists fighting against evil. That’s why we get really into sports rivalries. That’s why outrage addiction is a thing.
It’s easy to imagine a good game-theoretic reason for this. When we do have enemies, we want them to know that hurting us is a bad idea. One way to do that is to make them believe that if they attack us, we will fight back way out of proportion compared to how we would handle other threats – their belief that we will overreact is useful to us. And the best way to make people believe that is for it to be true – to really be willing to go to extraordinary lengths to fight back against and thwart our enemies.
And beyond that, having enemies is memetically adaptive. But there’s more than just a feedback loop between two social groups going on – you can get similar effects with an enemy that doesn’t actually exist.
IV. The Enemy That Wasn’t There
You know how humans tend to see faces on everything, even when there isn’t one?
It’s because we’re all walking around with these hyper-sensitive face-detectors in our heads, and it’s really easy to give them a false positive. Detecting faces was and is a key part of surviving as a human. And so for enemies: Knowing that someone is out to get us is an incredibly important skill to get by day to day. And sometimes we’ll get false positives on that, too. If it happens constantly, we call it “paranoia”. The Satanic Ritual Abuse Panic of the 1980s gave millions of people a horrifying, if imaginary, enemy to unite against, and it was extremely effective. The entire “Witch Hunt” pattern is basically this – desperately seeking out an enemy who you believe just has to be there.
Having extremely-sensitive-enemy-detectors in our brains makes sense: Enemies are motivated to conceal themselves, so we put extra effort into detecting them – and early humans who were bad at detecting the growing conspiracy to cast them out of the tribe probably didn’t have amazing life expectancies.
Combine this with the aforementioned adaptive overreaction, and things can go very wrong very fast. There’s no amount of overreaction you can give to scare off an imaginary enemy. And this gets way worse when your enemy detector is dealing with a miniscule signal-to-noise ratio, as in an outcast arms race.
So, in a shocking twist for this blog, I’m claiming that a lot of things are broken.
How can we put this to work for us?
V. Giving the Enemy a Face
If we could ask knowledgeable experts in human behavior modification how to get people to pay attention to a neglected area, what would they say?
The “Mayhem” series has won over 80 advertising awards, Since launching in mid-2010, Allstate’s stock has more than doubled.
This blog’s thesis is that our truest, worst enemies do not have faces. They are the incentives and systems that optimize for things that we don’t care about. It’s disease and coordination problems and maladaptive thought patterns. It’s bad defaults and broken perceptions. The enemies that hurt us most, we pay the least heed to – as evidenced by our stories and our budgets.
It’s hard to change how minds work.
It’s much easier to change our narratives.
Meditations on Moloch names the perverse incentives that chip away at happiness and freedom and everything we truly value – and gives it a face. What was an abstract conception of broken systems becomes the evil god who will give you a slight momentary advantage if you just sacrifice what you cherish most. By trapping the abstract evil in a (metaphorical) anthropomorphic form, we can level all of our mental enemy-attacking machinery at it, confronting it as the monster it is.
The pattern is known. I’m just making it more explicit, and asking you to apply it to the problems you find in the world. Does your opinion or approach to a problem change when you face it as an enemy? Are you facing the vast and formless villains of the world with the same resolve you reserve for mere mortals?
Would the plan to compromise polio eradication to hunt down Bin Laden have gone forward if our leaders treated faceless enemies with the same weight they afford human enemies?
Think of someone you really, truly despise. Someone who has hurt people, who makes the world a worse place. Now imagine they killed 600,000 people every year, injuring hundreds of millions more – and you knew how to defeat them. Would you ignore it – or would you…
VI. Fight Back
The enemy is out there. The enemy does not know love, or hope, or anything of what it is to be human. The enemy does not mourn its countless victims. The enemy has outlived a thousand empires, and never paused in its ceaseless campaign of suffering and despair. The enemy murders children.
The enemy can be weakened.
The enemy can be killed.
You have this power. You have this choice.
But aren’t the game-theoretic reasons for treating human enemies as much higher priority than natural ones, um, totally valid?
In the ancestral environment? Maybe. When network tv and twitter exist? Almost certainly not.
Surely the game theoretic reason is that we want to ensure that attacking us is never a successful strategy for achieving your goals and that requires that we spend what would otherwise be a disproportionate amount of resources in ensuring attacks on us (by rational agents) fail. I don’t see how network tv and twitter enter into it.
Our heuristics were evolved in an environment where we would never meet more than a few hundred people, any one of whom could have a decently large effect on you. The average rage-target in the age of mass media has a much lower expected impact on your life than your instincts think they do.
I don’t see what twitter has to do with it. If it is sensible to prioritize sentient enemies in the ancestral environment, I honestly fail to see why it should be any less urgent to prioritize them in the modern age when, say, you are in a nuclear arms race with them, or they openly vow to exterminate or subjugate you.
I understand that the US military budget is bit more than $600 billion. But no enemy has staged a military attack on the US for 70 years (AFAIK). It is true that some part of that budget is used for military adventurism overseas, but it seems doubtful that this saves any lives: rather the opposite. So looking at the lives saved (or lost), that $600 billion looks like really terrible value. Now I certainly think that the US overspends on defence, but I don’t think that’s the reason. The army you never use is the best army. Equally, I’m not really sure that counterterrorism expenditure can fairly be assessed by a calculation of the lives saved. Isn’t there also (a) a genuine societal benefit to inflicting condign punishment on wrongdoers and preventing them from achieving their goals and (b) a benefit from a small but widespread increased sense of security (or reduction of terror)? Again, that’s not to say that the US spending on counterterrorism is proportionate. I’m sure it’s not, but I don’t think as straightforward as saying they could save more lives if they spent the money on road safety instead.
Also, now I think about it, I’m pretty sure that the reason the US government was willing to risk increased deaths from polio in order (maybe) to slightly decrease the risk of a smaller number of deaths from terrorism is that the deaths from polio happen in Pakistan while the hypothetical deaths from terrorism might happen in the US. The US government has never even pretended that its policies are intended to promote the well being of all humans equally regardless of their nationality.
Very eloquent, and I loved your smallpox post last year.
Donating…
That’s fantastic! Thank you for joining the fight.
Awesome post! Gets to the core of what I like about a lot of rationalist writing. Making our true enemies clear and real is a powerful skill, and I’m glad some people have it.
Great article. To steelman a bit, wouldn’t you say that security related threats have a significant potential to take a much higher toll than the raw figures might suggest? If say there is public panic and a break down in law and order, or it triggers a large-scale ethnic conflict and/or civil war, or the attacks use some technology that has a much higher destructive power, perhaps then the calculations could look quite different? Although given that at least some of the threat comes from perception (panic), it seems like it’s important to not be too alarmist either. You’re right that the polio thing, without knowing the details, looks pretty awful. I like your writing style and your appeal to good causes, keep it up!
I like the analogy, but the argument on automotive safety misses the mark. The current Dept of Transportation value of life (as of 2015) is $9.1 million dollars. https://www.transportation.gov/regulations/economic-values-used-in-analysis This cost is not included in the DoT budget but is borne by manufacturers and passed on to consumers. It’s why the DoT is forcing car makers to spend hundreds of millions of dollars to reinforce roof supports that will save tens of lives over the life of the cars. In other words, the US government values regulation that will reduce automotive deaths at exactly 9.1 million dollars per life saved.
OK.
Now go do stats on world-wide abortions…