The Problem with Utilitarianism (Featuring Covid-19 Mass Psychosis)

Eliezer Yudkowsky once proposed an interesting through experiment. Imagine the most minor incidence of pain that a human being can experience, which is accretable over multiple people; that is to say, that it’s measurably worse for two people to experience it than just one person. He suggested a stinging dust mote in the eye. Now imagine the worst sort of pain somebody can experience – say, 50 years of torture. Then a hostile, God-like AI gives you a choice: either one person gets tortured for 50 years, or a googolplex of people suffer a dust mote in the eye. Which do you choose?

“Do the math!” says Eliezer, “The solution’s obvious: torture that one person to avert the eons of misery which the other entails!” And within the confines of the thought experiment he’s correct: if you’re given a choice between two people suffering a broken bone, vs one person, the humane option is to select the latter. If your choice is one billion people receiving a black eye, vs one person breaking their hand, the one billion black eyes are probably worse, and you should select the broken hand. All is well and good so far in this trolley problem, until we try and apply it prescriptively.

The first problem with Yudkowsky’s question is the unexamined premise: the hostile, God-like AI. We can replace the science fiction concept with Gnosticism’s Demiurge: the cruel god of this world, who creates the material illusion in order to mess with us. Whether or not such a being exists is a statement of metaphysical faith: you cannot prove a Gnostic wrong, or prove a Christian right. But the Christian adage “Ye shall know them by their fruits” is a very useful one. It’s another form of Pascal’s wager. Even if the Gnostics are correct, and reality is an illusion designed to make us miserable – and the Christian claims about a loving God, who’s scrutable and consistent, laying out reality in a predictable and comprehensible manner are incorrect – despite all of this, Christianity provides science, architecture, purpose, and music – while Gnosticism provides asceticism and degeneracy, and corrodes the society which adopts it.

The Gnostic position ends up with an infinitude of imaginary threats – take Roko’s Basilisk for example – if reality is ruled by a God with an occult agenda, then anything is possible. You might as well start sacrificing Albinos to the Demon Juju, who knows what the Demiurge wants! If, on the other hand, the Ruler of Reality is fair and consistent, then you can stick to worrying about sitting on a tack or playing in traffic, and forget about being presented with an impossible Sophie’s choice.

The second problem comes when we abandon the Demiurge aspect of this dilemma, and simply attempt to apply its underlying philosophy of Utilitarianism to reality in general. In a crisis we’re often presented with similar problems. During a mass casualty event, who do you help first, Person A or Person B? There’s already a well-developed ethical theory known as Triage. Categorize the patients as follows:

• Those who are likely to live, regardless of what care they receive;
• Those who are unlikely to live, regardless of what care they receive;
• Those for whom immediate care may make a positive difference in outcome.

And then apply care as appropriate to save as many as possible.

Unlike the Utilitarian approach, however, Triage assumes the gut instinct and snap judgments of First Responders, so while it may be Utilitarian in its outcomes (maximizing measurable good, minimizing measurable harm), it’s not Utilitarian in an intrinsic sense.

Utilitarianism as a political philosophy is markedly different from Triage in that it assumes perfect knowledge, and fails to accord for the web of interactivity inherent to reality.

Perfect knowledge: Utilitarianism attempts to measure reality, so that choices can be made quantitatively. Choosing economic policy X will benefit this industry, but harm that industry; policy Y will do the inverse, but with different numbers. Which solution benefits the most number of people? There is a time and a place for such considerations, but always with the caveat that there are an infinite number of facts to consider. Models are useful, but never perfect; they aren’t augurs. They’re a simplified version of reality that leaves most things out, and even the best of them can sometimes be wrong. To revisit the question of dust motes – nobody can guarantee that a googolplex of people will suffer as part of this decision, and to then presume that torturing one person will prevent it is the height of intellectually-driven monstrosity. This was the philosophy adopted by the architects of the Holodomor.

The web of interactivity: social policies are not a science experiment with independent variables. When you yank on one string, the whole web moves. Let’s say you try and make the roads safer by proliferating the number of playground zones: the Utilitarian numbers say that this will prevent a certain number of deaths every year. The unintended consequence, however, is that there are too many zones for the police to enforce, and drivers become cynical about the signs, leading to reckless driving in areas where children are present, and an uptick in deaths.

Utilitarians love their statistics, and pointing out facts which are counter-intuitive. Having a pool in your backyard is more dangerous than owning a firearm. “People are stupid, you should trust the facts!” they say. But this winds up being the premise which eats its own tail. If people are too stupid to accurately assess risk – then they are definitely too stupid to apply statistics correctly! This should be particularly worrisome since our society has transitioned from one ruled by leadership, to one ruled by a managerial elite. As laid out in this twitter thread by Jon Hayward (H/T Patriactionary):

The pandemic highlighted how decades of pushing for socialism have utterly destroyed our ability to measure costs against benefits and evaluate risk. We’re down to people refusing to take vaccines until they’re 100% effective and demands for lockdown until we have 0% coronavirus.

This neurotic hysteria is a result of pushing people to demand 100% safety and security in all things, and convincing them only bigger maternal government can make the safety blanket bigger. The law of diminishing returns means each 1% increase in “security” now costs billions. (Source: https://threader.app/thread/1356221782559256579)

As soon as people start thinking in a Utilitarian manner, they start demanding impossible absolutes. They fail to account for their heuristical biases – obsessively researching every dubious fact they can find about Covid-19 until they see it everywhere. They think linearly, ignoring all of the unintended consequences from the policies that make us ‘safer’. They start to ignore the evidence right before their eyes, in favour of the ‘scientific’ consensus about a topic which, by its very nature, is inscrutable to the scientific method (you cannot run an experiment on social policies happening in real time), and from this we get mass psychosis.

In other words, they drive themselves utterly mad in their attempt to behave rationally.

Don’t underestimate your gut instinct. There are cases where you shouldn’t trust it – it turns out that flying in an airplane is pretty safe – but for the infinite complexity of policy decisions, where an entire web of interconnectivity defies our predictive power, your mind’s innate perception filters will be far more effective in apprehending a useful version of reality than the statistical conjurations of technocrats.

The bad news is that we’re ruled by cowards and idiots who think they can count the number of angles dancing on the head of a pin. The good news is that their own intellectual hubris will be their downfall… eventually.

I have one final thought for you, a free-verse poem that my grandfather liked to say, a stark rejection of Gnostic fear mongering:

There are only two things to worry about:
Either you are sick or you are well.
If you are well there is nothing to worry about.
If you are sick there are two things to worry about:
Either you will get better, or you will die.
If you get better there is nothing to worry about.
If you die there are only two things to worry about:
Either you will go to heaven or go to hell.
If you go to heaven there is nothing to worry about.
If you go to hell, you’ll be too damn busy shaking hands with friends to have time to worry.