Reading Time: 9 minutes

Resources:

Welcome, Wayfinder.

Yesterday we introduced Moloch and multipolar traps.

Today explores illustrations of multipolar traps, and how they manifest through the incentives structures of modern civilization.

Synchronize.

Moloch here is not being used as a symbol of an ancient idol of sacrificial worship.

Moloch refers to the body of the momentum, system, and incentive structures embedded in and expressing through civilization.

Agency is key too—it is not required for Moloch to have a body and influence.

Systems and outcomes can be entirely without agency, and still have impact.

Example → Dictatorless Dystopia

Nick Bostrom pointed towards an illustrative scenario of this kind of non-agentic system.

Imagine a world where every person (including leadership) is part of a dictatorless system where, by tradition, there are two rules that each person reasonably expects enforcement of:

(1) Everybody must give themselves electric shocks for 8 hours per day.

(2) If anyone does not, speaks out against, or fails to enforce either rule, all citizens must unite to kill that person.

Sounds crazy right?

But… you can almost feel into waking up, begrudgingly getting out of bed, and since your survival depended on it, taking your shower and getting your cup of coffee and going to sit in your shock cubicle.

You shock yourself, annoyed, but dutiful.

And you watch, like a hawk, for anyone around you who is so self-righteous and self-important that they might be slacking—while you, the good and upstanding citizen doing your solemn duty for society.

How selfish must someone be to think they’re above shocking themselves while the rest of us endure?

Nobody is going to slack paying their dues on your watch.

Your 8 hours complete (not including your 60 minutes of lunch and bathroom breaks), you commute back home.

On the way, you receive an alert:

“CITIZEN, YOU ARE IN REASONABLE PROXIMITY TO AN UNLAWFUL EVENT. REPORT TO FULFILL YOUR CIVIC DUTY. YOUR PROXIMITY HAS BEEN LOGGED. FAILURE TO REPORT FOR DUTY WILL BE LOGGED AS UNLAWFUL BEHAVIOR.

You report in. You fulfill your duty. Death by hanging. Quick. Uneventful.

You hate it. Everyone hates it. But what are we to do?

This is just “the way things are” after all.

Clearly contrived, but not so hard to imagine, is it?

We’re all familiar with doing things we dislike because we have to.

We all have heard the line “that’s just the way things are” before.

The multipolar trap here is that:

(1) It’s in every agent’s self-interest to follow the laws (i.e. not die), shocking themselves and participating in the collective killing of law breakers.

(2) Although everyone would be better off if they cooperated to refuse to shock themselves and refuse to kill anyone that broke the law, it is of great personal risk to attempt to subvert the system. An individual would need sufficient trust that enough others would coordinate that a sufficient majority of defectors could defend themselves against the compliers.

At the root, it’s a coordination problem.

Let’s explore a few examples of multipolar traps observed in the modern world.

The Prisoner’s Dilemma (Cooperate or Defect?)

Generalized, this is any situation where there are at least two parties, and each party has two choices: (1) cooperate and (2) defect. The incentive structure is such that if both parties cooperate, they will collectively be better off. Both are incentivized to defect, however, and if only one defects, the cooperator will be significantly worse off. If both defect, they will be collectively worse off, but less so. Both parties are aware of situation, but unable to coordinate.

The classic The Prisoner’s Dilemma describes a situation where they are two people arrested for a crime. They are independently interrogated, and unable to communicate or coordinate with one another.

The prosecutor makes the following offer to both people:

  • If you both confess (defect), you’ll both be convicted but with a reduced sentence.
  • If you confess (defect) and they don’t, you’ll go free and they’ll get a full sentence. The reverse holds too.
  • If neither of you confess and remain silent (cooperate) then you’ll go free or receive some trivial punishment.

Although choosing to cooperate produces the best possible outcome, doing so would require trust that the other party would choose not to act in their self-interest (defect).

This seed of doubt is likely to poison the chances of dual party cooperation.

Imagine one of the arrested has young children—they sit in their cell unable to sleep, for days on end.

All they can think about is their kids, them growing up with a parent in jail.

They say to themselves “well, if I confess, then at worst I’ll have a reduced sentence. I might even go free. I can’t afford to risk staying silent and them confessing, there are people counting on me. They don’t even have kids. They can handle a full sentence and me going free. That’s better for the kids anyway, I’m sure they’ll understand some day.”

In the other cell, the thoughts may go “well, I didn’t even want to be a part of this. They talked me into it. I bet they’re thinking about their kids, and will probably confess to save their skin. Fuck that—I didn’t even want to do this, it was a mistake and I shouldn’t pay full price for it while they go free. I don’t have kids, but I want to start a family and I won’t be able to if I’m in here for a full sentence. Especially since they’ll probably confess… I’ll confess too.”

You can see how things are likely to go.

If there’s a single, significant seed of doubt in the trustworthiness of the other party to cooperate, the likely outcome will be defection by at least one party.

It’s possible that altruism will be ingrained deeply enough into one or both parties for cooperate to be chosen by one, or maybe even both.

But not likely.

Most importantly, it’s not incentivized.

And that’s just two parties.

They may even be in deep, trusting relationship and choose to defect.

Things get infinitely more complicated when there are multiple players, with no personal connection, and non-agentic systems in play responding to incentives that invalidate choice grounded by human ideals of altruism.

Capitalism (Multiplayer Prisoner’s Dilemma)

Let’s take a look at what happens in a free market economy with substantial competition.

We can think about each company at play in the industry as an individual prisoner (from the above example).

Cooperating here would be to collectively hold prices steady or increase them.

Defecting would be lowering their prices to undercut the competition.

Market dynamics are such that the industry is sufficiently competitive with enough options and minimal differentiation between suppliers that buyers will choose the cheapest price above all else.

So, our dutiful capitalist cuts costs. They cut corners. They pay workers the minimal amount required by law, with the minimal conditions required to keep producing. This isn’t even necessarily their choice, this may be the demands of being a player in that industry since the biggest players have all competed down to these price and production standards.

Maybe the capitalist would even like to pay the workers more, maybe they would like to improve the working conditions.

But they can’t—because increasing prices would result in losing business to more competitive rivals, and it’s so tight already that it would result in the businesses going bankrupt and shutting down.

“It may be suboptimal, but at least its something. At least it’s enough for the workers to survive on, and for me to survive. And everyone else is doing the same. If I don’t do this, someone else will, so why not me?”

The best case would be for all competing companies to collectively raise prices (cooperate), which would not only increase their profitability but also allow for improving worker conditions.

Unfortunately, if a single company lower prices to undercut the competition (defect), then all others will be forced to either defect too, or go out of business.

Remember, “one bad apple ruins the bunch”.

And now coordination is even harder than in the prisoner’s dilemma—these are non-human entities with no emotional ties to one another, and now the risk calculation gets even darker for any individual player.

As the number of players increases, the chances of all parties cooperating gets lower.

Don’t forget, in the United States CEOs are legally required to prove they acted in the best interest of the company, which certainly would not be altruistically going out of business on principal.

They board would find someone else who would be willing to lower prices and keep the business from bankruptcy.

So now the decision maker has a personal incentive to defect in addition to the business incentive to defect.

As illustrated, corruption is not required.

Evil is not required.

Incentive structure is all that is required to produce what appears evil and corrupt.

(Not to say evil and corruption don’t emerge or aren’t embodied in these systems, simply that they aren’t fundamentally required for the systems to produce observed outcomes.)

Moloch need not be evil to appear evil.

Moloch can simply be a mindless, agency-less outcome of agents responding reasonably to incentive structure.

And truly, how different is the sweatshop smartphone buyer from the sweatshop smartphone producer?

Through this lens, there are more similarities (of responding to self-interest incentive structures) between us and the “bad guys” than we’d like to admit.

Tomorrow, we’ll explore a few more manifestations of multipolar traps, because seeing how far this web extends and how many threads are connected to the multipolar source is crucial.

More than the ability to see the world around us, we need to develop vision for its mechanics, myths, and levers.

Godspeed, Wayfinder.

~ Michael