Science and Capitalism

In case you’re confused why there seems to be an ongoing assault on both science and capitalism, let’s clear things up a bit. At its heart, capitalism is just a special case of scientific inquiry, which is itself a tool for finding truth. These are both types of epistemic tools, the things we use to learn new things about the world.


The core of science is a question – how can we prove what we think is true? You start by assuming you’re wrong (null hypothesis), design a test, and see what happens. Make sure you take good notes.

Capitalism is basically the same thing – what do we think people will want, and how can we prove it? You design and try to market a product. The null hypothesis is important here as well – assume you will lose any money you invest.

The idea isn’t complicated. Get idea, test idea, learn things about the world. Simple and elegant. But the world is a messy place, and you have to be careful. First, make sure you know precisely what you are measuring so you can compare your results with others. Second, tailor the experiment to isolate the one variable you care about while holding everything else constant.

For a business experiment, if I invest X dollars in company Y, how much value will be unlocked as a return on investment?

Similarly in thermodynamics: if I apply X heat to Y chemical, how much energy will be released in the form of rapidly expanding gasses?

Anyone who says science is boring has never had a final exam question where you calculate overpressure vs shrapnel effects of a 200 lb JDAM

In our thermo example you can measure the energy inputs and outputs in joules or BTU, then subtract the ambient heat to arrive at a solution. Easy day.

In capitalism you should be able to measure dollar inputs and outputs, then subtract ambient dollars (inflation) to arrive at a solution. This proves to be substantially harder.


The first problem is the unit of measurement. In science, we use a standard measure of energy, maybe BTUs or joules. One BTU is the amount of energy required to raise a pint of water 1° Fahrenheit, while a joule is the amount of energy required to raise 1 kilogram of water 1° Celsius. Different units are annoying, but easy to overcome. We’re all talking about the same thing.

As long as we have a shared definition of reality, we can do this. The sky is blue means el cielo es azul. 2+2 = 11 in base 3, and 4 in base 10. You’re being a little pedantic if you insist on the former at a dinner party, but you’re not wrong.

A shared definition of reality is critical if we want to compare things across time and space. Unfortunately this isn’t the case when it comes to money. A dollar is whatever the Fed and Treasury say it is. Very postmodern.

A dollar today is a very different thing from a dollar in 1971. We typically think of this difference as inflation, and can do a quick translation using a CPI calculator. But this misses a whole lot of weirdness. The price of some goods have deflated considerably thanks to technology, while the cost of others have far outpaced nominal inflation. God help you if you want to compare value across both time and central banks.

Remind me which sectors are the most heavily regulated and subsidized again?

The second problem is ambient energy

In theory, the fiat economy means we could have more or less inflation. Same way you could run your thermo experiment in the desert or in the arctic and need to subtract different levels of ambient energy, for a business experiment you can subtract different levels of inflation to find the real result.

In practice, money printer go brrrrr.

There is a LOT of money printing going on right now, which adds a lot of noise to the data. It gets hard to separate out the signal, the same way it’d be hard to get an accurate reading on our science project if we ran it inside an active volcano.

The real trick is to not try too hard, and instead convince everyone around you that you’re the genius is causing money to explode out of the ground. A lot of people will be incentivized to believe you too. Getting this kind of consensus is certainly a skill and you can do quite well with it, but it’s not one that helps us learn anything new and useful about the world.


To wrap things up, what we want is the truth. One way to find certain kinds of truth is by creating social incentives to experiment with different ideas. Like science, this process is messy and filled with failure, but that’s life.

These kinds of experiments have lifted the world out of abject poverty and created everything we value today. Things have gotten so good in fact that we’ve started taking fundamental underpinnings of the system for granted. There are certainly short term political benefits to playing games with truth and value, but those experiments tend to crash down to earth rather violently. We avoid this by assuming we know very little, and designing experiments to learn just a little more every day.

Incentivizing people to experiment with products to discover truth is capitalism, but this only works when we can effectively measure value across time and space

So how do we fix this? Start using an established benchmark, and move forward.