Jeff Bezos wants to own all the robots

If he succeeds, we’re finished

by James Kaelan

A woodcut for the 1868 novella “The Steam Man of the Prairies”

In the salad days of infotech — the late 1980s/early 1990s, as the PC was becoming ubiquitous, but the internet was still something you accessed through a landline (if you logged on at all) — software wasn’t separable from its physical container. If you wanted a new program, you went to a store and bought a cardboard box that contained an actual disk. This disk was replete with information that was cheaply copyable. But the vector for the code was still tangible: a square of plastic and metal that you inserted into your beige, Intel-powered tower.

Once the internet got fast enough, though, and the quantity of data that could be stored on a hard drive increased exponentially, our relationship to software — to digital data of all sorts — began to change. In the late ‘90s, file-sharing started taking off. Suddenly you could go to Napster and download, for free, an album that cost $17 at Borders (R.I.P.). It felt sort of like stealing, but it definitely didn’t feel the same as walking into a store and shoplifting.

Why was that? Why did ripping an album off Napster feel way different than pocketing a CD at The Warehouse (R.I.P.)?

In his indispensable book, Postcapitalism, Paul Mason argues that this disjuncture — feeling differently about physical and digital commodities, even if they both contain the same content — is rooted in an old concept that fell out of fashion in the mid 1800s: Adam Smith’s Labor Theory of Value (though AOC used LTV to troll Ted Cruz last year 😍😍😍).

Back when Smith was writing, to make a plow, humans had to mine the ore, refine it, and transport it — and then a blacksmith had to pound the raw slug on an anvil for hours. A plow’s monetary value wasn’t based on the quantity of plows on the market, or even the new work the plow allowed the farmer to accomplish. A plow was valuable because of all the human labor that had gone into making it. Cheap things were easy to make. Expensive things took time.

But Smith’s Labor Theory of Value gets usurped in the 19th Century by a new “fundamental” law of economics: supply and demand. With supply and demand, the price of a plow doesn’t depend on the labor that went into making it. The price of a plow depends on how many plows exist in the marketplace, and how many people happen to need plows at any given moment.

Supply and demand, like any great theory, is elegantly simple. And for about 200 years, it more or less accurately described the price of commodities on the increasingly global market. Whether you’re talking about raw materials or finished goods, when demand goes up, the price goes up. When the supply goes up faster than the demand, the price goes down.

But fast forward to the 1990s — the infotech era — and suddenly, Paul Mason demonstrates in Postcapitalism, supply and demand starts looking like a weird metric for establishing prices.

The price of a grain of rice depends on how much rice there is and how many people want to eat it. But what about the price of a byte of data? How do you determine what information is worth?

One way to establish the price for data on the market is to determine how much it costs to store and replicate each [peta]byte. Factored into that calculation are the servers and personal hard drives and fiber optic cables and satellites that store and transmit the data. And there’s the energy required to produce, transport, install, and power that storage and transmission infrastructure. Those are the most tangible factors in the price of information — recognizable to anyone in the business of producing and transporting rice.

But that’s where data stops acting like rice. Because unlike rice, once you’ve built the infrastructure for transmitting and storing data, you can replicate that data infinitely at a cost approaching zero.

Imagine if that was true with rice? Imagine if you could just copy and paste a grain of rice, instantly, more or less for free, and send it, more or less for free, anywhere on earth? That sudden infinite abundance, per supply and demand, would destroy rice’s market value.

Well, that’s where we are today with data. Data will never quite be free to replicate, transport, and store. But the cost of data, unlike the cost of rice, continues to fall exponentially.

So how is it that companies make money from a nearly-free-to-replicate commodity like data?

There are two main ways.

The first is to make data act like a scarce commodity.

Before digital security barriers and regulations were as robust as they are now, we lived in a golden age of piracy. You could Torrent everything from Hollywood films to pay-per-view sports to the latest version of Photoshop with near total impunity. For years, Torrent host sites, bouncing around from one permissive country domain to another avoided lawsuits and regulation. And consumers of Torrents ate free at the buffet.

At its peak, Torrent traffic accounted for 25% of overnight internet activity.

To get people to pay for a thing that’s free to reproduce, therefore, the “owners” of the data must lock their “property” inside security systems so the data can’t be freely copied — or at least not easily. If someone has exclusive control of a certain type of data, and you want that data, you must now pay to access it — not because the data is expensive to replicate, but because one entity controls the drawbridge.

The second factor influencing the price of data is the strategic advantage that data can give an enterprise competing with other enterprises in a given market.

Let’s use Facebook as an example. In order to profit off data, Facebook must “own” proprietary data that no one else has free access to. And “own” proprietary data, they do. As-we-know-all-too-well-but-somehow-STILL-USE–THE–PLATFORM–MOM, Facebook watches all its users, vacuums up huge quantities of behavioral information, and then extracts valuable psychographic profiles from that observation.

Why are those behavioral analyses valuable?

Because they allow third-party enterprises to sell shit to Facebook users.

If Facebook knows your exercise schedule — because your FitBit (owned by Google lol) posts to your profile — Facebook can charge Nike to serve you an ad for running shoes when you’re literally the most physiologically susceptible to their messaging: right after you’ve finished a run.

So, to reiterate, companies like Facebook (and Google and Microsoft and Amazon) can make data valuable by

  • Proving that “their” data gives an enterprise — or country — a competitive advantage in whatever market they operate in
  • Putting up a wall limiting access to “their” data

But without those walls, without the legal right to collect our information — as we saw during the Bittorrent golden age — data trends toward free.

And based on the idea that the price of nearly-free-to-reproduce commodities trends toward zero, Mason, writing in Postcapitalism, makes a radical claim: As Free Machines — fully-automated, machine-repaired machines (something Karl Marx hypothesized in his “Fragment on Machines”) — proliferate, as human labor is removed from the production process, physical products could start to act like digital ones 🤯🤯🤯.

The last company on Earth

The Spheres at Amazon Headquarters in Seattle, WA

The centerpiece of Marxism is the idea that collectivized labor — a union at one end of the spectrum (think the United Automotive Workers), state ownership of the means of production at the other (think Soviet Communism) — counterbalances the influence of capital. If workers unite and refuse to produce anything until their demands are met (or just overthrow the government), the proletariat will topple the bourgeoisie .

Sounds pretty good in theory.

Given enough time — and a moratorium on new machines — labor might eventually triumph over capital. But as we determined in our definition of capitalism, in order for capitalist enterprises to grow (and grow they must), they seek new technological efficiencies that reduce their labor and material costs. The natural effect? The more automated the production process, the more power the proletariat loses.

Why? Per supply and demand, which holds up so long as jobs and workers are finite “resources,” the fewer jobs there are (because robots eliminated them), the more precious those jobs are to workers, and the less likely workers are to unionize.

(To see all these factors roiling in the iron cauldron of modern manufacturing — proletariat, bourgeoisie, American capitalism, Chinese communism, union organizing, union suppression, and the very visible hand of automation —watch American Factory.)

Since the dawn of modern capitalism in 18th Century England, workers have correlated technology with their eventual obliteration (for good reason). The term Luddite, which we now use to mean a person skeptical of technology, comes from a secret organization of laborers in Nottingham who destroyed the textile machinery that was depressing their wages. What would they have thought about Amazon’s Kiva robots???

Almost 300 years later, the pace of the development of new efficiency technologies has accelerated exponentially. And workers have lost influence proportionally. Marx was pretty certain that, by the end of the 19th Century at the latest, the world’s laborers would overthrow their capitalist overlords. During the 20th Century, revolutionaries succeeded in Russia and China, Vietnam and Cuba (and might have in Guatemala, Haiti, Brazil, Uruguay, Bolivia, Chile, Argentina, El Salvador, Panama, and Peru — to name just 10 countries where the CIA suppressed socialism).

But capitalism, naturally-aspirated or force-fed, has prevailed. China, the world’s only “communist” superpower, has nearly as many billionaires as the United States.

(It might be worth noting, here, that I don’t identify as a Marxist — or even as a socialist. I caucus with the Democratic Socialists of America, and I support many of the policies and candidates they endorse. But Marxism is a reaction to capitalism, and therefore exists within a capitalist, industrialist paradigm. State control of production doesn’t mean a healthier planet. In 2017 China emitted twice as much carbon as the United States. But more on why I’m not a Marxist in coming weeks!)

As we enter the era of automation — when an ever-vanishing minority of actual people are making the things the ever-expanding majority of people need to survive—whomever owns the automated means of production wins. Forever. Or as long as they (he) can maintain ownership.

But there’s a fascinating, destabilizing phenomenon that fully-automated production might unleash. And since I’ve just used enough Marxist jargon to make a college junior in a Che Guevara t-shirt sleepy, I’ll pass the mic over to Memoji Jeff Bezos and let him explain!

In the first section of the video, Bezos describes the transition from mercantilism/feudalism to capitalism. In the second he explains the effect that automation could have on the price of Free Machine-made physical goods.

But fast forward roughly 500 years to 2019

And now you’ve got an abstract concept called the economy

Which is a term that attempts to describe the global exchange

Of all goods and services produced by both man and machine.

And that works pretty good for us as long as the metrics don’t change

As long a service done by a computer accessed through a screen

That was built by a robot in an automated factory

Is valued the same as a Washoe basket handwoven from reeds.

I bring this up because the concept of supply and demand

Which underpins our economic thinking — like, the Invisible Hand —

Is predicated on the idea of scarce natural resources:

Limited quantities of raw materials, labor, capital, and land.

But shit gets kind of fucked when the economy expands

Into infotech, where you step into ideological quicksand

Because the commodity — data — isn’t governed by traditional forces:

It’s free to reproduce, something Adam Smith wouldn’t understand.

But he would understand something called the Labor Theory of Value

Which says machine-made products are valuable when they’re brand new

Because the machine contains human labor, but doesn’t take human labor to run.

But if the machine operates for more hours than it took to construct

Then once the industry adopts the machine and it becomes ubiquitous

Physical products start to act like things made of zeroes and ones:

Prices crash, and the margin gets too slim to pursue

And after a decade of full automation, the free market is done.

To revisit our Labor Theory of Value explainer, Adam Smith posited that plows aren’t valuable because of how many there are on the market or how many farmers need to till their fields. Plows are valuable because of all the human effort that goes into mining, refining, shipping, and forging the iron into usable tools.

If plows grew on trees — unless someone figured out how to own all the trees, turning the plow-abundance into an artificial plow-scarcity — plows would be free.

So what happens if machines are doing the mining, refining, shipping, forging, and plowing? Over time, the price of the iron (or high tensile steel), the plow (or autonomous tractor), and the (machine-grown-and-harvested) food, like hyper-abundant digital information, should trend toward zero.

That is, unless like Facebook and Google and Amazon do today with data, someone manages to keep a wall around those abundant, machine-made goods and forces them to act like scarce commodities governed by the law of supply and demand.

As Memoji Jeff Bezos says to mollify his panicked board members:

The only way to make information scarce is to have a monopoly. We’re going to be the last company — the only company on Earth. We’ll be fine.

Facebook’s monopoly on our data allowed the Trump campaign, in conjunction with Cambridge Analytica and Russia’s Internet Research Agency, to undermine the integrity of the 2016 American Presidential election. If Amazon has a monopoly — or near monopoly — on the automated means of production, Jeff Bezos et al can set the price of the stuff machines make.

Now imagine if Amazon, using data gathered from our purchase histories — and all the things Alexa has heard us say—could manufacture, market, sell, and ship completely machine-made-and-moved products. We’d be as susceptible to Bezos’s manipulation as we were to the Russian trolls. And he’d be even richer. (That’s why, Spencer, I want to preëmptively seize Amazon 🙃)

In the age of the COVID-19, a corporation capable of delivering food and durable goods to your door has a tremendous opportunity to further its reach and solidify market control. Amazon has already hired 100,000 new employees to keep up with accelerating demand.

The upside is you get the things you need quickly and safely (so long as you aren’t the one working in the fulfillment center or delivering the dildos). The downside is the control you cede now to Amazon doesn’t just revert back to you when the pandemic ends.

Jeff keeps it.

Eject from Capitalism >> Survive the Landing 🚀 Subscribe to the newsletter:

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store