We’re Still in the Industrial Revolution

by Jasper Gilley

A little under 200 years ago, one of the most important events in human history began:

This graph is really boring, until about 200 years ago, when human population suddenly goes exponential. Clearly, something interesting is going on.

The Industrial Revolution

Historians generally think of the Industrial Revolution as a thing that happened in the late 1700s and early 1800s. If you ask a historian to tell you interesting things about the Industrial Revolution, they’ll probably mention things like the invention and subsequent adoption of power looms, steam engines, gas lighting, chemistry, and metallurgy. Those were all interesting things, but saying that they defined the Industrial Revolution misses the forest for the trees. One might be suspicious that something bigger is going on by looking at the world’s historical GDP per capita:

For pretty much all of human history, the average human’s wealth is stagnant or declining (due to population growth.) But suddenly, around 1850, it goes exponential, directly correlating with population growth. (If you’re unfamiliar with logarithmic graphs, a straight line on a logarithmic graph denotes exponential growth.)

The important thing about the Industrial Revolution is not so much that new technologies like power looms developed, but that those new technologies were used in entirely new ways. The power loom, for instance, in automating a task that previously required a semi-skilled laborer, allowed for the development of an entirely new form of manufacturing in factories and cotton mills; the key thing about the Industrial Revolution was not that humans gained new capabilities so much as it was that they gained a significantly cheaper and quicker and better way of doing things.

Yet the same could be said about any time in the past 200 years. The rise of railroads in the mid-1800s made long-distance transportation cheaper and quicker and better. The rise of automobiles 100 years ago made short-distance transportation cheaper and quicker and better. The rise of computers and the internet is currently making communication cheaper and quicker and better. The exponential gains in world GDP per capita seen 200 years ago have shown no signs of abating since. We’re still in the Industrial Revolution.

Exponential Economics

In the post-Industrial era, a number of distinctive economic trends have recurred. Perhaps the most significant of those trends is that of automation: the substitution of capital for labor.

Consider the pre- and post-Industrial methods of farming:

 

If you’re a pre-Industrial farmer, you’re not happy because you have to do a lot of manual work, in addition to owning and feeding an ox. In short, you do a lot of laboring. The only upside to your situation is that you didn’t have to raise much capital to get started in the farming business: since everyone needs them, oxen are relatively inexpensive, and can be fed using the food you grow. If you’re a post-Industrial farmer, though, your work consists solely of driving the tractor and having someone to repair it when necessary. You probably had to take out a sizable loan to buy the tractor and other equipment, however; you needed to raise a substantial amount of capital.

Prior to the Industrial Revolution, it would have been impossible for any non-state actor to raise the amounts of capital required to farm on an industrial scale, even if it were possible to buy a John Deere tractor in 1700. This points to a significant macroeconomic shift that occurred in parallel with the process of industrialization. Growth early in the Industrial Revolution created investment opportunities, which began a positive feedback loop that gave rise to the modern capitalist economic system:

During the Industrial Revolution, the economy began growing appreciably again for the first time in thousands of years, so it was newly possible to make money by not “doing anything,” but simply by funding others. Thus, the modern investor was born. The availability of investors’ capital enabled more capital-intensive projects to be undertaken, further driving growth. For the first time in human history, the factor limiting economic growth was the slowness of scientific progress, not the unavailability of capital.

Micro-Industrialization

One way to view the ongoing Industrial Revolution is as a series of micro-industrializations, the first being in textiles and the most recent being in information technology. Each micro-industrialization follows roughly the same synopsis:

• Machines begin doing a task that until recently was done by humans, undercutting humans on price
• Human jobs are lost, Luddites protest
• In the long run, automation creates more jobs for humans in adjacent sectors

Take the micro-industrialization of ATMs in the 1970s. When ATMs began to be installed in banks, biological tellers became obsolete. Job losses led to widespread protests from Luddites. Yet in the long run, ATMs significantly lowered the marginal cost of opening a new bank, allowing more branches to be opened, which increased the total number of humans who worked in bank branches.

Or textiles. The power loom made cotton much more cheaply and quickly than any human could. As a result, human weavers were no longer competitive, and had to vacate the industry. But the power loom led to cheaper and better clothes for all, which increased the purchasing power of the poor in particular. As a result, everyone had more money to spend, and the economy grew, more than absorbing the job losses from textiles.

Our Industrial Future

GDP per capita is still rising exponentially. Therefore, we have no reason to suspect that the Industrial era is finished, or that society will not continue to see micro-industrializations that adhere to the above synopsis.

Likely one of the next major micro-industrializations will be in the operation of motor vehicles. About 3.5 million Americans currently drive a vehicle for a living¹, or about 1.2% of the total population. There are a number of very serious, well-funded startups and companies that will almost inevitably put 100% of these laborers out of work. The adoption of autonomous vehicles will likely be very swift (once regulatory approval is given), and there will almost certainly be the strongest Luddite backlash since the original Luddites who protested against the power loom (given the current political climate, this seems especially likely.)

Yet consumers will very quickly notice a funny thing happen. Since trucking is the dominant method of transporting goods in the United States, the price of almost every consumer product will go down by at least 15%. The economy will rally along with consumer spending, and the unemployed truckers will find something to do. The micro-industrial cycle will be complete, and the next micro-industrialization will arrive.

AI

Another pending related micro-industrialization will be that of artificial intelligence (AI.) This is a particularly difficult one to think about for several reasons:

• We’re currently (early 2017) at the height of a speculative bubble surrounding AI.
• Many quasi-legitimate thinkers (e.g., Ray Kurzweil) are really onboard with the speculative bubble, possibly for dubious reasons (e.g., it makes them money.)
• AI has been the subject of many apocalyptic science fiction movies and books (most of which are as unrealistic as all science fiction.)
• The term artificial intelligence is extremely misleading. The things AI algorithms can do may seem “intelligent” to us today, but so did power looms to Luddites. In 50 years, AI will seem as obviously non-intelligent as standard computers do today: cool, but nothing to lose your mind over.

That being said, the implementation of AI (the term is misleading but it’s the most commonly used, so let’s go with it) will be a very serious micro-industrialization, though it will still fall inside the parameters of those previous.

Before we analyze the synopsis of AI’s micro-industrialization, it will help to provide a clear definition of what an AI algorithm actually does. Since Wikipedia sets the world record for the worst possible definition (“Artificial Intelligence is intelligence exhibited by machines”), it won’t be difficult to do better. Broadly speaking, AI algorithms analyze patterns in data, and use those patterns to make predictions about further data. For instance, one type of image-recognition algorithm takes in photos of an object as input, and as output, can recognize (imperfectly) that object in new images it is fed. Thus, AI algorithms act without being explicitly programmed to do so, hence the misconception that they are somehow “intelligent.”

The problem is that there are a lot of laborers whose job consists of analyzing patterns in data and using those patterns to make predictions about further data. (I use the term labor to denote a job that is automatable at a given point in time, and human capital to denote a job that is not automatable at that particular time.) In 10 years, there will be a large number of newly unemployed individuals who until recently held a stable, white-collar job. Due to the upmarket nature of the jobs to be displaced, I’d expect less of a Luddite backlash against AI in general than against the specific application of AI to automate the operation of vehicles. Still, if Ray Kurzweil keeps up his rhetoric, anything is possible.

Yet just as with every other micro-industrialization, the end-result of AI will be greater productivity across the global economy (as long as its gains are widely distributed, not concentrated in the hands of a particular set of corporations.) Data will be analyzed faster and better, effectively putting every human out of work who finds patterns for a living, but this automation will allow humans to hold more meaningful occupations than pattern-finding, growing the economy in the long run.

Beyond Industrialization

As alluded to above, the Industrial Revolution isn’t the first time in human history with exponential gains. The dawn of modern agriculture and urbanization in 10,000 BC produced an explosion in Earth’s human population that contemporary humans would no doubt have considered as important as we consider the Industrial Revolution to be. If the Industrial Revolution has a predecessor, we can also expect it to have a successor.

What might that successor be? It is likely as difficult for us to imagine the Industrial Revolution’s successor as it would have been for ancient humans to imagine the Industrial Revolution itself. But a key constant in all forms of exponential growth is a long period of stagnation before takeoff. Workers’ productivity was stagnant for thousands of years prior to Industrialization, at which point it rapidly increased. The next round of exponentialism will revolutionize an area that is stagnant now.

Though, of course, thinking about such deep-future events, while fun, is hardly relevant to us who are in the heart of the Industrial Revolution. We have our own exponentialism to discover.

---

If you enjoyed this post, consider liking Crux Capacitor on Facebook, or subscribing to get new posts delivered to your email address.

---

¹ The American Trucking Associations

Featured image originally appeared in The Economist