The Flywheel

Chapter 6 of The Techno-Humanist Manifesto

Previously: Solutionism, part 1, part 2 and part 3. Part of The Techno-Humanist Manifesto.

---

If the dramatic progress of the last few centuries is the great boon of history, then the great tragedy of history is in all the centuries prior, when that progress didn’t happen. For tens of thousands of years, people toiled, starved, suffered, and died until we finally achieved modern economic growth.

Why? Why was progress so slow, for so long? Did it have to be? What caused it to finally accelerate in the modern era? And were the last few centuries a fluke, a lucky windfall of progress, after which we should expect a regression to the mean of slow growth? Or were they part of a trend that we can expect to continue?

Seeking the origin of progress, one is naturally drawn to the Industrial Revolution. This brief span of two generations ushered in a new age of humanity and began to shift the basis of the economy from agriculture and crafts to energy and machines—the most fundamental economic transformation since the advent of agriculture and settled societies some ten thousand years prior.

A singular focus on the Industrial Revolution seems to be bolstered by looking at world GDP per capita, which stays relatively flat for thousands of years, only increasing noticeably in the early 1800s, as industrialization began to be broadly felt:¹

Looking at this and several related charts with the same pattern, one astute essayist remarked, “my one-sentence summary of recorded human history is this: Everything was awful for a very long time, and then the industrial revolution happened.”²

But the seeker who looks rigorously is unable to find the origin of progress in the Industrial Revolution. Agricultural improvements in England began at least a century earlier; the higher productivity that resulted freed up labor for industry.³ That labor was pulled away from the farms by the growth of London, which was noticeable starting in the mid-1500s.⁴ London’s growth was driven by foreign trade, which was made possible by improvements in ships and navigation made in prior centuries. Fertile ground for the Industrial Revolution was laid by the new scientific culture that was well under development in Europe by 1600,⁵ and which itself was powered by the Gutenberg movable-type printing press invented in the 1400s. The Industrial Revolution was not the beginning of progress, just the beginning of a new and more intense phase of progress.

We find progress as far back as we look, even well before our own species, Homo sapiens. Technological progress is evident in the archaeological record of stone tools: the earliest tools, from over 3 million years ago,⁶ are little more than broken rocks with sharp edges; middle-era tools from hundreds of thousands of years ago are more deliberately formed, with a cutting edge all the way around; by the end of the stone age they are made with precision and in a wide variety of specialized forms.⁷ We can even measure this progress by the length of cutting edge created per weight of material, which increases from 10 mm/g with some of the oldest techniques, to 25 mm/g with some of the most advanced.⁸

Left: chopping tool found in the Duero Valley in Spain, from the Oldowan Industry, 2.9–1.7 million years ago. Center: “handaxe” found in the same valley, from the Acheulean Industry, 1.95 million–130 thousand years ago. Right, top: projectile point found in Eastgate, Nevada, less than 5,000 years old. Right, bottom: awl or piercer, found in Cornwall, late Neolithic, dated to 4500–2100 BC. Adapted from: José-Manuel Benito Álvarez and Portable Antiquities Scheme via Wikimedia, CC BY-SA; Heizer and Baumhoff, “Archaeology of Two Sites at Eastgate”

Progress continued into the agricultural era, and even picked up the pace—which is to say, the pace accelerated from glacial to molasses. Yet some of the most important inventions of history were created in this era: crafts such as metalworking, pottery, and glass blowing; textile equipment such as the spinning wheel and the loom; agricultural implements such as the plow and the harrow; watermills and windmills; gears, levers, and pulleys; wagons and ships; writing and printing; swords and cannons; law and corporations; credit and loans. The world of 1700 was vastly advanced beyond the world of 1700 BC, let alone the world of 17,000 BC.

All of this is obscured by looking at world GDP per capita, for two reasons. First, modern economic growth began in the West, which “broke out” and raced ahead while much of the world stayed behind until recent decades. Looking at Western countries, even GDP per capita shows signs of gradual acceleration before and through the industrial era.⁹ Second, until the modern era the world was stuck in a “Malthusian trap.” Wealth grew so slowly that population was always nipping at its heels—population growth, in fact, was limited by how much food we could produce. When productivity increased, more children could be born, and more survive, but this excess population quickly grew to consume all the excess wealth. Because agricultural productivity was limited by land, and because technological advancement was so slow, any additional labor supplied by the growing population faced severely diminishing returns beyond the carrying capacity of the region. As population thus kept pace with economic growth, GDP per capita grew very slowly, if at all, for long periods of time.¹⁰

But the growth in knowledge, technology, wealth, infrastructure, and civilizational capabilities was real—even if at first, it mostly just enabled a larger and more powerful society, and transformed the lives of a small elite, rather than creating broad-based prosperity. And those capabilities were a necessary foundation for the economic system that finally would create wealth faster than people and improve the lives of nearly everyone.

Thus, we can see the long-term pattern of progress more clearly if we look at total GDP rather than per-capita GDP. And when we do that, we find gradual acceleration over all of human and even pre-human history.

It’s crucial to understand the exact shape of this acceleration, so let’s examine it carefully.

Consider the GDP of the UK over a span of 200 years. We see steady growth—depressions, world wars, and even the 1918 influenza pandemic are barely noticeable as small dents in a fairly smooth, upward-bending curve:¹¹

The curve is exponential. We can tell because when we plot it on a logarithmic y-axis, it becomes a straight line:¹²

That straight line indicates a constant growth rate, the hallmark of an exponential curve. The steeper the line, the faster the growth; in this case, about 4%/year.

Here is world GDP over the last 2,000 years, also with a logarithmic y-axis:¹³

Here, even on a logarithmic chart, the line is not straight: it bends upward. This is not exponential growth, but super-exponential; not a constant growth rate, but growth even in the rate of growth itself.

This super-exponential pattern is found as far back as anyone has looked, back to the early stone age.¹⁴ Below are estimates of growth rates plotted against the size of the economy, each on a logarithmic axis.¹⁵ Each data point is labeled with the year it is from. I have annotated this chart with groupings for each major era of economic history. Roughly, annual growth rates were less than a hundredth of a percent in the stone age, a fraction of a percent in the agricultural age, and single-digit percentage points in the industrial age:

Chart taken from David Roodman, “Revisiting Hanson’s 'Long-Term Growth as a Sequence of Exponential Modes'”; annotated with ellipses to show groupings.

We might question the data, as most of these figures are reconstructions based on scant evidence.¹⁶ But Nobel laureate economist Paul Romer proposes a simple thought experiment:

Suppose the modern rate of growth of real GDP per capita (that is the growth rate after taking out the effects of inflation) is equal to 2% per year and that income per capita in year 2000 is $40,000. If this rate had prevailed for the last 1000 years, then in the year 1000, income per capita measured in the purchasing power of dollars today would have been $0.0001, or 0.01 cents. This is way too small to sustain life. …

Reasonable people can differ about what the future holds, but the simple calculation that first got me thinking about this … leaves no room for doubt about what happened in the past. The rate of growth of GDP per capita has increased over time. The rate of progress in standards of living has increased even more.

Because the logic is so clear, there has never been any serious debate about the historical fact that is the basis for the question about speeding up.¹⁷

Romer and Chad Jones, in a 2010 paper, cited “accelerating growth” as one of six major facts that should motivate the research agenda of economics in the coming decades.¹⁸

The constant growth rate of an exponential curve is generated by a system where the mechanism of growth is invariant with scale. A classic example is a biological population (in the absence of resource constraints): the mechanism of growth is the reproductive faculty of the individual organism. Each female gives birth to a certain number of offspring, which does not depend on the size of the population. Similarly, a savings account grows exponentially if each dollar invested earns a constant rate of return, regardless of the size of the account.

This is also a good approximation of an economy over the short run. But over the long run, the mechanisms of economic growth are themselves improved by growth.¹⁹ Progress compounds.

Consider some key drivers of growth, and think about how they have been strengthened over time:

• Total investment in research, development, and new ventures, both financial capital and human capital

• Science and technology, the base of knowledge and processes we have to work from

• Industrial infrastructure, in energy, manufacturing, and transportation, the base from which we turn ideas into products and services

• Communication and computation, our ability to spread, process, and recombine ideas and information

• Market size, the potential demand for any new product or service

• Institutions that make progress possible, from government to corporations to universities to banks

These factors are self-reinforcing, in a system of multiple, overlapping feedback loops. The growth of the world economy creates more surplus wealth to plow into R&D, which creates better technology, which raises productivity, which creates a bigger economy and more surplus. Better science helps us create new technology, which helps create better scientific instruments, which help advance science. Better communications and transportation infrastructure expands the scope of markets; larger markets support greater investment in all kinds of infrastructure, including better communications and transportation.

Our institutions are part of the feedback loop, too: the development of corporations in the 1600s, research universities in the 1800s, and venture capital in the 1900s were all in response to needs generated by the advance of science, technology, and the economy; and these developments in turn helped accelerate that advancement. Even our ideas about progress itself are motivated by prior progress and help steer and drive future progress. When Francis Bacon exhorted his contemporaries in 1610 to believe in the possibility of scientific progress and to devote their efforts to it, he pointed to past discoveries and inventions as evidence, such as the compass, silk, gunpowder, and the printing press;²⁰ his philosophy motivated scientists for generations to come.²¹ When Vannevar Bush called for investment in basic research, it was justified by the idea that science was the foundation of national prosperity and security.²²

The nature of such self-reinforcing feedback loops is that they are very hard to get going, because you start from a low level of capability—but as they speed up, they gain almost unstoppable momentum. Envision a massive flywheel, gathering energy from the turns of a small engine. Due to the inertia of the wheel, each turn of the engine makes only a small impact on its speed; at first not much seems to be happening at all. But the compounding investment of energy into the wheel eventually accelerates it to enormous velocity.

So, why was progress so slow for so long, only recently reaching modern levels of growth? Because that is the nature of accelerating growth: it is relatively slow for a long time before reaching any given level; no matter where you are in the process, it will be the case that most of the progress was made recently. As early as 1780, Benjamin Franklin was feeling the acceleration when he remarked on “the rapid progress true science now makes,” yet even he would be astounded to see what we have done since with his electricity.²³

To make this mechanism vivid, let’s look at a case study of one invention and its struggle to be born: the threshing machine.

Threshing is a crucial step in the processing of wheat after it is harvested: breaking open the hard outer shell of the wheat kernel in order to get at the grain inside. For thousands of years, this was done through manual labor with hand tools, or using draft animals.²⁴

It was not a stroke of genius to imagine that a machine could lighten this load. Water mills had been used since antiquity for a later stage of wheat processing: grinding grain into flour.²⁵ By the Middle Ages, mills were used for many mechanical tasks, such as sawing lumber, pumping bellows, or fulling cloth.²⁶ The idea of a threshing machine specifically is mentioned in English royal documents as early as 1636.²⁷

What is difficult is to make a threshing machine work reliably. Whereas grinding, sawing, and fulling are tasks requiring mostly brute force, threshing requires a certain amount of delicacy and precision. When inventors began seriously attempting to solve the problem in the 1700s, some of their machines bruised the grain, or broke the ears off the wheat instead of opening them up.²⁸ Other machines simply broke.²⁹ To make a working, reliable machine took careful construction and craftsmanship.³⁰

By the late 1700s, there were working threshing machines, but there were only a few craftsmen who could make them, and so they were available only to a few local farmers. One source reported as of 1800 that “almost all the threshing machines in England” had been built by one “industrious workman, of the name of Stevenson.”³¹ It would be a few more decades before threshing machines gained wide adoption.

If the threshing machine was not hard to conceive, then why wasn’t it invented much earlier—say, in the 1300s?

First, there was no established professional class of inventors, engineers, or entrepreneurs. Farmers themselves, busy with farm work from dawn to dusk, had little time to tinker, and little spare cash for materials. Skilled craftsmen, especially those with mechanical skill such as millwrights, were relatively rare.³² Even they had little chance to educate themselves beyond apprenticeship: printed materials and mechanics’ institutes would come centuries later.

If an inventor did come up with a workable design for a threshing machine, there would still be the challenges of constructing reliable devices mentioned above. There were no machine tools to create precision parts. Wood is too soft for precision work, but metal was expensive, and cast iron probably didn’t even exist in Europe at that time.

Beyond manufacturing, there would be challenges of distribution. At the time there were no newspapers to advertise in, and no reliable postal service by which to receive orders. Shipping product by wagon over bumpy, rutted roads, or even by canal, was slow, expensive, and hazardous—especially for a delicately tuned machine.

Even if these obstacles could be overcome, to create even a small factory would require investment capital. There were no VCs or investment banks to provide the capital, and no way to set up a corporation to receive it. There was no patent office to apply to for intellectual property, to protect the investment; although if our inventor had favor at court, he might be able to snag a royal monopoly.

Social forces, too, created headwinds for invention. There were taboos against labor-saving devices.³³ New machines were fought and opposed by the laborers who felt threatened by them, who in many cases smashed and burned machinery.

Because of all these factors, medieval craftsmen with mechanical skill found it more prestigious and lucrative to make clockwork novelties for the aristocracy than to address mass markets with practical inventions. Even in the 1700s, when inventors were turning their attention to the threshing problem, they didn’t attempt to set up manufacturing and distribution: newspaper advertisements of the time announced that they would sell the plans or parts; the customer was expected to employ their own workman to do the construction. By the early 1800s, threshing machines were being constructed in machine shops originally set up to build steam engines, with the specialized tools and skilled workers needed for precision manufacturing.³⁴ And by the mid-1800s, railroad networks were built out that could distribute centrally manufactured products to a wide region. This is right around when threshing machines became widely adopted.

So an array of factors—human capital, manufacturing facilities, transportation networks, communications channels, legal and financial institutions, social attitudes and conventions—conspired against the threshing machine and held it back until the 1800s.

We can see some of the same factors at work against other inventions conceived ahead of their time—such as Babbage’s computing machine, dreamed up a century before the electronics technology that would make it practical, or many of da Vinci’s sketches, which are suggestive of machines such as helicopters or tanks that would only be possible centuries later. And the patterns are even stronger the further back we look in history. The wheelbarrow, for instance, is a useful, simple device, but it was missing from Europe for over a thousand years;³⁵ it seems clear that the same factors can explain this gap.

If we think all the way back to the stone age, it’s not surprising that a million years might pass before a major change to stone tools. Not only were there no inventors, no businesses, and no markets—there was no writing, no cities, no large-scale human communication at all. The main way for an idea to persist was to be taught as oral heritage, and the main way for it to spread was through the descendants proliferating. Better tools were almost biological features, and progress was more like natural evolution than human technology. Indeed, we can imagine that better tools were invented many times and then lost, either because someone failed to pass on the tradition to their descendants, or because a tribe died out instead of spreading.

Today, in contrast, it’s impossible to imagine any gap such as the wheelbarrow or the threshing machine persisting for more than a matter of years. There’s a whole professional class of entrepreneurs looking for such problems, and data on industrial processes and cost drivers to highlight them. There are professional engineers supported by a deep manufacturing infrastructure, from standard parts to computer-aided design software. New products can instantly address a global market using the internet and worldwide shipping networks. New ventures can be created and funded in days through standard legal documents and payment mechanisms.

It is the compounding effect of all these factors that explains why progress was so slow in the past, and the gradual acceleration of progress along a super-exponential curve.

The rapid progress of the last few centuries, then, was not a fluke or a lucky windfall. It is simply the continuation of a trend that goes back to the dawn of humanity. The causes are fundamental: accelerating growth is created by structural mechanisms.

Progress is not automatic or inevitable. Trends do not always continue. Despite the overall pattern of progress, history also contains many episodes of stagnation, regress, and even collapse. And as we’ll examine in later chapters, some of the drivers of progress have already gone into reverse. Progress requires choice, effort, and vigilance. But we should expect that it can continue. And perhaps it can even continue to accelerate, creating progress in the future that is faster than anything in history. Unprecedented growth rates may turn out to be the entirely precedented future.

---

For more about The Techno-Humanist Manifesto, including the table of contents, see the announcement. For full citations, see the bibliography.

1. OWID, “Global Average GDP Per Capita Over the Long Run.” This and the other OWID charts carry a footnote: “International dollars are a hypothetical currency that is used to make meaningful comparisons of monetary indicators of living standards. Figures expressed in constant international dollars are adjusted for inflation within countries over time, and for differences in the cost of living between countries.” More here: OWID, “What are PPP Adjustments?”

2. Meuhlhauser, “Three Wild Speculations from Amateur Quantitative Macrohistory.”

3. Allen, “Tracking the Agricultural Revolution in England.”

4. Howes, “Age of Invention: The Pull of Cities.”

5. Mokyr, The Gifts of Athena, 28-40; Jacob, Scientific Culture and the Making of the Industrial West, 105-111.

6. Harmand et al, “3.3 Million Year Old Stone Tools.”

7. Whitaker, Flintknapping, Kindle location 338-590.

8. Muller and Clarkson, “Major Transitions in the Evolution of Lithic Cutting Edge Production Rates,” Fig. 2.

9. Look at OWID, “GDP Per Capita, 1000-2019,” or Jones and Romer, “The New Kaldor Facts,” Figure 2.

10. Roser, “Breaking out of the Malthusian Trap.”

11. OWID, “Gross Domestic Product.”

12. Ibid, annotated with dotted line.

13. OWID, “Global GDP Over the Long Run,” annotated with dotted lines and growth rates calculated from the underlying data for the periods 1–1000, 1000–1500, 1500–1820, 1820–1950, and 1950–2023.

14. Delong, “Estimates of World GDP, One Million B.C.-Present”; Hanson, “Long-term Growth as a Sequence of Exponential Modes.”

15. Roodman, “Revisiting Hanson’s Long-Term Growth.”

16. Guinnane, “We Do Not Know the Population of Every Country in the World for the Past Two Thousand Years.”

17. Romer, “Speeding Up and Missed Opportunities.”

18. Jones and Romer, “The New Kaldor Facts.”

19. Others have expressed this theory before, including Kurzweil, “The Law of Accelerating Returns,” and Roodman, “Modeling the Human Trajectory.”

20. Bacon, Novum Organum, CX.

21. Mokyr, A Culture of Growth, 70ff.

22. Bush, “The Endless Frontier.”

23. Franklin, “From Benjamin Franklin to Joseph Priestley, 8 February 1780.”

24. Quick and Buchele, The Grain Harvesters, 11, 39-40.

25. Usher, History of Mechanical Inventions, 161–70.

26. Usher, History of Mechanical Inventions, 184–6; Sass, The Substance of Civilization, 149–50, 152.

27. English Patents of Inventions.

28. Somerville, General View of the Agriculture of East Lothian, 76; Ransome, The Implements of Agriculture, 140.

29. Somerville, Agriculture of East Lothian, 75; Quick and Buchele, The Grain Harvesters, 45.

30. Crawford, “Why Did We Wait So Long for the Threshing Machine?”

31. The Commercial Agricultural and Manufacturer's Magazine, 3–4.

32. Morgan et al, “The Mechanics of the Industrial Revolution.”

33. Jacob, Scientific Culture and the Making of the Industrial West, 68.

34. Crawford, “Why Did We Wait So Long for the Threshing Machine?”

35. Medieval sources attest to the wheelbarrow only in the late 12th century; although it was known in ancient China and perhaps ancient Greece: Lewis, “The Origins of the Wheelbarrow.”