What will it take for a great power today to remain one? Based on research from Jeffrey Ding, a professor at George Washington University, it’s probably not the flashy technology everyone is focused on. We discuss his recent book, Technology and the Rise of Great Powers, as he challenges conventional wisdom about how technology has shaped the rise of great powers like the United States.
What will it take for a great power today to remain one? Based on research from Jeffrey Ding, a professor at George Washington University, it’s probably not the flashy technology everyone is focused on. We discuss his recent book, Technology and the Rise of Great Powers, as he challenges conventional wisdom about how technology has shaped the rise of great powers like the United States.
Click here for the episode transcript
Jeffrey Ding:... think about some of the most exciting emerging technologies today. They're oftentimes in the headlines. Think for example, like electric vehicles. These are at the cutting edge of innovation. We're all talking about US-China competition in electric vehicles, but for all the importance of electric vehicles and how significant they will be, how fast growing they are as an industry, ultimately, there's a single use case, which is transportation. Whereas for artificial intelligence or technologies that are more general purpose in nature, they have all of these different types of use cases. So even among these cutting edge technologies that dominate the headlines, I think it's important to differentiate between ones that are more like general purpose technologies and ones that are not as enabling.
Brad Smith: That's Jeffrey Ding. He's a professor at George Washington University who I've gotten to know the past year. He challenges the conventional wisdom about how technology has shaped the rise of great powers like the United States. Jeff and I discussed the lessons from his recent book, including for artificial intelligence, what it will take for a great power today to remain one tomorrow. It's not necessarily the biggest breakthroughs that everybody's talking about. The story is more interesting and much more urgent than that. Let's dive in. Jeffrey Ding, up next on Tools and Weapons.
Brad Smith: Jeff Ding, author of what I think was the single most important book about technology published in 2024, Technology and the Rise of Great Powers: How Diffusion Shapes Economic Competition. Great to see you again. Thank you for joining me today.
Jeffrey Ding: Thanks so much, Brad. Really glad to be here. And that's a really high praise coming from you. I appreciate it.
Brad Smith: As you know, I loved your book so much, we bought 600 copies of it, literally. Started by giving it to Satya Nadella, our CEO, to people across the company, to people who worked for me. I think oftentimes in life you accidentally discover a book, and that was the case for me. My son, who works in Washington DC at the Rand Corporation, was working in Seattle for a couple of days and he had a copy of it, and I walked over and I picked it up and I said, "What is it?" And he showed it to me and I said, "Can I have this?" And he said, "You can take it. I'll buy another one."
And I absolutely devoured it because in part, to me, it speaks the history of technology and how it has changed the course of nations over three centuries. It speaks to very much what I have seen from my perspective at Microsoft in the third industrial revolution around computerization. And it speaks to how everything needs to come together for a country truly to succeed. So let's dive in and let's walk through some of the pieces here. I want to start with one aspect. Your book is all about GPT, but not ChatGPT, not GPT in the technology sense of a generative pre-trained transformer. It's a more long-standing use of the acronym GPT. Tell us what it stands for in the world of economics.
Jeffrey Ding: Yeah, in the world of economics and economic history, GPT stands for general purpose technology. And so this refers to what some scholars have called foundational transformations that are engines of growth. They have this potential to usher in huge ways of productivity growth because they are so foundational and they have a wide variety and range of uses, which gives them the potential to spread across the entire economy. So think electricity as a prototypical GPT or the computer or the steam engine. And excitingly for today, AI, as this potential next GPT.
Brad Smith: And when you think about it in life, we go to the store, we order online, most things that we buy do one thing and one thing only, and they do it very well. A lawnmower cuts the grass. A light bulb illuminates a dark room. But as you point out something like electricity, it actually ultimately can diffuse across every sector of the economy and change everything.
Jeffrey Ding: Yeah, and even think about some of the most exciting emerging technologies today, they're oftentimes in the headlines. Think for example, electric vehicles. These are at the cutting edge of innovation. We're all talking about US-China competition in electric vehicles. But for all the importance of electric vehicles and how significant they will be, how fast growing they are as an industry, ultimately, there's a single use case, which is transportation. Whereas for artificial intelligence or technologies that are more general purpose in nature, they have all of these different types of use cases. So what's happening in electric vehicles, all these new advances in electric vehicles, they're not fundamentally changing what's happening in the field of artificial intelligence. But what's happening in the field of AI is fundamentally changing what's happening in the field of electric vehicles. So even among these cutting edge technologies that dominate the headlines, I think it's important to differentiate between ones that are more like general purpose technologies and ones that are not as enabling.
Brad Smith: One of the real conclusions in your book is that it is fundamentally general purpose technologies that have been the fundamental fuel for each major industrial revolution. Tell us a little bit about that.
Jeffrey Ding: Yeah, so as you alluded to, the book goes back through these past three industrial revolutions, starting with the first industrial revolution around the 1760 to 1840 time span, the second industrial revolution, the 1870 to 1914 time span, the third industrial revolution, which refers to US-Japan competition in the 1960s through to 2000. And so, across all of these periods, I was trying to trace which technologies were crucial, which technologies were the critical drivers of why one country was able to sustain economic growth at higher rates than its rivals.
And so a lot of the usual suspects came up. In the first industrial revolution, you have the cotton textile industry, you have the steam engine, you have all these advances in iron machinery. Now, I was tracing through some of these, and in each of these periods it was the general purpose technology, not necessarily the leading sector, the fast-growing industry like cotton textiles in the first industrial revolution, it was more this gradual industry of improved iron machinery that went through incremental advances, that ushered in this process of mechanization across all these different industries, not just in cotton textiles, that spurred Britain to technological leadership in the first industrial revolution. It was a similar story that I found throughout these other two industrial revolutions as well.
Brad Smith: One of the things that I find so interesting about that is because a general purpose technology can drive innovation across the entire economy, because it can boost productivity across the economy, it can therefore drive higher rates of economic growth. And so a country that uses the technology more broadly is likely to benefit if it does so better than other nations at the same time. And that is, I think, what jumps out about the nature of different even general purpose technologies. You look back at say the United Kingdom in the 17 and 1800s and said, "Gee, iron working and iron based machinery could be applied across every part of the economy, but cotton and textiles was always, by definition, going to be more limited," just as you were saying, electric vehicles are important, but they will never have the breadth of impact of say, AI, at least as we look to the future. And so it actually helps us think about the role that different technologies can play.
Jeffrey Ding: Yeah, and that was the starting point, because the story doesn't end with just technology, right? Technology is not the sole determinant of which countries can be successful in adapting to these industrial revolutions. But you have to first understand the different types of technological trajectories. Just the idea that innovations in cotton textiles would make their mark through a very different pathway than innovations in iron based machinery. And so one pathway is more about which country can monopolize leading edge innovations in these single sectors. The other pathway, the GPT pathway, is about which country can facilitate and enable the widespread adoption of these GPTs across a wide range of productive processes in all different sectors of the economy.
Brad Smith: I think to some degree, you are challenging conventional wisdom because oftentimes conventional wisdom in national capitals is if there is a piece of technology that is defining the future, then everybody focuses all of their attention on how they can be and remain at the forefront of it. If the future is about computer chips, let's make sure that our country's chips are better than anybody else's. Let's make sure that we're producing our own computer chips. Let's not have to buy anybody else's. And what you point out is actually, that's a nice thing to do. Nothing wrong. It can be a useful thing to do, but what matters more is making sure that these computer chips are used across every part of the economy so that every part of the economy can accelerate based on it. And you actually don't even have to have your own as long as you have secure source of supply. And that really, I think, upends some of the traditional national security based thinking, certainly in Washington, D.C. I mean, how have you found as the person who has put this forward, those conversations in that community?
Jeffrey Ding: There is some credence to the conventional wisdom. That being said, I think in conversations with policymakers in DC I think the usual pushback is to say, why can't we do both at the same time? And I think you even mentioned that in your comments and your questions, is that we can walk and chew gum at the same time. We can try to ensure that, for example, in D.C. or in Beijing, that we're self-reliant on these key technologies or that we prevent these key innovations from leaking out. And then at the same time, we can also pursue this approach of running faster and ensuring that general purpose technology is diffused throughout the entire economy. I think a few things I'll say to that, one is that sometimes it is hard to walk in chew gum at the same time. Sometimes there's limited political capital, resources, time, and talent to spend on a problem.
And a lot of that time and talent and resources, a lot of that is being spent on what I call the leading sector model, the innovation centric model. And I don't think as much time is being spent on the diffusion model. And then I think the last point I'll make is sometimes those two pathways do come into tension. So for example, China right now faces tension between its focus on this indigenous innovation strategy where it's trying to create a self-reliant technology stack that's in line with this innovation centric model. But that's going to slow down its ability to adopt general purpose technologies at scale, where you should be open to a diverse range of suppliers. You shouldn't be that concerned about who is supplying the technology, at least according to my theory.
Brad Smith: And I'll even take it maybe half a step further by saying, if you have the world's largest economy and it's vibrant and productive, then you will have the resources to invest in the leading technology, including leading military and security-based technology. But if you don't have the world's largest economy, ultimately over the long run, it becomes very difficult to sustain the technology investments that were required. And I would say that is the story of the 20th century in the Soviet Union. The Soviet Union could not match American economic strength.
I would even say, and this connects more closely with your book, I think to some degree it is the story of the United States in comparison to the United Kingdom. In World War II, the United States could build a navy of vastly superior size to say, the UK. And in part, it was because the United States had surpassed the United Kingdom in the second Industrial Revolution, first harnessing the power of electricity more broadly. But the point you make in your book that is not yet appreciated, I think by most folks, is that what the United States really did was invest in this new general purpose technology called machine tools. Tell us about the role of machine tooling and how it built the US manufacturing behemoth really, of the 20th century.
Jeffrey Ding: Yeah, I think this was one of the most interesting parts of researching this book. And what was fascinating going through all these different histories was discovering technologies and sectors that never became the largest sector or the flashiest sector in an economy, but still played a really crucial role in incubating these general purpose technology pathways. And so machine tools in the case of the second industrial revolution was an example of such a case. So innovations in machine tools, which allowed one to shape and cut metal and wood in more precise ways, that process lets you have standardized and interchangeable parts. And so when a bicycle broke down, you didn't have to recreate the bicycle from scratch, you could just replace the part that was broken. And you can see how that's not just limited to the machine tool industry. That process of interchangeable parts manufacturing could apply to any manufacturing industry. And so this system became so widespread in the US that it became known as the American system of manufacturing. And I argue that this was the crucial GPT that enabled the US to become the preeminent economic power.
Brad Smith: And what I find so interesting is it does connect back with sort of leading sector or leading edge innovation. It's easy to look back at, say, the first half of the 20th century and say, well, that's a 50 year period that gave birth to an automobile industry, it gave birth to an aircraft industry. It gave birth to frankly, all of the industries that were needed to make everything that was critical to winning the second World War. But they all were based on this same advance in machine tooling, interchangeable parts, the ability to manufacture at scale and at lower expense. That is what made Americans success possible.
The other part of your book, which I think is of equally profound importance, really, is the piece that says, well, if you want to get a technology like this, a general purpose technology adopted or used or diffused, the word that economists used, in every part of the economy, you need to focus on a skilling architecture. Basically, the ability to ensure that there are people across the economy who know how to put the technology to work. And you looked at this in all three industrial revolutions. Take us through what you saw emerge from your study.
Jeffrey Ding: Yeah, I think across all three industrial revolutions, what really stood out to me was that in each of these cases, the country that eventually became the technology leader did not necessarily have a comparative advantage in terms of training the best and the brightest, the top scientists, the expert scientists in all of these different fields. Actually what emerged was the secret sauce was located in the institutions that widened the base of average engineering talent.
Brad Smith: When you talk about an average engineer, what do you mean by the word average?
Jeffrey Ding: Here I'm not saying that this person's ability or potential is run-of-the-mill or anything. I think more the meaning I'm trying to get across is the average technical literacy among your country's pool of AI engineering talent, or whether a person can meet the baseline level of being able to implement a general purpose technology, say implement a large language model or an AI advance in their particular sub-industry or application sector. And so I'm not referring to the skills and talent required to develop foundational breakthroughs in general purpose technologies, but I'm looking more at the talent and the skills required to implement these technologies across a wide range of sectors.
Brad Smith: Now that makes a lot of sense because that is the critical ingredient to the diffusion or adoption of technology on a widespread basis.
Jeffrey Ding: That was an insight that really struck me when I realized that after a GPT emerges, oftentimes you have an entirely new engineering discipline. So think about mechanization and mechanical engineering and then electricity and electrical engineering. You're very familiar with the computer and the skilling infrastructure that came to be entrenched in this computer science discipline, which even though it doesn't have the word engineering in its name, it's very much an engineering oriented discipline. And so these institutions that can broaden the base of engineering skills and knowledge are so important because they systematize and standardize knowledge in mechanical engineering in terms of adopting interchangeable parts manufacture.
They systematize and standardize knowledge in software engineering principles that are required to spread computers across different sectors of the economy. And that also lowers the barrier to entry for small sized businesses, for individual developers, for individuals to gain knowledge in these fields and ultimately spread information between the GPT sector and all these different application sectors that are trying to figure out where is the GPT going.
Brad Smith: And you point out that the United States really was the global leader in terms of economic competition in both the second industrial revolution based on machine tooling and that manufacturing economy and the third industrial revolution based on computerization. And in each case it was the American educational system that played such a fundamental role in the country's success. Let's start with the second industrial revolution. You talked about the United States and mechanical engineering. Share with us what produced all those mechanical engineers across the country.
Jeffrey Ding: The US was better equipped than Britain and Germany, its two main industrial rivals at the time to cultivate this broader base of mechanical engineering skills that was connected to industrial requirements and needs. And so in the first part US institutions with a large proportion of them being these land-grant institutions that were set up by the Morrill Act were really instrumental in terms of training sufficient numbers of mechanical engineers. So one of the US's rivals at that time, Britain just simply did not produce enough mechanical engineers. Its universities were slow to adapt to these trends.
The key point of advantage for the US over Germany was not in the number of mechanical engineers. Germany did produce enough mechanical engineers, but its mechanical engineering training in its technology institutes and in its universities was divorced from practical applications. And so I looked at evidence of what did the curricula at US institutions of mechanical engineering education compared to German ones, what amount of time was devoted to practical exercises in the lab and shop versus purely theoretic instruction. And German institutions devoted very little time to those practical real-world exercises. Whereas US schools were much more connected to the industrial requirements and demands. And so that knowledge, that mechanical engineering knowledge was much more standardized and systematized to enable the spread of interchangeable parts manufacturing.
Brad Smith: Well, and it's interesting because there are similarities to how things played out in the third industrial revolution. I think it's easy for people to forget today that in the year 1980 Americans were obsessed with Japan overtaking the United States. In fact, I think one of the points that you mentioned in your book is that there was a public opinion survey of the time that said that around 58 or 60% of Americans thought that competition with Japan was a greater threat to the United States than say competition with the Soviet Union. Interestingly, it's about the same percent today that focus on competition with China. I find the public opinion polls remarkably similar, and yet the United States ended up overtaking Japan when it came to computers, especially when software became, I'll say as important as hardware for innovation. And as you point out, it's because computer science education spread across the United States, especially across American colleges and universities.
Jeffrey Ding: Yeah, I think you did a great job in terms of setting the table there. And I just want to note that it wasn't just the public, it was also very prominent scholars and commentators that did speculate, and there was a widespread consensus that Japan was on the verge of overtaking the US as the number one technological power. But again, those predictions were rooted in Japan's success in these leading sectors. These single sectors such as consumer electronics, they didn't account for, as you were saying, the US's superior ability to spread computerization across all these different sectors. I think they're the story that I tell in that chapter is about how the US's decentralized approach to computer science education in which don't have what was happening in Japan, where it was much more of a top-down centralized approach to build up key centers of excellence.
That approach was not as successful as the one that the US took, which was allowing its universities to experiment with different models of computer science education. And that more decentralized approach allowed the US to develop a software engineering skill base that was much more tailored to the needs of the nation. And I think the other factor that has to be mentioned too was the US was able to tap into international sources of software engineering talent and attract software engineering talent from around the world. Japan was not as effective in doing that. It was a much more insular approach to training software engineering talent.
Brad Smith: One of the things I find so interesting about this, and I think one of the reasons that your book resonates so strongly for me and others of us who work in leadership positions at Microsoft, is we saw it all play out. We were part of it. In fact, it even to some degree reflected our strategy as a company. I cut my teeth in the software industry in the early 1990s in Europe. And fundamentally the strategy was very simple. It had two parts. You had to protect software under the copyright law so that companies would have to pay for it rather than just pirate it. And second, you'd invest in skilling, have all of these programs so people could learn how to use the technology so that they could support their organization and then deploy it in everyday life. And I think that's even part of your personal story as the son of people who came to Iowa City from China. You saw this growing up in your own home, didn't you?
Jeffrey Ding: Yeah, I think oftentimes we academics like to pretend that all of our writing is super neutral and adheres to this time-honored scientific method, but I tend to believe that all forms of writing and research are also personal. And so for me, that connects with my parents' story actually. My dad came to Iowa City from China and studied computer science, and now has worked as a software engineer for over 25 plus years. And so part of the implication of the book is that US success in these emerging technologies might rest more on people like my dad, then the heroic inventors that we see valorized on the covers of magazines and in the front pages of newspapers.
Brad Smith: So if we take all of this and then we look to the future, because that's ultimately what you help us do through your book, it's a new century. We're entering the second quarter of it. It's a new technological age. We, I think, both think based on AI's potential as a general purpose technology. So what is your prescription for the United States or any country that wants to put itself at the forefront of economic opportunity and growth using this technology?
Jeffrey Ding: I think the starting point is it seems like so many technology strategies, it's almost as if you could find and replace the specific emerging technology and just use the same boilerplate language and apply it to some other emerging technology. It doesn't matter if it's or electric vehicles or blockchain or AI. And I think for me, the starting point is to recognize that these technologies are all different, and AI is unique in part because of its potential to be this general purpose technology. And so based off of that starting point, the key move is to not solely focus on achieving and monopolizing cutting edge innovations in AI, but move towards a more diffusion-centric approach.
And that could mean different things for different countries. I think some of the specific policy proposals I outlined for the US case involve giving more backing to community colleges as an alternative pathway to training AI engineers. I think with some of the legislation that's already out there, such as that landmark CHIPS and Science Act, there's already different planks and starting points in those pieces of legislation, especially the STEM workforce development planks of the CHIPS and Science Act that we can build on. I've been excited looking at some of the things that other countries are doing. I know that Canada has this very cool internship program that is a public-private partnership that allows students to get internship experiences in emerging technologies. And so I think there are all these different policies that countries can explore, but for me, the starting point is almost a mindset shift away from the only thing we can do is invest in cutting-edge R&D and make sure technology secrets don't leak out towards what does a diffusion-centric mindset, what space does that open up for all these other really exciting policies?
Brad Smith: And I think that is the race ahead. I think it is a race that any country can enter if it focuses on how to put technology to use as rapidly and broadly as possible. And Jeff, what I'd love to do is sit down again and compare notes in six or 12 months because this to me is the conversation that the world needs to have, the country needs to have and is starting to have. And the questions that you just mentioned, those are the critical questions. What is the future of training and education for the use of AI? How does one advance it? How does one encourage its adoption across the economy, and not just in one or two sectors? How does one do it in a world where one needs...
I think to think not only about as we think about this new technology and what I love so much about your book, what I found consistently in more than three decades at Microsoft is the best way to predict the future is actually to start by learning from the past. That to me is what your book does. So thank you. I know I'll see you in the coming months. I look forward to that. Let's talk some more and bring everybody up to date.
Jeffrey Ding: Yeah. Thanks so much, Brad. Really fun to engage with you on these topics.
Brad Smith: You've been listening to Tools and Weapons with me, Brad Smith. If you enjoyed today's show, please follow us wherever you like to listen.
Our executive producers are Carol Ann Browne and Aaron Thiese. This episode of Tools and Weapons was produced by Karina Hernandez and Jordan Rothlein. This podcast is edited and mixed by Jennie Cataldo with production support by Sam Kirkpatrick at Run Studios. Original music by Angular Wave Research. Tools and Weapons is the production of Microsoft, made in partnership with Listen.