This week our guest is Mark Mills, a senior fellow at the Manhattan Institute who recently authored The Cloud Revolution: How the Convergence of New Technologies Will Unleash the Next Economic Boom and A Roaring 2020s.
In this episode, Mark and I discuss how the latest advancements in materials, machines, and information are unlocking a profound new paradigm represented by the cloud, the impact of which Mark argues has been severely underestimated. As part of this conversation, we also explore the ways such tech will impact automation, governmental regulation, and our on-going tension between climate and energy production. As Mark describes it himself, he’s a realist and an optimist, and both of those things are conveyed in this information rich episode.
Music by: Amine el Filali
Mark Mills [00:00:01] Well, automation is what human beings have been trying to do for all of engineering and industrial history. They want to take labor. Human labor out of accomplishing a task, whether it's making something or performing a service. And that's been going on since the dawn of engineering. And so people are. What's different this time? Yeah, it is. It's better.
Steven Parton [00:00:37] Hello, everyone. My name is Steven Parton and you are listening to the feedback loop on Singularity Radio this week. Our guest is Mark Mills, a senior fellow at the Manhattan Institute who recently authored the book The Cloud Revolution How the Convergence of New Technologies Will Unleash the Next Economic Boom and a Roaring 2020s. More specifically, Mark and I discuss in this episode how the latest advancements and materials, machines and information are unlocking a profound new paradigm represented by the digital cloud, the impact of which, Mark argues, has been severely underestimated. As part of this conversation. We also explore the ways that such technology will impact automation, governmental regulation and our ongoing tension between climate and energy production. As Mark describes himself, he is indeed a realist and an optimist, and both of those things are conveyed here in this very information rich episode. So without further ado, everyone, please welcome to the feedback loop. Mark Mills. Right away with authors and anyone who's spent the time like you've spent investing themselves into a subject so deeply. Can you just talk a little bit about to start? What motivated you to write The Cloud Revolution?
Mark Mills [00:02:04] Well, what I tried to look at what I spent a lot of my life doing, whether it's writing about technology or investing in it, but especially writing about it is to look at the sort of step back. And it's it's a trivial thing to say, but it's actually hard to do. So to step back from the details, look at the architecture, the pattern of what's going on, but that requires that much you think you've seen the pattern to dove back in the details to figure out if the pattern is true without engaging in confirmation bias. That's obviously hard because we're all we're all creatures of our own biases. But I think there's something really big going on in our economy. And it's it's more than more than just computers. Computers have got faster. I mean, yes, they have. That's like saying airplanes got faster, which they did for a long time, you know, and then it stopped because physical limits were hit. And we'll hit physics limits on physics. We know computing as well. But that won't end the impact of computing any more than. You know, people thought we'd be having daily supersonic flight today. If you looked at the trajectory of the rate of improvement of the average aircraft from 1920 through about 1960. At 40 years, it was pretty amazing. And so a lot of people extrapolated that thought we'd all be flying supersonic on everyday flights and doing, you know, hypersonic around the world. And it turned out that was harder than people thought. Doesn't mean what. Never do. It was harder than people thought. A computing's very much like that where everybody's focused on speed. Remember the early days of computing? All the advertisers talked about how many gigahertz the CPU was. Then they stopped talking about that because the clock speeds didn't increase very much. Computing power went up, especially what you got for your dollar. But that didn't get much faster. Same kind of physics reasons. So I wanted to look at the implications of the semiconductor revolution in the context of everything else across the society. And I divided the world into the three buckets that are the only three buckets of what? Everything. Things that make up everything in civilization. Civilization is made up of just three big buckets of things, machines that we make to do stuff, grow, move, fabricate, transport, whatever. Machines of all kinds. Lots of kinds of machines. Materials that we build everything from. The quality materials we build stuff from is change or dumped materials because you can't instantiate anything without materials. We live in a material universe. The virtual reality exists. Physical machines. Just like in The Matrix. They're real anyway. And the third domain, of course, is information. Understanding the universe, the forces of nature, but also understanding the materials and understand the machines. It's obviously a symbiotic relationship among the three spheres. But when there's a revolution in all three spheres simultaneously, when there is. Profound advances that are materially different, no pun intended, than what we've been doing for decades. And they happened contemporaneously, more or less. That convergence is unusual and it's happening now. It's only happened once before in recent history, a century ago. From 1920, we had a very similar revolution in the three spheres of. And then of course, it was easier. The machine revolutions were, you know, the airplane, the car and electric power plant, those kinds of machines. And there were others, but that was the big ones. And then, of course, in materials, the revolution then was in alloys far superior metal alloys and polymers and pharmaceuticals, which was the diet of the chemical synthesis era and information we had. The dawn of instruments and ideas about science in the 1920s were truly remarkable, and we professionalized science roughly around that time. It had begun earlier. Same things are going on today, same same domains, same kind of confluence, which I think is, you know, profoundly optimistic for the world. That's not a naive optimism. You know, as I wrote in my book before, which was published before Russia invaded Ukraine, I said the obvious, you know, wars will always happen. Human beings are what they are. They keep fighting. I mean, I don't like that observation. But saying that we have an optimistic future is not a way of predicting that that we're going to live in a utopia without conflict, without political conflict, without racial conflict, without economic conflict. All those things are still locked into human nature. So they happen in parallel. So that's that's a long way of why I wrote it. But I wrote it because those things were in my head for some time in my work and my reading, and I've written a lot about each of those things separately and then. You know, I felt like it was an opportunity to synthesize it into a sort of one magnum opus about what's what's going on in the three domains and what it means for jobs, for health care, for entertainment, education, things that matter to people.
Steven Parton [00:07:22] Yeah, well, to start, can can you just kind of first show us how that convergence matters to the cloud? How how's that connection take place between this revolution of these three domains and the way the cloud is rising to, I guess, ascendancy?
Mark Mills [00:07:38] Yeah. Well, the thing that the thread in the book, of course, is the cloud, which is why I ended up titling it the cloud revolution. So the cloud epitomizes. The confluence of the three spheres is built from materials that are really different from anything we used 20 to 50 years ago. It's obviously an information machine, but but far more important is it's an infrastructure and we don't make new infrastructures very often. So if you map out the history of civilization, let's stick in moderately modern times. I mean, the canals were the first sort of significant transportation infrastructure in the world and certainly in America, followed by railroads and, you know, telephony, telegraph and telephony and highways. Airways and the Internet. That. I mean, I've just about covered the whole that the whole schmear, right? That those infrastructures are incredibly important because the reason to call them infrastructure is that they infuse all of the structure, structure of society, and they're foundational. People use highways for much more than going to work. In fact, only about a little more 20% of total road miles in America are devoted to go in commuting pre-COVID lockdown and post-lockdown. It's only a couple of percentage points lower. In fact, it's lower still, but not a lot lower. Surprisingly, traffic around the world is back to where it was. Is any of us who've driven lately have figured out. But the highways are used for moving goods, for entertainment, for, you know, McDonald's was created, fast food restaurants were essentially created because of the highway. But you wouldn't call McDonald's a transportation technology. So we call a lot of things tech today that are built on information highways. But they're not tech. They use tech just like McDonald's was made possible by the highways. So infrastructure is a really, really important the cloud is an infrastructure, but it's not an infrastructure of communications. It's an information infrastructure that uses communications. When I say that, it's obvious when you state it, but it's important to have that in one's head because the cloud is different from the Internet as the Internet was different from telephony. It's a non-trivial step function change in what it represents. So it uses telephony, people use telephone, but, you know, it's actually using the infrastructure of it, telepathy, both wired and wireless. It uses the Internet, obviously, but it doesn't have a lot of the clouds. Functionality isn't on the Internet. There's more intra cloud traffic in a data center than there is traffic to and from the data center. And the more we do complicated things. A.I. to do machine learning to understand how to make that what how a virus is operating. For example, all that data churning doesn't isn't on the Internet. The queries are on the Internet. The knowledge it's gleaned from doing the turning comes back on the Internet to the scientist. But the majority of the data traffic is inside the cloud. So it's using the Internet as a way to connect to the world where it collects data and to human beings that want to know what the data is. You know, data crunching maids. So the cloud is different and it's different in that structural sense, but it's also different. If I were measuring it in physical or dollar infrastructure sense, it's really different. I mean, if you measure it in miles, which you can, because it's connectivity, it may be invisible. There are there are visible cables, you know, fiber cables being dominantly now. But if you measure it in ten miles, it's orders of magnitude bigger than either highways or airways. I mean, it's an astonishing network of hundreds of billions of miles of connectivity, which continually expands because the virtuality, the virtual nature of the connections, keep expanding because we can vote. That matters. I mean, I would just say it's consequential when you have a network that is so big and expanding so fast. Or if you measured it in dollar terms, which is another way of doing it, because all infrastructures cost money and we build stuff and we have to spend capital on it. It's actually fascinating to note that the annual spending on annual capital spending to expand what we call the cloud is now bigger than the annual capital spending globally to expand the electric utility industry. Pretty it's this is consequential. And if you sell, which is going to grow faster in the future, well, you don't have to be an analyst. You don't have to be an engineer. Economists know the cloud is going to grow faster because simplistically speaking, electric demand grows roughly with population and wealth, roughly speaking, and the invention of new things to do with electricity, of course. All right. So there's that. But the the clouds growth grows with our appetite to do something with information. And the if there's one thing that's infinite in effect, then are very few things that are infinite. The physics of the world we live in, it's probably information, the data, because the magnitude of the scale of what we want to collect information about and the granularity with which you might want to study it, it's essentially unlimited. And as I make it cheaper and easier to do to access and use a data. I'm going to have to expand the infrastructure. And of course, the key the last key thing about the cloud, which is phenomenal, largely different and everybody knows this. But I don't think people really this is a case where the hype is less is less hyperbolic than the reality, which is pretty unusual because the tech community is is very, very accustomed to hyping things, you know, disruptive innovation. This changes everything. No, it doesn't. A lot of things don't change everything. And some things don't disrupt what you think. In fact, the disruptor gets disrupted. I mean, it it's complicated, as they say. But but here's here's the thing that is quite remarkable. Obviously, a marginal dollar spent on a highway, but at best gets you that much more highway. Sometimes it gets you a little more highway. We get more efficient construction and more efficient at producing the raw materials. On a marginal dollar spent on a airplane, airways get you a linear increase in the airlines, roughly speaking. And as airplanes get more efficient and they have, you get a little bit of a bump. But everybody knows the marginal dollars spent on information systems get you much more than the last dollar, because we're still on this declining cost curve, the so-called Moore's Law, but it's more than Moore's Law. So the analogy I made in my book is, is if we consider a measure not of speed, but of what people buy, it's let's use transportation infrastructures. If we look at the dollar spent like a fair to go somewhere, some distance, not not downtown but to go to another city or go to another country. The dollars per foot per second. But because speed matters, the precious thing is our time. We want to get there quickly. More quickly. So how many dollars you spend per foot per second of transportation service? Well, that's actually improved 10,000 fold in the last century. But that's why there's so many people traveling in the world. That's why tourism has become a huge industry. It's why. Because it's gotten cheap to do by transportation. If measured in the in the metric that matters, dollar in dollars per foot per second. And it's remarkable when you're map that out from the sailing ship to the stagecoach, to the car, to the airplane. And the next leap, which I write about in a book, is actually the drone. It's not Elon Musk's airplane. It doesn't spaceship. Rather, it doesn't get to there. It actually is. The reverse is the trend. Space travel. Turns out it's a lot harder than the hype would say. We don't have we haven't conquered the physics of gravity yet, but we have done a lot more on the other stuff. So you got a 10,000 fold improvement in the metric that matters dollars per foot per second for the transportation service in a century. The the era of computation that was pre-computer got better at 16 fold per decade. So which is not nothing. You got almost 2,000% more computations per second per dollar spent. Computation of second dollar got better. 16 fold per decade. 2,000% per decade. The computing era, the first computing era. Jump that up a remarkable amount. Right we went all the way up to let's say was. A 2,000% per decade sort of like electromechanical error, was a 700 hit per decade misspoke and the computing error was almost 12% per decade again, and the measure was computations per second per dollar. The cloud era is growing at a thousand fold, not 1,000%, but a thousand fold in computations per second, per dollar per decade, or put differently in ten years. The measure of merit computations per second per dollar set of travel feet per second per dollar. In ten years, it's gone up 10,000 fold. Took a century for transportation to go up 10,000 fold. So that has you don't have to be you know, that has to be consequential. There has to have a have meaning. How does it get realized that what what forms does it show itself? Well, it shows itself and how we discover on how to use materials. It shows how we build and manage machines. It shows how we do education, how we communicate. It shows it how we do research and shows it, how we do how we do medical development. It shows that how we can change the very structure of how we allow health care to operate and democratizes, democratizes knowledge and information in ways that have to be unprecedented. So we're all we're seeing hints of all that already. But I think they're just hints. I think that's why I wrote the book and the things that we're seeing or begin to flower in the next decade. That's why the subtitle is Roaring 2020, despite the mess we're in now. Yeah. The 1920s began with a mess, and we're just sort of copying the 1920s.
Steven Parton [00:18:02] But those hints you're talking about, it seems to me like they are the seeds of a positive feedback loop that once it catches its traction and, you know, kind of crosses the elbow of the curve, we're going to see that massive explosion that you're kind of talking about with that kind of awareness in mind, with the awareness that when you build an entirely new infrastructure like you've talked about, this is an entirely new infrastructure in a very small chain or list of infrastructures that we've done over the past century or two. Why do you think there is this pessimism? Why do you think this isn't something that's more hyped? If you know that these infrastructures tend to launch entirely new industries like an entire fast food industry like the highway did? Why is there not this very why? Why is there pessimism and not just pure optimism about the fact that like, oh, we're about to launch a whole new industry with a feedback loop?
Mark Mills [00:19:01] Yeah, it's just this is this the social psychology question as opposed to the physics question. Okay. Which I have opinions on because, you know, I'm out in the public space. I speak and write and talk and argue with people. You know, it's interesting. I have again in my book, I write a bit about this, the psychology of of communications. I quote Marshall McLuhan, who was a fellow Canadian who was famous for the statement, the medium is the message. But what he did is, as a clinical psychiatrist and researcher, was look at these big picture phenomenology about how communications affected people, how they reacted socially. So you know pessimism is is an easy is easier than optimism because if you're optimistic you get accused of being a Pollyanna and not aware of the factors, bad things in the world as bad people. That's the first thing that people comes in their heads is, oh, you're just not realistic. And also, you know, Niall Ferguson wrote a new book. I apologize to him for not getting the title. It's about the catastrophizing catastrophizing things. His new book and he writes about the history of catastrophizing. Human beings like to catastrophize things first. It makes a better story, whether it's a news story or a science fiction story. The apocalypse is a much better story than things are going to be. Okay. And, you know, I mean, that's I mean, I like apocalyptic stories. Who doesn't? I mean, if you think about all of the all of the great stories over all of history, they're about things that go wrong. And, of course, saving the world or saving the hero heroine or whatever. So, I mean, there's that. There's also, you know, my appendix of my book is right about forecasting. I give I spent some time and have for a long time. And I may convert the appendix into a book, in fact, about forecasting because others have written books on forecasting. Forecasting is an interesting problem because all of us suffer from presentism, which is one of the things I point out. We we look at where we are and we think what we live in is a unique time and uniquely bad. Or you. And you know, things are, you know, this time it's different. Yeah, well, some things are different, but the pattern may not be different. What's you know, the war itself is different, but it's still a war. And it started because somebody is stupid or somebody was really a predator. I mean, that's the patterns are there. The I think politicians generally do better. I mean, you know, the political zeitgeist that we're in is not new if you read history. Being an alarmist again because of a human nature, reacting to negative news is sort of the political leaning. So if you put it all together and then when things are actually not going well because things don't always go well and everything goes well, we have high inflation now. We've got a war in Europe. We've got, you know, the racial division and social debates about all kinds of stuff. And it feels kind of depressing. I mean, when I wrote my book, I didn't read the news in the morning because it was very hard to be focused on the big picture and not be depressed if you get the news in the morning. So but read it afterwards. So I so we are in a a time and it's it's never smooth. Sometimes things are commerce and things. Things are messier. We're in one of the messier times then rather than a calmer time. So it makes people depressed. I mean, that's just the political storm and the rank we have, which is could be depressing. The I don't care what party around. I think pretty much all my friends of both stripes are not happy with the state of affairs. I think only people pay no attention to the political news, are happy with the state of affairs, and the polls don't show that there are many of those out there, so it doesn't really matter. You feel sort of enervated as opposed to energized. And there's another factor, which there's a camp that I call the New Normal Lists in my book. These are the people and they're serious people. I don't mean to make fun of them as new an enormous that's a they believe that we live at a time of a new normal of slow growth that the innovations we have coming now are incremental rather than foundational, that once you've invented the airplane, you can invent it again. Once you've invented plumbing, it's a one shot deal. Fresh water and plumbing changed more lives and save more lives than probably any single invention, more than antibiotics possible, while probably comparable. This sort of say certainly, certainly more than vaccines. More people were saved just because of clean water and sewage. So these are these are what you've done at once. How do you how to get more? The mass production line is a one time invention. You got that? The car is a one time event. So the computer one time deal. So that those are all old inventions. They were a very big change, productivity propelled wealth. And since there's nothing more foundational event except better apps, better zooming, better video games. But you're all. Make money for people, but they're incremental and not society shifting inventions. That thesis, in fact, is true. We have we have lived through an interregnum of foundational inventions. It is true. You can't point to. If you look at 1920 and you look at your world today, there are very few things that exist today except that they're better than existed in 1920. One of them is the computer, which actually is consequential. The other is the cloud, which is really consequential. But if you took those two things out and you didn't understand what they meant, you would say, not much has changed. You can play video games and you and I can talk on a Zoom chat. That's nice. It's not world changing because we could have picked up a phone that Alexander Graham Bell invented and it's like 80% as good as a crappy zoom call. You know, we can hear each other just fine. And if we want to get together for business, I can get on a train instead of a plane with a lot of people still prefer and go see somebody or a car because there were cars and things. So that argument has truth that all those things are better, which is consequential, but they're not. The equivalent of them haven't been invented. What I wanted to look at was is is the answer to the question, are there things comparable to that which have been invented or not fully utilized? Not things you could pretend I could invent one day, but have been invented or not fully commercialized, not fully utilized, not deployed into society. The radio, for example, was a big deal, but it was invented 20 years before people had radios in their homes, Marconi and had made the transatlantic transmission very early after the dawn of the 20th century, and ships had radios for almost two decades before RCA introduced the home radio. And then it took off. And then within a period of about eight years, we went from no homes with radios to about 80% of homes. It was just as fast a rise as the rise of cell phones or smartphones. Very, very fast. Very similar. The knee and the curve when it hit, to your point, happened very fast. But it took two decades of engineering development to get to the knee, and it took the decade before that. After the fact, the idea of a radio wave predated the first radio by a couple of decades. That pattern is very common, which I write about my book. What I try to show is that we are very close to the inflection point of many things that were recently made commercial but are not a significant use by that. Telemedicine is not really a significant use yet, but everybody post-lockdown understands what telemedicine is now, but they don't know really fully appreciate what its power is. Except for a handful of people I was with a. An older woman recently who had an AFib attack. She is, I think, late eighties. And, you know, so she's not very tech savvy. She had one of the carry around EKGs, which she put her fingers on and used her smartphone to look at and can to can tell if she feels bad, whether she's having an affair with that. And as you know, the new Apple Watch can do the thing for you automatically. This is that kind of thing, plus more. It's really consequential. That's very different than where the world was ten years ago. And but it's the share of the population that have access and are using what I would call democratize diagnostics for for medical health conditions. Trivial, right. But the tipping points is visible. It's just in front of us. And that will bend the cost curve in health care, that will make health care different and democratize it. And these things will be consequential because if you bend the cost curve on health care, we free up the money for other things like fun, like tourism or PayPal, or for buying a ticket on a, you know, air taxi, which. You know, it's been hyped a lot, but it's it's it's a little harder than a lot of people realize. Harder in a regulatory sense, not harder. Harder in the engineering sense. But it's around the corner too. Now, we don't have to guess anymore. It is probably 300 plus variants on piloted and unpiloted air taxis based on various versions of a quadcopters drones, multi rotors, all kinds. So there's not just one design because there's a lot of ways to approach this, but that quantity. And there's billions of dollars of money flowing into that space with venture funds and not just venture funds. Airbus is all in some ways Rolls-Royce and Shell. So as Boeing, they will they will they will come up. There will be a product, a service. And you don't have to guess that's going to happen. But it's not obvious who the winner will be because it's like all engineered things. No, nobody knew. If we were using the Internet cloud as an analogy, nobody knew that the world would see Amazon being the successful dominant cloud player back in the nineties. Nobody knew that. I mean, I we're all around. Then you can read what people said. You can Google it up. Nobody knew Google was going to dominate in lots of players and lots of contenders, but you knew that it was a big deal, which is why Wired magazine was practically frothing at the mouth right about e-commerce in the early days, if you remember, it was just. Irrational exuberance over eCommerce. And it took a while, though, you know, where we went from like 0.1% retail and ecommerce at that time to probably if you do all retail at 12, 14% now, but that's that's 100 fold increase in those 20 years. That's that's consequential.
Steven Parton [00:30:08] Yeah. Are there are there things right now that you're seeing maybe in one of those three buckets that you can give us examples, either of convergence of new foundational shifts or maybe something that you'll think in ten, 15 years it's going to be the Amazon and Google of today we're going to look forward and say, Oh, that's I should have known that in 15, 20 years that was going to dominate.
Mark Mills [00:30:33] You know, this is that is I am going to eventually do a sort of investor map, sort of take my book and do a parallel map of how to think about investing in each of the trends. And of course, investing too far ahead of the curve isn't worth anything right for an investor 15, 20 years, nobody cares. And it's really hard to guess the next 15 to 20 months. That's where all the money is. May Of course, but so I don't want to get into the Robinhood camp of a meme investing in this space, but yeah, sure. I mean, if you think about some examples in each of the three domains, the machines and materials in information and I and I didn't try to cover all possible examples of my book, but I tried to focus on were iconic examples. That were demonstrably and clearly real and not notional. As opposed to this could happen if somebody did. Why? No, this is what's already happened. And people, businesses are doing it on the machine space. It's the answer is mobile robots, anthropomorphic robots. I mean, we've liked robots. Bui I mean, society, the apocalyptic writers, the rise of the machines. I mean, robots are actually in Greek literature going back to 800 B.C. The idea of mechanical, you know, animatronic men, women, animals. Very old idea. The word robot is what people know was created by a pole. Who Coppock who wrote, you know, restrooms, you know, robots. And he created the word. So we'd like this out for a long time. It's. It's hard to make an anthropomorphic robot. The reason you would make an anthropomorphic robot is that where humans were anthropoid were we live in the Anthropocene and helping humans in their human environment is maximized. If the machine can operate in your environment with you without hurting you being an aid, whether it's on a factory line, in a warehouse or in a hospital or at home. And so it have to be anthropomorphic. It's if it's on wheels, it's it has to have roads or flat spaces. So that's always been the goal for that class of robot for obvious reasons. It's taken a confluence of materials, machines, better motors. The electric motor itself has gotten fivefold, which is, you know, 5,000% better in the last 35, 40 years, which is really remarkable. Shrunk down in size, better batteries of lithium battery, better actuators, better sensors, the constellation of things. They've all come together and. Wait. We don't have to. Yes, we can. We can use YouTube or Dr. Google and look for videos of pro commercial robots doing pretty remarkable things. Boston Dynamics, of course, gets the gets the prize for this for the best stunts with robots. And of course their spot many which was prodigious commercially earlier this year is a commercial anthropomorphic Robert a dog but it's you know it's anthropomorphic in that sense. This this is no longer speculation and since robots to manufacture them set aside their application which that's a little harder to think about, some there's some easier and obvious applications. You know, the Navy and DARPA funded it because they want firefighting robots to go where people operate, to put fires out without putting people at risk. That was their motivation. So it has to be has to be walk where people walk open doors like people open and put out fire. So but the fact that there are machines are the closest analog to manufacturing. A robot was manufacturing a car. The automobile is a single, most complex and expensive product that human beings buy. Your house is not a product, it's an asset. So whereas the products are depreciable and you know, that's what they are, the car the robot is is more complex than a car but is very similar is so I would I propose in the book and I think it's. And I think it probably happens within a decade, but may take a little longer. Manufacturing robots will be as big an industry as manufacturing automobiles. It's an industry that doesn't exist today. Wow. There aren't really. In terms of the mobile robot universe, there is industrial robots, the arms chair, and that that's consequential. So if I if I switch the material side and say, well, what would be a revolution? We we already know is happening materials. It's a it's a class of materials called metamaterials. There's lots of new materials like graphene, the new classes of semiconductors for high power control, electronics. Gallium nitride is soft and carbide. And there are a whole set of what we call smart materials, self-healing materials, self-assembling materials. But if you're going to pick a material class, which is remarkable, is this these metamaterials are using by using the kinds of machines we use to make. Not microscopic but nanoscale speak transistors which can become manipulate nature at the atomic level the way physics Feynman imagined you know the room at the bottom and make materials that exhibit properties that don't exist in nature like invisibility. The favorite one, of course, is invisibility and making a surface invisible as light appears to pass through it. But it's not, actually. Glass is no longer theoretical. It's, of course, being done. That has really interesting implications for all kinds of products. The other material that's maybe more immediately consequential are the class of materials called Bioelectronics, or so-called suntan recall trends in electronics. These are biocompatible. Electronics, therefore biocompatible computing and sensing that is programed to do something and then disappear self-destruct harmlessly so. Put simply, too simplistically, the idea that you could include in your vitamin pill or whatever homeopathic thing you want to take each day in its side, it would be a biodegradable computer that you would take each day. I want a1a day computer. Why would you eat a computer? Because it would measure and tell you lots of things about your body's microbiome and your state of health. Wirelessly connect with your personal body area network and tell you a lot about your your own, your own health. And give data to your physician when you see them, when you don't no longer have to ask the question, What did you eat two days ago? Who the hell remembers? But if you have the data right, this kind of thing is not crazy. So biocompatible materials are already in production, already FDA approved. So we don't have to speculate that there's an industry there. That industry will be as big as the silicon computing industry at some point. And the information side, the obvious candidate is a I. Which is profoundly misnamed. And I think Professor McCarthy is still I would probably, I don't know, ever probably agree. I think I've seen it right as much that it was a bad term in the term. Artificial intelligence has salience, roughly to calling a car, an artificial horse or an airplane artificial bird. We can make artificial birds. We have made them right. You can see Aurora, Optus, they fly. They look pretty cool. Very, very realistic. They're not very useful. They're stunts. Airplanes are more useful than birds, but they perform a similar function. A.I. is very useful and can do things that we can't do. But it's not it's not intelligent. So part of the misnomer is in in the language and there's a large swath of the engineering community, as you probably know, that's it's trying hard to flip the nomenclature around instead of AI's because it's called like intelligent automation, where the the automaton is not whether it's a virtual automaton, which would be an AI answering a question or whether it's a physical automaton, a robot walking. Doesn't do a computing. Because that's not what you want. Computing implies a precise answer. In the real world, we just have to have advice that's probably right in the right direction. It's how you drive cars. It's how you answer medical questions. That's intelligent automation. So A.I. machine learning are extraordinarily powerful tools. And we've just begun to figure out first how they operate, how they how they work and how they'll work. But we don't have to guess that it's getting cheaper fast. We don't have to guess that it's important and useful because it already is being used. And we don't have to guess that we're on the early edge of the curve of making the stuff better, because there's an incredible arms race going on now in the semiconductor world to make better, faster, cheaper A.I. chips. And so we're sort of, I would say, for picking an analog to the silicon CPU era versus the A.I. era. Not GPUs, but inference chips. Let's call them. It's kind of feels kind of like we're about 1995. Hmm. You know, we're past the first stage of seven. The first intel CPU was 20 years old at that point. And it was pretty, you know, computer chips were pretty good. But, boy, have we made the better since 95. I mean, holy moly.
Steven Parton [00:40:09] Well, this this tells me that, you know, as you're saying here, we don't need to guess with robots. We don't need to guess with machine intelligence or AI or AI. With that being said, it seems like we don't need to guess about automation in general. What do you feel about the impacts of that in terms of, you know, bringing it within the scope of the Roaring Twenties or of the societal impacts? What is automation as it seems inevitably, as it seems to be inevitably happening, going to do in the next ten, 15 years to society, work, you know, what have you.
Mark Mills [00:40:46] Well, automation is what human beings have been trying to do for all of engineering and industrial history. They want to take labor. Human labor out of accomplishing a task, whether it's making something or performing a service. And that's been going on since the dawn of engineering. And so people are. What's different this time? Yeah, it is. It's better. That's how it's different. But we're accelerating it a little faster than any time in history. Again, because in large part because of the cloud, the cloud brings information capacity to do an automation. What was very difficult to do in most of history. So the automation is a wealth creator. So the simple the simple, simple reality is that we have arrayed forces to allow us to automate far more things than ever before, both knowledge tasks and physical tasks that will that will accelerate faster than in the past, which is wealth creating, because the very definition of productivity is fewer inputs for more same outputs, which is what creates wealth always has, however, they create these anxieties. You know, this is not the first time people have claimed automation will kill labor. We won't have any jobs for anybody. And I caught some of this in my book that there's lots of books that have gone back and found. You can find hundreds of quotes. You don't have to be a researcher now with Google because it automates that function. You can find lots of people claiming there be no more work for anybody because of automation. In one of the graphs to put my book is an obvious one. If automation were job destroying, then the one trend you would see for the last century in America. Well, just pick the one centric a pick two centuries because we've been automating farming in manufacturing for two centuries at least would be the unemployment rate would continually rise. It would just keep going up and up and up as as automation progressed, because we've been automating everything I mean, everything from steel to farming to, you know, to services. You can't there's nothing that we aren't trying to automate and haven't successfully automated compared to years ago. And instead what the unemployment rate oscillates based on other factors. It has what's gone up over the last century or two is is per capita wealth because of productivity, which in effect is almost entirely because of automation. So that doesn't mean we don't disrupt jobs, that 60% of the kinds of jobs, the classifications of jobs that existed in the 1960s don't exist as jobs today. So automation, in effect, destroyed those jobs. That's obvious, right? Where from the typing pool to the the famous and infamous teller. If you go back further in time, you got. So the Luddites were right, referring to the famous Luddites who broke all the automated looms because it destroyed their jobs. It did. It did. It took their jobs away. So there are there are jobs that are eliminated. And it does cause it can cause social disruption. And this is where social safety nets come in. That's why we have unemployment insurance. But the overall result, our society is vastly better. The main thing different today from all other periods were automation and it goes through so spurts and cycles and it's not linear. You know, you automate things and then and then you can't do better and you have to wait a while to I know how to automate something different so it's fits and starts. But for the first time in history, the tools that allow better automation are the very same tools which makes it easier for people to find jobs, get retrained. Because the the thing that we all know is true is that if you need a different skill or you have to go somewhere else to find a job. Yeah, I can date myself. I mean, I was obviously a practitioner of the business world before there was an Internet. I mean, because the Internet wasn't useful till the eighties. So you had only people born in the eighties have lived in the era when there's never been anything about an Internet. But prior to that, if you want to find a job in an adjacent city, your job disappeared in your city. It's very hard. You'd have to ask somebody to send you a newspaper, or you had to go the library and look at the newspaper for the a city somewhere else. You had to do it every day and go through all the different papers in different cities to find a job for the skill you had. This is beyond obvious, easy to do now. Now whether you should or want to move. Americans are very Americans, very mobile. So that's not a big impediment for most people. But this has been made easy because of the Internet. It's been even easier because of A.I., because most job search stuff now runs through an AI engine, both on both ends for you and for the for the job seeker, the job, your job, the employer. And the training gets easier because of virtual reality and automation and computing. So. We've never had a point in history where the. The disruptor was also an enabler to smooth the disruption, which is, I think, a remarkable thing. And it's promising in this sense, because I think the disruption will go faster this time, which means you we better hope we have a better way to ameliorate the negative effects. There will be there will be negative effects. It's not it's not a it's not avoidable, as you know. And I don't say that in a callous way. I mean, again, it's like saying there's going to be wars in the future. I don't wish there's wars. There's going to be wars. We'll have depressions again, recessions. Those aren't caused by technology. They're caused by politicians.
Steven Parton [00:46:12] Yeah. And what do you think? Speaking of politicians, of the regulatory landscape, for all of these things, how do we maybe, like you said, ameliorate some of the negatives? Do you think that the regulatory bodies are even going to be able to keep up with something like this convergence when it hits its inflection point? Is the human condition ready to navigate that fast of a transition, a phase transition? Nick, what are your thoughts on the the regular regulatory aspect?
Mark Mills [00:46:41] Well, I think the idea that the pace of change is happening faster than we can manage is I'm not don't I'm not in the camp. I think that's that's true. And I think no, that's not true. Just requires reading history. So you have to sort of categorize what we mean by a pace of change. I mean, but if we look at what was going on in the 1920s and the pace of change that occurred then, contemporaneous with a lot of social disruption. Race riots, the horrific race riots in the twenties, really grotesque. People can read their history. They should read what was going on. That was pretty, pretty, pretty, pretty grotesque. If we look at the social disruption, the pace of change that occurred, what was called the second steam age in the late 1800s. And the same was true in the early 1800s. And by the way, the same was true in the 15, 14 and 1500s, which was the first industrial revolution. So and by pace of change, I mean the rate at which new things emerged that changed how we could do commerce, how we could war, fight, or how we could communicate, or how they they've happened frequently over history at a very rapid pace. So human beings are pretty resilient, adaptable. And I like it. We may fight, argue and, you know, vote for people that other people don't like on all that stuff. That's always true. But we but we we manage through it. It's not an unique thing. So I'm not pessimistic about that. And I don't believe that we're at any kind of acceleration that's not manageable. So I said before, I think we're we're going to see an acceleration more like we had in the twenties, which maybe that makes some people uncomfortable because we went through an interregnum where the you know, the biggest change was that the World Wide Web came along and AOL, you got mail, and that seemed really like a big deal. Oh, things are changing fast. And then we got cell phones. Ooh, things are changing fast. Still, the telephone is more telephones. Okay, but we went from no telephones to every home having a telephone just as fast as we went from everybody having no cell phone, ever having a cell phone. So these are not new sort of cycles, but they are it's new to maybe our particular time this decade because we had a slower period for a while, got we got used to the slower pace of new stuff and it's picked up again. We'll be fine. You know, I just I think we'll well, I think it'll be exciting. That's why I think it will energize people. I think what happened in the 1920s, the Roaring Twenties, the flappers, the jazz age, this to my mind, it was no accident that those happened contemporaneously because the automation created wealth, the wealth created time for other things. People were kind of excited about what it meant and what the future was. They still argued about the negative effects. Yeah, I mean, of course. And it and we had a Great Depression after that too. But that wasn't caused by technology was caused by politicians. So. The political. But the question on the regulatory space is an important one because that that is a that is a new feature in our economy that wasn't as extant, let's say, in the 6070s. We have a much bigger regulatory state. For better or for worse, the good stuff is sort of we all agree there are things we like to regulate. But clean air, clean water, all that stuff. Okay, fair enough. But then the regulatory state sort of what, run amok? A little bit. I wrote a piece a while ago pointing out that the reason we have less manufacturing in America is not because we can't manufacture, is we chase them out of the country with regulations. They went to other places where there were regulatory burden was less onerous, not just the labor. Cheaper labor was. If you look at the data honestly, it might have been half of the factor, but it's probably a third of it. Two thirds was just the time value of money. If it takes me six months to build a factory versus three or four years, I'm going to go where it takes six months. And that's a purely a regulatory feature or the permissions. So the the regulatory state itself. Will eventually adopt the tools of artificial intelligence and computing. They will more automate or get better. That scares some people because, you know, the man is watching you. I don't think it's as much that the man is already watching you. But I do think some efficiencies will come into the regulatory state. What'll come faster is on the consumer side. The business side is that if the regulations are complex, if I if I have to hire a team of accountants and a team of lawyers, 3000 pages of different documents to figure stuff out, that's a lot of sand in the gears. But if I have an engine that can read it and pass it and say, for the new factory you're going to build if you do these things and you do it this way with this agency, these are path outcomes. It'll look better. The complexity is made simple by AI and analytics, and that's beginning. I mean, a lot of that's already already happening in infrastructure projects. It hasn't solved the problem in the sense eliminating regulations, but it's certainly taken a lot of the friction out and will take a lot more of the friction out. So I again, I'm, you know, expressing optimism about A.I. doing that, but I base it on. Evidence and facts. I know, as you know, in way of full disclosure, I'm a a nonoperating partner in a venture fund that looks only at software for energy. And one of the most common startup companies we find are companies that automate sort of the regulatory labyrinth, if you like, the paperwork labyrinth of systems and industrial systems. So you have fewer people spending hours doing that. Does that eliminate jobs? Yeah, sure. But they're jobs that most people you don't you don't want you. Right. Somebody do things that are productive. This is an unproductive use of time and labor.
Steven Parton [00:52:37] Yeah, I know we're getting short on time, but I'd love to briefly touch on something you mentioned there, which is energy. I know there's a lot to be said for energy and climate change, and that's that's a whole nother hours of conversation. But I know you mentioned before that the energy costs of the running the cloud is enormous. Given given that and what I would say is potentially an energy crisis for for lack of a better term, where we're at right now. Are you concerned about our ability to sustain the energy needs of the cloud?
Mark Mills [00:53:12] So the one know short answer. No, I'm not worried about it in the macro sense, but yes, I am worried about it in the micro sense. In the short term will make some we may make some really dumb decisions that will increase costs or degrade reliability by by being, I guess I'd call it, ham handed about aspirations to make changes faster than they can be made in industrial systems. Because energy is a big, big industrial supply chain infrastructure that involves lots of big machines, big trucks, lots of steel, lots of concrete glass. It doesn't matter what you're building, you use lots of materials, minerals, concrete, glass, mines. It takes time to mobilize for move, build and trying to do it faster and not impossible, but extraordinarily expensive and difficult. And so a lot of the. A lot of a meme is out there about an energy crisis are self-inflicted because they're they're means created by self-reflection. We don't have an energy problem in energy. Just spoken for physics. Perspective is infinite supply. It was the subtitle of my earlier book on this from years ago. Yeah, there's lots of energy flows available and materials in the planet. It's all about machines and materials and what we want to spend and where we want to put the machines and materials in. They're different choices. They have different qualitative impacts and different quantitative impacts. So the cloud is another measure of its scale and all of its manifestations. You know, building operating roughly twice as much electricity as the country of Japan uses for all purposes. So it's a pretty big infrastructures in energy equivalent terms. It's already surpassing global aviation. Hmm. So it was a non-existent energy using system 30 years ago because there was no cloud. Now it's now it's like aviation and it's growing faster, both in energy consuming terms and in productivity terms, in aviation. But the the desire to stop using hydrocarbons is sort of the center point of what's going on all over the Western world. Not so much. Not so much in Asia and Africa, setting aside what they say, what they're actually doing. You know, China's electricity is two thirds coal fired. That's true in Africa. And cold burn is going up this last year, not going down, gone up by almost a billion tons per year. And that's because the world needs electricity, not just for the cloud, but for everything lights, airconditioning, factories. So they you know, the world's not going to easily reduce its use of hydrocarbons. In fact, I've written about this many times. It we're not going to reduce it at all, in my opinion. What will happen is the net increase in energy needs for the world will be increasingly met by the addition of wind and solar and largely wind and solar, but also other things, but mostly wind and solar. There's not many other options and especially nuclear bringing up the rear, but I think taking over because the only phenomenology from that energy materials perspective and as you know, my book, I have a chapter called the The Energy Materials Nexus. Everything about energy is about materials. Materials are used to build machines and the materials are used to operate the machines. So you use steel and largely to build combustion burning machines. But, you know, if you if you build machines, they don't have combustion solar arrays, wind turbines. Okay. But you use a lot more materials to build those machines per unit of energy. When I say a lot more and most people don't really understand. It's amazing when you tell them this is not a design flaw. It's about 1,000% increase in mining of materials, metals and minerals to produce the same unit of energy from wind turbine or solar panel compared to a combustion turbine. Thousand percent. Okay. Yes you're not burning gas are call but or oil but you're you're mining a lot more stuff burning oil with the big machines. But maybe one day they don't have to burn all the right now they do. But you have to disrupt a lot of land and you have to get it from somewhere. That's in the physics of energy. I cover that not from a viewpoint of being pro or anti anything particular, but from a physics and technology perspective, profoundly optimistic about finding better, different ways to both use and produce energy than we have today. But they take time just back to talked about earlier by 20 over 20 2020 rule, you know after discovery of new physics to do something like a new form, a new battery storage chemistry. A new superconductor that's, you know, low tech, high temperature at room temperature. It's 20 years before you get to the first realization of that invention is a potential commercial execution. Then it's 20 years after that, roughly before it actually shows up as a commercial product, and it's almost 20 more years before it significantly impacts markets. Lithium ion battery. The chemistry for that was invented in the seventies, mid seventies by a chemist at Exxon, which is kind of ironic by itself. Exxon was planning to make a build ultium batteries for cars because that's they propel cars. They don't care. They invented it. The three people got the Nobel Prize for it because there are two other scientists at up perfected. But if you measure it from the mid-seventies, the first commercial lithium ion batteries, which are very expensive and used for electronics products, showed up in the mid-nineties. And then it was almost 20 years before Elon Musk's first Tesla s before the lithium battery had scaled and manufacturability. And we understood the chemistries and the designs well enough to make a car. Now, we're almost 20 years since of that, right? We're pushing not quite 12 or 15 years since the Tesla s launched and the global electric fleet of all, you know, all light duty vehicles is, you know, almost near 1% now. 1%. Yeah. Will we accelerate that over the next 20 years? Yeah, yeah, yeah, sure. 20 fold. Pick a number. But that'll get you to 20% of all cars. But the number of cars in the world will have gone up to 2 billion instead of 1.2 billion. And we'll still be burning oil for all the other ones. So what I write about in my book is that that's the nature of trajectories. And so we don't have this magic wand to change those things. So I'm I'm both a realist and an optimist. I mean, you can be both that and people are confused by like there's kind of like the patting your stomach and rubbing your head thing. It's possible to understand inertia of economic systems that are huge, but also understand that that if you're far enough along the path, then, you know, significant changes can happen. I think the biggest changes in energy we're going to see are not going to be in wind and solar. There'll be lots more of those electric cars. Electric cars use energy elsewhere. That's self-evidently both. And making the machine, it uses a lot more men. The average electric car uses 300% more copper, for example, going to have to mine a lot more copper in the world, which is got got consequences as well. But the the resurgence of interest in nuclear energy is, is maybe the single most optimistic feature of our future energy infrastructure, because we've spent 50 years trying to figure out how to make nuclear plants cheap, reliable and safe. And we're pretty much on the cusp of being able to do it now. Yeah. And it's not for want of trying. It's just turns out that that engineering physics has been very hard hit. Not not as easy. Gone a little slower than airplanes, but airplanes went pretty slow because, you know, if you compare to when it became a common a common way to travel is six days from from the first aircraft and is 40 years from the first aircraft to make it a fairly common travel mode. It was 50 years before a lot of people flew in the seventies, so it was 50 years. Space spaceships sort the same thing. Space is actually a lot harder than everything else because because the gravity thing. But, you know, Elon Musk is getting at Jeff Bezos for getting us closer. And but we'll solve that problem of nukes, by the way, not with chemical rockets. We're going to get to Mars on a nuclear propulsion in ways that are sensible and that'll help things on Earth. So and we'll get better faster because go back to A.I. The ability to simulate the design of a machine that's hyper realistic and what's called a digital twin has been imagined by engineers ever since the first computers. They said, Oh, this is great. I can put a digital twin of a human organ, I can put a digital twin of a nuclear reactor and make it operate like the real thing. Yeah, you can, but it takes a supercomputer to do that and it takes a super supercomputer. In fact, we know that it takes computers that can operate at exa flops, not petaflops, which is a really big number. Well, the first X a flop. The computer is now operating. And, you know, just ten years ago, people thought it would that computer would take 500 megawatts to run. That was the estimate. 500 megawatts is a nuclear power plant. I mean, it's enough to run a city of roughly a half a million people, but it turns out it takes 50 megawatts with a design we have 50 megawatts is essentially 747. That's what the power that's a lot of power for one computer, but it's a lot easier to do that 500 megawatts. And so we'll build lots of those and the next one will be five megawatts and we won't build ten times as many of those. We'll build 100 times as many, which is why we got to where we are with the cloud. Using Japan with electricity efficiency gets better. Our appetite to use data expands. But with that we'll find better photovoltaic cells, better batteries will event. All those things will happen, but they won't happen in the political timeframe that our our political class would like to have happen by throwing money at it. But it's good we need it. I mean, Silicon Valley lives and breathes on money. Engineers need money to build big machines. But some of these things are like that, you know? I don't think it's a a sexist joke because they say that you can't you can't have, you know, three women have a baby in three months instead of nine months by concatenating it. It doesn't work that way. You know, it's biology. It works. Everybody takes nine months. Engineering's very similar. When you build machines, you have to build the prototypes. Yeah, it works. Then you build the improvement of the prototype. You can parallel pass some things. But again, the digital twin allows you to build the first prototype in a virtual space, which can take some years out of it. Yeah. That's a long answer to the energy one you say. I mean, right. Energy is consuming everybody these days because we're we've done some really dumb things in Europe which is causing pain for everybody. And the damn thing was obvious. Trusting Russia. I mean, I wrote about this at the time saying I, I, I would prefer to ship American natural gas to Europe at a slight premium than trust Russia. And that is it not as a replacement for the windmills that they got that whatever the windmills aren't solving the problem. Yeah, they can't. They can't.
Steven Parton [01:04:19] Yeah, I would love to have that conversation. I already am taken taking you over time here, Mark. So I want to honor honored the step aside time we have here. This was a very information rich and optimistic conversation, which is a combination of things that I really enjoy and I feel like you don't get enough of. I will give you one quick chance if you'd like to leave a final word. Obviously, they will promote the book in the show notes and whatnot. But if you'd like to leave a final message or talk about anything you're working on, feel free.
Mark Mills [01:04:48] Well, I would say that would come back to where we started about. Why do people pessimistic? And the reason I wrote the book is I think we can be realistic about the political challenges that we have, that we have to deal with those problems the world has. But the reason I wrote the book is there's reason to be optimistic and therefore a reason to solve the challenges. Getting the politics right matters that means something different to everybody. But typically we start out I'm Canadian, that I emigrated to America. I think it's a great country. I think it is, in fact, one of the most promising provinces in the world to advance things for all of humanity. This not mean other people are smart, do good stuff. We just we just have a lot of advantages in America for a lot of reasons. So I'm I'm as an immigrant that adopted this country optimistic that will sort out our political differences eventually come to some reasonable compromises and and that's will be great for the country but profoundly good for the next 50 years. I think this is America century. And I don't mean that as a sort of a as xenophobic, you know, rah rah America, although I, I feel that way about this country. That's why I came here. You know, it's not like a, you know, people that come here come by accident, they choose to come here. I'm among them. Although I was a documented alien. Just for the record, I did walk across America today. I got my papers. But that's a whole separate argument. And we've always had documented undocumented aliens in America. By the way, this is a very old it's a very old a very old thread. But it is I think it's important to to have a sense of optimism about what we can happen. In fact, my favorite economist, John Walker, who's a Nobel class economist, his most recent book, he begins his first line with and I and I guess I'll quote a pretty accurately that economic growth depends far, far more on what people believe than most economists would admit. And he's an economist and believe in the future could be brighter is actually consequential. Yeah it drives growth it doesn't drive naive you know slobbering optimism it but it but it drives drives growth in reasons for compromise. So it's. And on that note only because I know in people's heads when I talk as an optimistic oh well having your revenues news today look what so-and-so said. Yeah I know of course I read it was I know what he or she said it's whoever he or she is. Doesn't matter.
Steven Parton [01:07:18] Well, I'll, I'll, I'll take that as an end note. I like ending on an optimistic note. So, Mark, thank you so much for your time, man.
Mark Mills [01:07:24] Thanks for having me. It's great.