Full Transcript for AI, Industry, and the Human Story with MIT’s David Mindell

books led you to that one, to give people like a flavor, but also who are your inspirations as like in, historian engineers or engineers who are historians that kind of. You read a book and said wow, this is I wanna do stuff like that and spend time documenting the process of technologies being born and sometime die and everything in between. David Mindell: I had some great teachers at MIT when I studied this stuff. Rose Smith, who was my dissertation advisor, is the preeminent historian of interchangeable parts Manufacturing. At the Springfield Armory in the 19th century. Actually Joel Moer is a big hero of mine who's written about. The enlightenment and the economic history of technology. That was a big influence on this book. And, he just won the Nobel Prize last week in economics for that work, Broadly recognized. Back in the seventies, there was a great show out of the BBC called connections, which was really. Most pe people who were technical at that time, remember seeing it on tv. James Burke had this great series about how these different technologies were developed and influenced each other in very non-linear ways. And that stuff always interested me. And then I was always interested in Thomas Pension's book, gravity's Rainbow, which is a great work of literature, but really about the. The V two and technology during the second World War and that, that got me interested in that. So lots of different reading about technology in American history, technology in European history. And my father was a kind of benchtop engineer who became a manager. That was never quite my ambition. But I did work for 20 years building. Autonomous vehicles and remote vehicles in the deep ocean and in aviation. But I always felt I like to do work that sort of captures the bigger picture about a technology and teaches people about how it evolves. Natan: So what are the books you wrote before that kinda shaped the journey towards the new Lunar society? David Mindell: So pretty much every book I've written has been something about how people relate to machines. My first book, which was 2025 years ago now was about the USS Monitor and Iron Cloud warship during the Civil War. Which was really the first time that sailors went to war inside a machine. It was almost a submarine, this machine. And I happened to write that when, at a time that I was going to see a lot diving in submarines, a lot. It was also soon after the first Gulf War when a lot of people were wondering about whether machines aren't gonna take over. Warfare and it won't be a human endeavor anymore. And those were topics that always interested me. My second book was about World War II and the history of control theory and computing, and how radar and computing and signal processing really influenced the origins of cybernetics. That was from the, early days of the internet getting really big. Then I wrote a book about lunar landing and the computers and the software and the lunar landing. Natan: The

fact that, I think the memory on that thing was like 4K or something like that. David Mindell: It was more like 30 2K, but it was an amazing computer that, the astronauts stake their lives on working perfectly for two weeks. They had 30 Natan: 2K on that thing on the David Mindell: apo. Yeah. And raised a lot of issues that are still very salient about how much of the task is built into the machine. How much should be left for the human, what role should the human play. What meaning does that have? Then our robots ourselves was 2015 and that was really about the coming age of robotics and AI and what we've learned from operating in extreme environments, deep ocean aviation space flight and warfare about what was then the coming technologies of driverless cars, industrial robotics, and ai and there are a lot of. Experience and lessons learned in those environments that at the time the kids playing around with driverless cars didn't really think that they were gonna have to relearn those lessons and turned out that they did. And then from 2018 to 2022, I led MIT's task force on the work of the future with Liz Reynolds and David ot. And we collected a whole lot of research studies by many folks around campus and then produced a book on the work of the future, the labor implications of robotics, automation, and ai. And that book came out right at the end of the pandemic. And. It just got me thinking about what is the origin of all this. The pandemic is a period where it became suddenly technologically possible to reimagine work. We'd all been playing around with Zoom, episodically and so on before then, but suddenly remote presence became a really viable way to conduct a workplace. And to me that's like a hundred year. For two Natan: years kind David Mindell: of change. Now it's gone back to some hybrid form. Yeah. And AI is also, changes in technology, changes in work, changes in global geopolitics, supply chains, economics just makes us feel like we're at a point where we really can and we really need to rethink some of the foundations of industrial society. So I went, let's go back to the beginning and the guys who thought about it the first time, is there anything to learn there? That's really where this. Lunar Society book came from. Natan: So let's dive in and start from the big picture. Studying technological revolution across centuries is fascinating, and it's actually a pretty hard thing to do and even harder to identify when it's happening. So what. If you juxtapose the moment, and maybe that's, you can introduce the concept of the lunar society and why the book is named that way. With what we are experiencing right now. Because I, I think it's like super interesting. People talk about the situation that we're experiencing oh, this is another.com or something like that. Which. I don't think it is. Maybe in many aspect it resembles it, but it's like more fundamental, given the proliferation of information technology, 50 years later and so on. But I don't wanna steal that thunder and want to get there but maybe you can talk us through lunar society up to this moment and like the tensions there. David Mindell: Sure to begin with, history is strange and the world that people live in and other times is pretty strange to us. And yet there's still people and there's a lot of things in common about technology how things get built, how they work, what makes them break, and so on. And so by going to such a different time. There's no sense that there's nothing new today. There's a whole ton that's new today, but we look at a different time and we can say, what are the things that are common into what these guys were experiencing and what we're experiencing? And that helps us understand what's new for us. And so the Lunar Society was a group of friends that met in England in Birmingham. Which was the center of metal Manufacturing in the early, mid 18th century. So 1730, 1740s, 1750s. And it was growing commercially quite a bit already without steam engines and without factories, just with a lot of, people with hammers really. And they made a lot of cheap middle products. Belt buckles, buttons, buckles for hats, even swords and small metal devices. And there was also actually a fair amount of gun Manufacturing going on. And one of those young manufacturers was a guy named Matthew Bolton, who was a kind of. Rising young, ambitious entrepreneur. He became very good friends with a doctor in town who was about his age named Erasmus Darwin, who was Charles Darwin's grandfather. And they actually got very interested in electricity. This is in the 1750s, and Ben Franklin came to town and he was already famous for his. Very important experiments and publications on electricity in the 1740s, and they did some experiments with Franklin, and they were so inspired by Franklin that they decided they wanted to. Kind of continue this set of interactions and so they began meeting to talk about the ideas of the scientific enlightenment. This is about 40 years after Newton's death, and the his ideas are really just beginning to proliferate through textbooks and other things. What of these new empirical scientific principles can we apply to our industrial pursuits? Bolton was a very innovative manufacturer already. He made great use of water, power and kind of mechanized tools in his shop. He built a factory next to a canal, then he extended the canal and they gathered a bunch of their friends. For the next 40 years, they met on the full moons, which is why they called it the Lunar Society, to talk about these different ideas, chemistry, geology, mining. Canal building, pottery, even experiments in living, how they raise their families. And very consciously sought to apply these enlightenment ideals to industrial and other parts of life. And one of those members was a guy named William Small, and he's a fascinating character. It's probably the least famous of the members of the Lunar Society and Small

was a physician from Scotland and the enlightenment was quite. Vibrant in Scotland for a bunch of interesting reasons. And small had just come from a stint teaching at the College of William and Mary in Virginia. And the day before he showed up to teach there, they fired the entire faculty and they said, you gotta teach everything. So for two years he taught every class at the College of William and Mary, and he did that using his enlightenment principles. He didn't know how to teach divinity or law. He taught history and Newtonian science and demonstration, and one of his students there was Thomas Jefferson, who took every one of his classes for his first two years. Small, then went back to England with Ben Franklin. He came to Birmingham Bolton set him up there as also a physician. Small, just was very good at organizing people and bringing people together. And one of the first things he did was say, Hey, I know this guy from back home in Glasgow. He's working on a improvement for a steam engine, but he is having a lot of trouble. And Bolton said, bring him here. I'd like to meet him. And that guy was James Watt. Bolton got together with Watt and they formed a partnership in 1776 that lasted for 25 years. And really. Perfected and commercialized steam power as applied to industrial pursuits. So in this little world, you have the. Incredible concentration of some of the founding fathers of the United States, founding fathers of the Industrial Revolution. Joseph Priestley, one of the founding fathers of chemistry. Modern chemistry is part of the group. Josiah Wedgewood, who's Manufacturing pottery at enormous scale for consumer goods, and it's just a remarkably fertile period where they. Bolton in particular, but the group overall really imagines an imagines a world driven by industry more than the world they lived in. And that's what I tried to capture in the book about what that meant for them and how they went about making it happen. Natan: The steam engine obviously changed the world. It's like the first kind of example of at-scale robust automation technology or mechanical power technology. David Mindell: Yeah. The steam engine was enormously important over the next a hundred years, if not more. Natan: Like you can't imagine Industrial Revolution without it. It David Mindell: depends on where you are. The American industrial Revolution was driven by water power more than steam power, actually. But that's 'cause we have the geography for it. But the steam engine was of obviously, of great importance or mechanization, largely because it added energy from the ground into the industrial system. Natan: But it also means you could, if you architected it correctly, like the mobility of that power that I think is so crucial. 'cause it's the first time that independent of geography, you could like. Put actuation where you wanted it, assuming you could heat water and David Mindell: Right. Natan: Pipe it and pump it. And David Mindell: so the steam engine really decouples the sighting of the factory from the local geography. Which is important. And like I said, Bolton built his factory near a canal, but he was very frustrated with it. The water froze in the winter. It dried up in the summer. He only had a limited number of months that he could use the water power. The first steam engines they made were sold to mines and Cornwall. But Bolton realized that they really should be used to drive factories and they started selling them to breweries and textile mills in London. And so for one thing you're able to move the factories to where they should be. They still need to get fuel, and that's not always so easy. But it's a lot easier than actually getting the water power. But again, really all industrial production up until that point is driven by human muscle power or animal power. A little bit of wind power, a little bit of water power. The steam engine just takes the energy of fossil fuels, quite literally, and injects that into the human industrial system. And that's been a story of growth. That's really what Joel Moocher studies is, how did we get this amazing growth in that following two centuries that just weren't around for millennia before? Natan: And before we dive into the new lunar society, I wanna fast forward for today for a second. The way I think about Industrial Revolution, 'cause we've heard, I think in the past decade people talk about industry 1.0 and 2.0 and 3.0 and four point the at 4.0 they lose me, yeah. I think that term negotiated itself out of meaning. No, I don't even know what it means today, I think we solidly arrived in this era where I feel like humanity is busy automating intelligence, or at least in love with the concept that it's automating intelligence and like we have different names for it, but basically all this AI stuff, it's like mechanizing, at least a portion of intellectual labor. Allowing what we perceive as reason or, design, diagnosis, code, writing, whatever it is to touch almost every aspect and it's like being implemented in various areas. So is there like corollaries to that very wide invention called steam engine and how it's applied to where, what we're experiencing today with ai, is that right? And is that gonna reshape our industry? David Mindell: I think the corollaries are, is that almost the definition of a machine is you're building some kind of human expertise and skill into a mechanism. I have a story in the book about Wedgewood who invented a really interesting way to measure. Temperatures above the boiling point of mercury, which is about 600 degrees, and he created this porometer with its own special temperature scale. The question was, did that replace the judgment, the eyeball judgment of the experienced potters? Who could tell how long the kiln should be fired just by looking at the color? And you know what it did do? It didn't actually replace the potters. That wasn't his goal. It enabled them to run much more

sophisticated profiles of temperatures. With better repeatability and to really explore their creativity in the other aspects of the process, which were much harder to do. And there's some, it's really a bit of a metaphor for AI in that most of the AI ais we have today are still ways of building human intelligence and human data and human experience into systems. They're still very highly engineered systems that human beings create and produce. Hopefully they enable people to do things that they would have difficulty doing otherwise. I was just at a big industrial facility. A maintenance facility on Friday, and they said to us, you know what? What the no coding and low coding AI tools allow us to do is now I don't need engineers to build our apps. I have the business SMEs, the experts building the apps, and we had just eliminated this layer of confusion and slowness, and now we can iterate very quickly with the people who actually know the business. They're the ones who are creating the software and the algorithms. Natan: Yeah, David Mindell: there are many instances like that. Again, not to say there's nothing new. There's all kinds of things that are radically new that we're just beginning to understand, but it is at least still true that most of these AI systems are extraordinarily good at aggregating massive amounts of human behavior, human knowledge, human experience, human generated data. Yeah. And to be able to have that at your fingertips changes the game in a whole bunch of different ways. Natan: When people compare, oh, this AI thing is just like the.com, I'm like, no. the.com was about information theory and mass distribution of media and access to that. If you think about it and how commerce evolved into that, and I think what we're experiencing now, especially if you think about it like from an economical perspective on this. This is about reorganization of labor and capital. I know these might not be popular topics to talk about nowadays, but we talk about this 'cause people are actually doing new. Work and it's just giving new definition to how we do work in a systemic way and affecting productivity. Whether this is work happening in an office or on the shop floor, assuming you can have access to it. And in that sense it has the potential to affect things like how we educate and like how we regulate and what should be our labor policy and how should we think about safety and all those things people were considering during the industrial Revolution, it's oh my God, this thing can kill us. We should not use this. Our horses are fine. David Mindell: Yeah. I think from a historian's point of view. It depends when you would locate a.com era, but if you call it the late nineties, up through the early two thousands, it's only 25 years ago. Yeah. So it's just the leading edge and the ais we have of today, you couldn't have without like massive. Population scale data being collected and absorbed and processed and stored and that's the breakthrough in some way that enabled what we have with AI these days. And so you might say the.com boom is just the leading edge of the changes we're experiencing now. And I think it's too early, but historians like to make periods and these are ways to think about the periodization. At the same time, these sort of fundamental ideas about intelligence and skill and labor and work and capital and how it gets built into machines, how that changes the nature of the work. And I think in particularly, who benefits from the improvements in productivity In what proportions? Those are crucial questions that. That are also changing. Those are fundamental questions. They've been with us for a long time, sometimes in different forms, and the new technologies as they emerge, find their way into these existing social financial structures. And historians like to debate like the technology itself, rarely. Comes with a set of social relations around it that are pre-coded into the technology. There are good ways or different ways to deploy it and there are ways to deploy AI that are surveillance corporate, very unfriendly to different types of workers at all levels of hierarchy. And there are other ways to deploy it that are very friendly and empowering and really enable people to be creative and do things with. Great, expanses of possibility that they couldn't otherwise. And sometimes two companies next door to each other have very different ways that they manifest these things. And I do think that's another one of the lessons from these longer stories is that how technologies are deployed and what ways by whom matters a lot in what the social effects are. And Augmented lean is a different approach than, a larger top-down corporate. Nineties lean is, as you well know. Natan: Yeah. David Mindell: And a lot of those details really matter and it behooves people to pay attention to them as they're rolling technologies out and Natan: deployment. It's so hard to affect change. And if we are experiencing like a new lunar society moment, and maybe we have new steam engines like AI and perhaps robotics or these are like, I think, specific examples of how the technology manifests and what can you build because you have. Powerful enough artificial brains and good enough actuation and sensing. So you can do that. But how do you see that applied to what's happening on the ground in the United States? We hear a lot of talk on the reindustrialization of the us. We need new factories. People talk about reshoring policy investments and there's like a ton of uncertainty around this. So how does that all come together? What's a meaningful reindustrialization look like in this sort of post Silicon Valley era? David Mindell: Yeah, no, that's a great question. And I think one of the challenges I see in a lot of places is the, like you said the adoption. What does it take for people to do things differently? Sometimes it's just generational change. The old guys retire and the

young people come in and they do things differently, and that gives you a kind of 30 year time cycle for changing things over. It's not clear we can afford that 30 year time cycle right now. And at the same time. We've hollowed out the middle of the workforce in a lot of ways. People, I call master chiefs who are. It's a sort of a Navy analogy that enlisted people who rise to the top and really understand how the systems operate. And those, by the way, are also the people in industries who adopt new technology. They have the credibility with the workers. They understand the day-to-day importance of the process and how things need to change. And they're very often the ones that management has to rely on to accept or reject or adopt energetically a new technology. And because we've hollowed out that middle of the workforce in this country I see that as one of the real bottlenecks about finding and making sure, how many startups do we know and see where. It's full of smart, bright people with a lot of energy, and yet their sales cycles all depend on kind of the people in the middle of these large organizations who are not well incented to change, have all the time in the world, sometimes counting toward retirement, and are not particular technology enthusiasts to begin with. It's that interface, which I feel like is one place that we really need to work on and make it easier for people to. Experiment and adopt new technologies and quickly understand whether they succeed or fail. Natan: It's like America has become post World War ii, extremely successful, both with its technology, but also economically while, exporting jobs overseas. But the result of that is it had this like false notion of, oh, we have time and therefore we can afford to be more. Conservative, which is weird. Like you go to Europe and like they're just comfortable being conservative, but I think this is a new phenomena, mid seventies and on in the US if you ask me. I'm sure people in Silicon Valley would not agree, but like we did export a lot of our Manufacturing innovation overseas. While teaching the rest of the world. Heck, this robot was invented here out of the industrial robotics. I'm speaking like a huge American robotics company, like they're German, they're. Japanese, they're Swiss, they're not American, David Mindell: right? There's a very long list of those technologies that were invented here. And then manufactured solar panels, electric cars, solar panels, semiconductors advanced chips. Natan: Yeah. David Mindell: Numerically controlled machine tools. It's a depressingly long list and it keeps going. Natan: So success leads to conservatism. And now what? Like how do we get out of this? David Mindell: And some people say never compete with China on Manufacturing. And the answer to that is probably we shouldn't on a lot of things. And what are the things that we should be competing on? We should be Manufacturing here. Aerospace and defense is easy, one because of

the national security dimensions of it, Natan: for sure. David Mindell: But there are other more strategic technologies, including semiconductors and obviously the software side of ai we've still been really good at. But even that, I wouldn't say is to be taken for granted for the next 10 years. Natan: But how are we doing that? We see the heartland versus the coast and there's so much policy and. Administrations are changing, but they're I think there's consensus around the need to do a lot of important infrastructure level work to, to change this reality. But it is so slow, and at least for the decade plus, I've been researching and working on in this domain. Yeah, there's been a ton of progress, but not so much like the big macro trends don't show that. Like we haven't doubled in a decade, our Manufacturing workforce. David Mindell: No, that's true. As my colleague David Otter says, United States actually has not reaped any gain in growth rates from the inequality in work. Other countries actually have much less hollowed out workforces, and they do just as well on GDP growth as we do. We're not particularly productive that way At the same time. We have been good at the kind of advanced r and d type of leadership. It's not clear that's gonna continue either. There's real threats from China there. And certainly not gonna maintain that by cutting our r and d budgets. And at the same time, there are aspects of a thriving r and d system where you pull all the smart people into what's gonna happen in 10 years, and you then undervalue. The folks in these process innovation roles which is really where the new technologies get adapted and adopted. We've encouraged a lot of students to do four year degrees, get jobs as HTML coders maybe when they might actually be happier being an electrician or a line manager working with their hands, actually making things and in contact with the material world. Those are actually pretty good jobs these days. Many of them. Natan: Yeah. I think people with more extreme AI paranoia, they're like, oh, we're going to tell our kids to be plumbers 'cause they won't need to worry about having a job. 'cause you always need plumbers and AI is not gonna do that. But then you think about all the robots, that could be great. Plumbers actually exist now. I don't know, maybe it's the people piloting those robots, but David Mindell: yeah. Natan: While somewhat pessimistic, what are we excited about? What do you see? Yes, China is there and I don't think supply chains are being untangled so quickly. And I also think I. Change takes time to really kick in. If you think about the stretch between 1995, emergence of internet at large scale to today, it's a few decades, good few decades. So what are you optimistic? What do you think is gonna go well? David Mindell: I use the word industry in the book in a very deliberate way. Because before it was associated with smokestacks and factories. It actually was and still is a human virtue. The industrious revolution actually preceded the Industrial Revolution. Industrial Revolution would not have been possible without the industrious revolution in England in the 1730s, forties, and fifties. And I think Americans are actually very industrious. They have a great deal of industry and the more we can promote that as a cultural value, which is really what the book is aimed at, I actually don't make any policy recommendations in the book on purpose, but rather talk about. How do we transform a culture to value industry again, and to value people's work in a way that encourages them to continue to work, but more importantly, to innovate in the processes that they're a part of, rather than removing them from those processes and expecting other people behind walls to innovate in again. Many of the technologies we're talking about are able to be deployed in a lot of different ways. There are plenty of jobs that people don't like to do. Nobody's really thrilled with driving a forklift around inside a freezer all day. That's an awful job. Natan: Sounds fun for a couple of shifts, David Mindell: There's no shortage of those things. And at the same time, it is interesting and worthwhile to figure out. Where the forklifts ought to go and how to organize their flows and so on. I watched a video this morning on, it was just a CNC machining training video, and when CNC machines first came out, people said, oh, it's gonna de-skill all the machinists, and they're all gonna be thrown out of the work. This video was really about the machinist sitting in front of a CAD station and incredible skill that machinist had in right. How to design and select the tool path to make a very complicated helical propeller kind of part. Natan: And then they became full. Full programmers basically. David Mindell: Yeah. Without writing code, but with they being able to interact with a well-designed interface that helps them understand how to remove the metal from the part to make the part you want. That's the part that's the interesting human experience skill and turning the little wheels on the tool was, wonderful craft skill, but probably not in the long run. As valuable as that. Insight into actually Frederick Winslow Taylor called it the art of cutting metals. Yeah. We still need people who know the art of cutting metals, but they can teach the machines to do it. Natan: That's a great insight. So David, thank you for the conversation today. Before we wrap up, I do wanna ask you something about. The New LU Society. What's your favorite feedback or questions that you got, like from people who read, interacted with the book that made you think about, what the moment we're in or how it should impact the way the US becomes more industrial or. How to build up a new lunar society that is real and active and working. David Mindell: I think some of the most interesting and maybe surprising feedback I got is from people who read the book who are not people who think about what you and I and lots of other people we hang out with think about all day. Who never thought about how a part might be machined, who never thought about where your iPhone comes from. And they were interested in the book. They enjoyed the history, but that was actually a way to draw them into some of these questions that, again, a lot of people in our town think about all the time, but most Americans don't. Everything has to come from somewhere. Everything has to be made by someone. And just providing folks who don't ordinarily. Think that way. A little bit of way in to some of these questions that are complicated and complex, but you don't have to be an engineer to understand them. They are about the world we live in is an industrial world. Everything we do is fed by industrial systems. We move by industrial systems. We eat through industrial systems, communicate through industrial systems. There's a chapter in the book called Supply Chains Are Us. They're not something that somebody else made. They're, every time we order a box or order a DoorDash, we're drawing on these supply chains, and it's been very gratifying to see people who don't ordinarily see our world that way to begin to think about it in those terms. Natan: That's a great note to end on. We're all part of an industrial system and rely on them. David Mendell, the author of The New Lunar Society. Thank you so much for joining us today, and thanks everyone who is listening to us on Augmented ops. I really enjoyed it and it was great. Great having you. We should do this again soon. Great. David Mindell: Thanks ton. Narrator: Thank you for listening to the Augmented Ops podcast from Tulip Interfaces. We hope you found this week's episode informative and inspiring. You can find the show on LinkedIn and YouTube. Or at Tulip dot co slash podcast. If you enjoyed this episode, please leave us a rating or review on iTunes or wherever you listen to your podcasts. Until next time.