Gabfest Reads: How Do You Solve a Problem Like Semiconductors?

Listen to this episode

Speaker A: Welcome to Gabfest reads for the month of March.

Speaker A: I’m John Dickerson, one of the hosts of Slate’s political gabfest.

Speaker A: This episode, I spoke with Chris Miller, the author of Chip War the Fight for the World’s Most critical Technology.

Speaker B: It’s a book about the history of.

Speaker A: The semiconductor, the wondrous process of how they are made.

Speaker A: And it’s a book that gave me a totally new view of global politics and economics.


Speaker B: It’s so wonderful, Chris.

Speaker B: There’s so much to do.

Speaker B: Like a good student.

Speaker B: I have lots of notes.

Speaker B: What, to you, is the best description of what a semiconductor does?

Speaker B: I mean, so people will say it’s ones and zeros.

Speaker B: Like, what is it at its most basic that we can then just imagine?

Speaker B: Okay, multiply that by 15 million.

Speaker C: Yeah, it’s it’s just a circuit turning on and off.

Speaker C: It’s like a light switch.

Speaker C: And when it’s on, you get the one.

Speaker C: When it’s off, you get the zero.


Speaker C: And it’s just lots of light switches.

Speaker C: Except rather than being the size of your finger flipping them up and down, they’re measured in billions of a meter.

Speaker B: And silicone was, as I recall from reading this and you’ll correct me was important because it could be both a conduit and opaque.


Speaker B: It could have different properties.

Speaker C: That’s right, yeah.

Speaker C: Semiconductor refers not only to the devices instead of our phones, but also to a class of materials.

Speaker C: So most materials are either a conductor, like copper wire, or an insulator, like glass.

Speaker C: And semiconductors can conduct when you apply an electric field on top of them.


Speaker C: And so the way the devices work inside chips is they apply an electric field, a current is opened, the switch is turned on, and that’s how you get the circuits flowing that produce the ones and the zeros.

Speaker B: And we’ll hop around here.

Speaker B: One of the things I love, just from the very start is you open up the book.

Speaker B: And of course, there’s a glossary that’s helpful for people like me.

Speaker B: But what I love was you start with a cast of characters.

Speaker B: What’s great about this is this is a book that is obviously rich with economic theory, world history, understanding.


Speaker B: There’s a fantastic series of sections on how semiconductors work and Euv works, and we’ll get to all of that.

Speaker B: But it’s a story.

Speaker B: And so you have these cast of characters which propel it as a story.

Speaker B: And I don’t want to give it away the way it starts, because that’s quite effective, though we’ll get into that topic later.

Speaker B: But tell me, as a place to start, Chris, is there a way to tell the story?

Speaker B: I know you do.

Speaker B: It in the whole book, but of chips and how semiconductors have become so central by the time you finish reading this book, I just see them everywhere now and think of them as this crucial central part of our life.


Speaker B: Is there a way to shorten that story?

Speaker C: Well, there’s one individual whose life really maps onto the chip industry.

Speaker C: And that’s Maurice Chang, who founded the Taiwan Semiconductor Manufacturing Company, which today produces 90% of the world’s most advanced processor chips, the chips in our phones and our PCs and the data centers, a company whose products that none of us can live without but that we are barely aware of.

Speaker C: And Morris Chang was present at the creation of the chip industry at Texas Instruments in the late 1950s.

Speaker C: And today he is still deeply involved in chip making in Taiwan, producing the types of chips that are, for example, driving forward progress and artificial intelligence.


Speaker C: And so his life has in some ways been the life of the chip industry as well.

Speaker B: And I want to get back to him because he seems to be kind of the one person who is able to ride the changes and now is in this position of extraordinary dominance.

Speaker B: But before we get there, was there for you when you started this project?

Speaker B: Were there a set of questions?

Speaker B: Was there a single question that you were in search of that you were hunting to figure out through this project?

Speaker C: I think there were three puzzles that I started with.

Speaker C: One was why is it so hard to make semiconductors?


Speaker C: They didn’t seem like they should be something that was so impossible to make, and yet just a tiny number of companies dominated their production.

Speaker C: And so that was a puzzle.

Speaker C: The second was why is it that China spends as much money each year importing chips as it spends importing oil?


Speaker C: Which is an extraordinary fact that I had no idea of when I began this research, but really seems to kind of redraw the map of what globalization looks like.

Speaker C: And the third is what explains the rise of technology across all aspects of society and economy.

Speaker C: And I guess I started with the hunch that we’d spent the last couple of decades talking a lot about social media and search engines, but had underrated the importance of the computing hardware that makes all of the big tech, the big tech firms possible.


Speaker C: And so those were the three interrelated questions that led me towards semiconductors and convinced me that they were actually sort of the hidden puzzle piece that explains so much of the modern world.

Speaker B: Why did Silicon Valley and America become the place where so much of the early work on this was done, where there were great innovations?

Speaker B: And Moore’s Law, which is almost a character in your book itself, comes out of the American chip industry.

Speaker B: What was a part of that original beginnings and that made it so successful in America?

Speaker C: Well, it was really a two step process.

Speaker C: The first step actually came from the government and the Pentagon, which was the biggest buyer of chips in the late 1950s and early 1960s at the start of the industry.


Speaker C: And the Pentagon wanted chips to provide miniaturized computing power in the nose cone of missiles to guide them more accurately.

Speaker C: And so the chip industry wouldn’t have existed.

Speaker C: It had not been for procurement from the Defense Department and from NASA.

Speaker C: But then the next step was taking these very expensive devices that were used in small production runs for military systems and making them an object of mass production.


Speaker C: And that’s what Silicon Valley did especially well.

Speaker C: It was turning these devices into something that was embedded into almost everything.

Speaker C: Not just computers, but eventually calculators and phones and today dishwashers and automobiles too.


Speaker C: And that was what the companies that came to define Silicon Valley companies like intel did particularly well take this high tech product and make it ubiquitous.

Speaker B: And that’s one of the things you say early in the book, which was really striking and a nice framing device, which is it’s not just the engineers and the brilliant people putting ever more transistors in ever smaller spaces, but in some cases, it’s manufacturing genius.

Speaker B: And then later, when you talk about all that goes into creating UV chips, the complexity of how Asml makes these high cutting edge chips, it’s mind blowing.

Speaker B: And I want to get there, but explain a little bit that idea that there are geniuses in this story that you tell that are not always just what we would think of as engineers, but there are these other kinds of skills that have been a part of the rise of this industry.

Speaker C: Yeah, I came away thinking that our way of conceptualizing scientific advances and technological advances put too much emphasis on scientists.

Speaker C: Certainly Nobel winners are important, but the pathway from a new invention to a device that changes society is a pathway that requires mass manufacturing with extraordinary efficiency.

Speaker C: It requires driving down the price so you can put it in everything.

Speaker C: And it requires visionary thinking about products.

Speaker C: And one of the things that really struck me is that none of the physicists who invented the semiconductors had any idea that these ships would be in cars or in smartphones or in data centers.


Speaker C: And it took a whole different set of people to have that vision of what are the new products that this type of technology can enable.

Speaker B: It seems that this is also it’s a story of making two things struck me.

Speaker B: Big bets, big gambles on do we go down the memory chip road or do we go down making CPUs?

Speaker B: Do we create generic CPUs or do we create bespoke ones for certain kinds of activities?

Speaker B: Euv, which I keep mentioning, and people will just have to wait till Chris explains exactly what that is.

Speaker B: But believe me, it’s worth it required this huge bet.

Speaker B: What to you are some of the big bets.

Speaker B: And then I should say the other thing that strikes me is there are also some huge mistakes or people fall off the map who were once considered to be the winners of this lottery for life.

Speaker B: But start first with some of the big bets.

Speaker B: And am I right about that?

Speaker B: That we’re talking about just enormous gambles at certain periods in this story?

Speaker C: Yeah, that’s right.

Speaker C: And today an advanced chipmaking facility can cost $20 billion.

Speaker C: So if you’re going to open up a new facility, that is indeed a massive bet.

Speaker C: These are actually the most expensive factories ever made in human history.

Speaker C: So the scale of the gambling is quite large.

Speaker C: I think there were a couple of key turning points in the industry.

Speaker C: One was betting that you could take technology that was designed for missile systems and put it in consumer devices and that was far from obvious to most people.


Speaker C: And indeed, a lot of the industry was perfectly happy trying to make a decent living on Pentagon procurement contracts rather than gambling that you could actually find civilian users.

Speaker C: That was the first one.

Speaker C: The second big bet was on creating international supply chains.

Speaker C: And I was surprised to find in this industry that the process of offshoring assembly to East Asia began in the early 1960s.

Speaker C: It was basically present from the beginning of the industry and it produced enormous efficiencies that we can talk about and also some interesting vulnerabilities that define the industry today.

Speaker C: And I think that the third big bet was the bet on concentration.

Speaker C: In the early days of the industry there were a number of different firms that could produce cutting edge chips of different types, whereas today there’s been extraordinary concentration, such that for many types of chips there’s just one or two companies that can make them and they make them in vast facilities with enormous capital expenditure involved.

Speaker C: That concentration has proven very, very efficient.

Speaker C: But the process of creating these vast companies was a big gamble and it would have been a very costly one had it gone wrong.

Speaker B: Why was the move to East Asia in the 1960s?

Speaker B: What encouraged that?

Speaker C: In the chip making process, the actual manufacturing of semiconductors is very capital intensive.

Speaker C: You need a lot of machines to do it, but then after you’ve got a chip, you put it in a package and that process has always been quite labor intensive.

Speaker C: And so in the earliest days of the industry there were efforts to scour the United States to find sufficiently low wage labor.


Speaker C: Some early chip companies opened assembly facilities on Indian reservations, for example, trying to find ultra low priced wages.

Speaker C: But nothing could compare it to Hong Kong, which was the first place that Silicon Valley firms opened offshore assembly.

Speaker C: And that dynamic has stayed present up to today, where there still is most assembly of chips into their packages.

Speaker C: Happens in East Asia and China and Southeast Asia.

Speaker B: Remind people what Moore’s Law is and it feels like one short way to tell this story is essentially everybody involved in the chipmaking business trying to keep up with or fulfill the promise of Moore’s Law.

Speaker B: And that that’s led to all of these, you know, the the offshoring, the innovation, the fact that we’re in this concentrated position we’re in today remind us what Moore’s Law is.

Speaker C: So Gordon Moore was an early engineer in the chip industry.

Speaker C: He would later go on to co found intel, and he noticed in 1965 that the number of components on each chip was doubling every year or so.

Speaker C: And that meant that chips were getting twice as powerful every year or two.

Speaker C: And that dynamic has been maintained up to the present for the past several decades.

Speaker C: Every two years, chips have gotten twice as powerful.

Speaker C: And what that means is that the chip industry has improved vastly more rapidly than any other part of the economy.

Speaker C: And I like to think about what’s an analogy in other sectors of the economy.

Speaker C: What if airplanes flew twice as fast every two years and that rate of change was steady for the past century?


Speaker C: It’s sort of impossible to imagine.

Speaker C: But that’s why remembering data on semiconductor devices costs less than a millionth of what it did 50 years ago because of that exponential growth rate.

Speaker B: I always assumed that the idea was you always went to cutting edge chips and that there was no use for anything that wasn’t right at the cutting edge.

Speaker B: And so I was interested to learn about the different kinds of chips that are useful in various different kinds of things.

Speaker B: But essentially keeping up with Moore’s Law means trying to fit more stuff in a smaller space, more powerful stuff in a smaller space, which requires extraordinary feats of design.

Speaker B: And it was hard to conceive sometimes about how small the etchings you were talking about were.

Speaker C: If you go to an Apple store today and buy a new iPhone, the primary chip on that iPhone will have 15 billion tiny transistors carved into it.

Speaker C: And so to fit 15 billion devices on a chip the size of your fingernail, each one of them is the size of a virus.

Speaker C: And they’re produced with basically perfect accuracy.

Speaker B: I think it was the Asml machine, you said, that could hit a golf ball and hit the moon.

Speaker B: What was the accuracy?

Speaker B: I’m butchering it, maybe, but the sense of accuracy of one of their lasers was what?

Speaker C: That’s right.

Speaker C: They could hit a golf ball sitting on the moon flowering that laser from Earth.

Speaker B: Oh, my God.

Speaker B: All right, so this notion of concentration which will get us why has the industry become so concentrated?


Speaker C: There are two main reasons.

Speaker C: The first is that if you produce more chips, you can gain economies of scale.

Speaker C: Costs get lower because you’re able to buy machinery more efficiently, acquire materials more efficiently.

Speaker C: So the cost is part of it.

Speaker C: But the second is the more chips you produce, the more you can hone your production processes and advance your technology.

Speaker C: And so today, it’s not a coincidence that TSMC, the Taiwanese company we mentioned, is both the world’s largest chip maker and the world’s most advanced chipmaker because they get data from every chip they produce and then use that to produce the next chip more efficiently.

Speaker B: And Chris, I want to turn us towards those sort of global politics right now and the current position before I lose you.

Speaker B: But I want to think about U.

Speaker B: S.

Speaker B: Semiconductor policy here now to kind of bring us up to the present.

Speaker B: It seems like the Biden administration and national security, as it thinks about chips, has kind of two prongs to it, and maybe they’re related, maybe they’re not.

Speaker B: One is recognizing there are vulnerabilities and shortages in this amazingly important thing in the world and trying to shore up those shortages.

Speaker B: So that’s one part.

Speaker B: And then the other is trying to basically deprive China of its chip making capacity.

Speaker B: Let’s take the first one first.

Speaker B: How do you assess both America’s vulnerability with this concentration?

Speaker B: You’re talking about the fact that TSMC is in Taiwan.

Speaker B: How vulnerable is the US.


Speaker B: Dependence on chips in the world right now?

Speaker C: The whole world is critically dependent on Taiwan for advanced semiconductors, smartphones, PCs, data centers, cell phone towers.

Speaker C: The entire tech infrastructure is basically impossible to build without chips from Taiwan.

Speaker C: And if there were a war in Taiwan or a blockade that prevented Taiwan from exporting chips, the impact on global manufacturing US.

Speaker C: Europe, China, Japan would be catastrophic.

Speaker C: It would be Great Depression scale in terms of its impact not only in our production of things that we think of as having chips inside, but also of dishwashers and cars and coffee makers.

Speaker C: And every device with an on off switch has a semiconductor inside, unless it’s the most simple of light bulbs, and a big chunk of those chips are made in Taiwan.

Speaker C: So we’re all critically vulnerable.

Speaker C: And as the risk of war in Taiwan has grown, I think people have become more and more concerned that we’re far from prepared to deal with that.

Speaker C: So that’s the real supply chain challenge.

Speaker C: The past couple of years, we’ve obviously had some pandemic induced supply chain issues.

Speaker C: They’re now mostly resolved.

Speaker C: The real problem is, if something goes wrong in the Taiwan Straits, we’re going to face an economic catastrophe.

Speaker B: All eggs are in one basket, but the Biden administration is trying to increase domestic production.

Speaker B: Is this right?

Speaker B: Is it TSMC is opening a plant in Arizona or something, I guess.

Speaker B: What is the administration trying to do?

Speaker B: And is it nibbling at the edges of the condition you just described?


Speaker B: Or is there a real effort here to build some sustainability or self sufficiency that could withstand something going sideways in Taiwan?

Speaker C: Yeah, I think self sufficiency is not the right frame to think of it because we’re still going to be buying materials and machines from the Netherlands and Japan far into the future.

Speaker C: So no country can produce advanced chips on its own.

Speaker C: But I think the Biden administration in Congress and this has been a bipartisan issue in Congress has been trying to reduce reliance on Taiwan precisely to deal with this scenario.

Speaker C: And the Chips Act, which is going to spend around $40 billion in incentivizing more manufacturing in the US.

Speaker C: Is going to have some effect.

Speaker C: But I think the key question is when we look at the big buyers of semiconductors companies like Apple, AMD Nvidia, they’re the companies that determine where TSMC build its facilities because they’re TSMC’s biggest customers.

Speaker C: And right now, they seem pretty comfortable buying their critical chips from Taiwan.

Speaker C: And unless they change and maybe they will, but unless they change their plans, the world’s most expensive chips and the world’s highest volume chips will continue to be produced in Taiwan.

Speaker B: So does this mean that Taiwan is both protected by, I think it’s what’s called a silicon shield, in other words, so vital no one can mess with them?

Speaker B: Or does it mean China, which could get feeling very inquisitive and say, well, we would like all of this to be in charge of all of this?

Speaker B: How do you read that?

Speaker C: So I think the Chinese leadership knows that if they were to attack, taiwan’s ship industry would be destroyed.


Speaker C: It’d be destroyed because the war would probably disable the facilities, because these plants need regular imports of materials and spare parts from Japan, the US.

Speaker C: And the Netherlands, because you shouldn’t assume that the personnel and the plants would stick around under Chinese occupation regime.

Speaker C: So I think it’s highly unlikely that China could grab Taiwan and inherit functioning plants.

Speaker C: But that doesn’t mean that I’m all that confident china is going to be deterred by the economic cost of attacking, because the reality is that history provides numerous examples of countries that have gone to war at disastrous economic cost for themselves and their adversaries.

Speaker B: And I’ve got to since over your right shoulder is a book on Stalin and the history of the Soviet I’ve got to get your I mean, when you say that, obviously Russia and Ukraine comes to mind, do you think that China, with respect to Taiwan, watches how the west has responded to the invasion of Ukraine and taken a lesson one way or the other about meddling with this place.

Speaker B: Because once you read your book and are kind of focused on the centrality of Taiwan with chips, it’s very hard for me to kind of not think about Taiwan all the time.

Speaker B: So what lessons do you think China, if any, is getting from what’s happening in Ukraine?

Speaker C: It’s a hard question to have a confident answer to it.

Speaker C: I think that the Chinese leadership is undoubtedly more skeptical of their military capabilities than they were a year ago, just because Russia’s military capabilities have so dramatically underperformed everyone’s expectations.


Speaker C: So that’s a positive if China feels less confident in its military, is less likely to attack.

Speaker C: But then I also look at the fact that the war has been going on for a year and Russia is still in the war.

Speaker C: It still controls territory on Net, and it probably will continue controlling more territory than it started fighting with.

Speaker C: I think that’s a lesson that’s much less optimistic from our perspective.

Speaker C: And a third lesson that I worry that China has concluded is that nuclear threats work.

Speaker C: And from the beginning, Putin has threatened nuclear war.

Speaker C: And from the beginning, Biden has understandably, in some ways said, we’re not going to intervene because we’re afraid of nuclear war.

Speaker C: And if you were to envision ramifications of that for a future Taiwan scenario, I think on day one of any sort of conflict between China and Taiwan, china would issue very dangerous nuclear threats and it would be tough for the US.

Speaker C: To respond.

Speaker B: Yeah, effective saber rattling.

Speaker B: So there’s the threat of taking what Taiwan has gotten.

Speaker B: But I was talking to Kevin Rudd, the former Prime Minister of Australia, and he had your view, too, which was basically you can’t just go in and take over what they do in Taiwan.

Speaker B: You can’t just turn on the lights and press a button.

Speaker B: What about China has been on a march, I guess, since 2015 to try to create its own domestic chip.

Speaker B: Correct me where I’m wrong here.

Speaker B: Production capacity, capability, how’s it going?

Speaker B: And how close is China to catching up to where they would want to be to have more independence from this incredibly interdependent system that you’ve described?


Speaker C: So I think China’s progress has been mixed over the past decade.

Speaker C: On the one hand, there’s a lot of different spheres in which China has substantially more capabilities than it did a decade ago in terms of chip manufacturing, in terms of chip design.

Speaker C: But it’s also the case that China’s industry remains really critically reliant on imported software tools and especially imported machine tools.

Speaker C: So the most advanced chipmaking facilities in China are reliant on the exact same companies to import tools from, asml in the Netherlands, a number of US firms, a couple of Japanese firms just like TSMC.

Speaker C: And so you can’t make an advanced chip in China without using machine tools that are imported from the US.

Speaker C: Or a couple of allies.

Speaker C: And that’s given the US.

Speaker C: A lot of ability to cut off Chinese firms from advancing or at least making it very, very difficult, very costly to make advances.

Speaker C: And so Chinese firms not only find themselves still a couple of generations behind what TSMC can do, and a couple of generations means a couple more laws behind, but they also find it increasingly difficult to make progress because the US.

Speaker C: Isn’t letting them get the machines that are needed to produce even more advanced ships.

Speaker B: And how do you evaluate what the Biden administration has done just on that last point you made in terms of choking the ability of China to meet its ambitions in creating chips?

Speaker C: I think the Biden administration’s rules have made it a lot more difficult because Chinese firms were relying on US.

Speaker C: Providers of machines to make their chips.


Speaker C: And as of last October, when the Biden administration tightened the rules, now there’s a whole suite of cutting edge machines that Chinese firms simply can’t import, not only from the US.

Speaker C: But also there’s been a deal between the US.

Speaker C: Japan, and the Netherlands to cut off chipmaking machines from all three of those countries.

Speaker B: Let’s finally talk about Asml, so that’s in the Netherlands.

Speaker B: Explain euv and why what Asml does is so important and why the Biden administration would want to get in the way of letting that get to China, that technology.

Speaker C: So, the way you produce a chip with 15 billion tiny transistors carved into it, each one the size of a virus, is by a process called lithography, which means shooting light through a mask that has the pattern you want to project onto the chip.

Speaker C: Then the light goes through the mask and reacts with chemicals on the silicon.

Speaker C: And where light hits the chemicals, they react in a certain way.

Speaker C: Where light doesn’t hit, you don’t get that reaction, and that’ll start creating the patterns on the chip that will eventually produce the 15 billion transistors your iPhone requires.

Speaker C: This process of shooting light is called lithography.

Speaker C: And in the early days of the industry, we used visible light for lithography.

Speaker C: But visible light has a wavelength of several hundred nanometers, which is now far, far too large for the manufacturing we need.

Speaker C: And so today, the most advanced tools have extreme ultraviolet light, which is a type of light close to an X ray in its wavelength.


Speaker C: So the specific wavelength is 13.5.

Speaker C: So the challenge is to produce light with this exact wavelength at sufficient volumes to manufacture chips.

Speaker C: And so the way you do it is you have a ball of tin 30 millionths of a meter wide falling through a vacuum.

Speaker C: You pulverize it twice with one of the most powerful lasers ever deployed in a commercial device.

Speaker C: It explodes into a plasma measuring 40 times hotter than the surface of the sun.

Speaker C: The plasma emits this light, which is then collected by the flattest mirrors humans have ever made and directed towards the chip.

Speaker C: And so this is hard to do.

Speaker C: Costs $150,000,000 per machine.

Speaker C: And today, the production of these machines is monopolized by just a single company.

Speaker B: And what you describe so well is the complexity of the machine, the fact that it has to be staffed by people around who know about that complexity, and all the parts preparing for breakdown, depending on how long it’s been running, and that it’s really a supply chain talent and task.

Speaker B: I mean, what you just described is an amazing, wondrous, kind of almost supernatural event, but it requires, like, those mirrors, for example.

Speaker B: I mean, it requires understanding how to get all of the component parts, I think.

Speaker B: Is it true that I think you wrote that this is basically the most complex factory or manufacturing process on the planet.

Speaker C: Yeah, that’s right.

Speaker C: There are hundreds of thousands of components in each of these systems.

Speaker C: The only thing that really comes close is a large aircraft in terms of the number of components.


Speaker C: Like, a 747 will have a roughly comparable number of components.

Speaker C: But the margin for error in UV systems is even smaller than the margin of error in aircraft.

Speaker C: We’re talking about nanometer levels of precision.

Speaker C: And one of the things that really struck me in learning about these systems was that if you’ve got a machine with several hundred thousand components, each of those components can basically never break.

Speaker C: Because if your average time to failure is one year per component, your machine never works right.

Speaker B: And debugging must be a h*** of a challenge.

Speaker B: When I read about the threat from China or what worries people, and the reason that the Biden administration is blocking this technology and doing everything it can is artificial intelligence is often brought up as the great.

Speaker B: What do people mean when they say that, that if China has the computing power it needs, it can use it in artificial intelligence?

Speaker B: What will it then do?

Speaker C: Or what’s the fear to train an AI system?

Speaker C: Whether it’s training Chat GPT to recognize patterns and text across the Internet, or training a computer vision algorithm to differentiate between a cat and a dog, you’ve got to show it a ton of data, a massive amount of data.

Speaker C: And to do this efficiently or to do it at all, really requires advanced chips that can actually crunch all the data.

Speaker C: And so there’s really only a small number of data centers in the world, hundreds of data centers, but not more than that, that are big enough to train advanced AI.

Speaker C: Systems.


Speaker C: And all of them have just a small number of chips inside, almost all at the cutting edge, produced by just a couple of companies.

Speaker C: And that’s because training AI is so data intensive.

Speaker C: And so when the US.

Speaker C: Government looks at AI in the future, there’s obviously any number of civilian applications, but there’s also a lot of military and intelligence applications, too.

Speaker C: And just as we’re learning to train cars to drive more effectively in autonomous mode, we’re also learning to train drones to fly more effectively autonomously.

Speaker C: And defense ministries around the world are increasingly thinking about how do you apply autonomy to military systems.

Speaker B: A couple of questions about the writing process.

Speaker B: How long did this book, how long did it take?

Speaker C: I started the research around 2015.

Speaker C: I started with no semiconductor expertise, so it took a long time to get up to speed with the technology, with the structure of the industry.

Speaker C: I probably started writing in earnest in 2020.

Speaker C: So only a couple of years to write, but a long time to research.

Speaker B: Yeah.

Speaker B: And when you were neck deep in the cave of writing and researching, did you start to have a semiconductor view of the.

Speaker B: World.

Speaker B: I don’t know.

Speaker B: If you’re writing about war, you see everything as conflict in the writing process.

Speaker B: It can sometimes shape the brain.

Speaker B: Did you have any of that while you were writing about seeing semiconductors just in everything?

Speaker C: Well, I did, but I came to realize that that’s actually how we should see the world because semiconductors are in everything.


Speaker C: And I found myself walking around my house from the moment I woke up and realizing that everything I touched had chips inside.

Speaker B: Yeah.

Speaker B: And what innovation did Apple make?

Speaker B: Because I think people intel and PCs and that sort of lane of the story.

Speaker B: But briefly describe the smart move that Apple made and where it fits in this vulnerability and concentration of the chip manufacturing world today.

Speaker C: So Steve Jobs had a great quote from the 1980s, before he was really famous, and he asked, what is software?

Speaker C: And his answer was, Software is something you didn’t have time to put into hardware.

Speaker C: And I thought that was actually really profound.

Speaker C: And Apple has stuck with that mentality and tried to turn everything it can into hardware, which is why it makes such beautiful hardware products.

Speaker C: And that’s meant investing in its own chip design from the very early days.

Speaker C: And so, unlike most smartphones, Apple phones have chips inside that are largely designed by Apple.

Speaker C: And that’s made Apple one of the world’s largest buyers of chips.

Speaker C: Around a quarter of TSMC’s revenue comes from Apple in many years.

Speaker C: And so Tim Cook, the CEO of Apple, has as much influence over the semiconductor supply chain as almost anyone because it’s dollars from his iPhone orders that are shaping how the industry is structured.

Speaker B: And as I understand that, he came up through I mean, he used to do supply chain before he was the top dog.

Speaker C: That’s right.

Speaker B: The final question is the way you’ve laid out the book is very pleasing.

Speaker B: Not only is it stories, but the chapters are not long, winding roads.

Speaker B: They are relatively short, as these things go.

Speaker B: They’re grouped in, I guess it’s eight parts.

Speaker B: Yeah, eight parts.

Speaker B: Describe the intention behind that.

Speaker B: And when did you do you always know you wanted to write it that way?

Speaker B: Was that something that came to you?

Speaker B: How did you design the book itself?

Speaker C: Well, I wanted to make the characters come through.

Speaker C: And there were so many fascinating individuals, bombastic personalities, wild gamblers that I came upon as writing that I thought that shorter chapters would let me introduce more of these characters and have them appear in more different places.

Speaker C: And so one of the fun things of the book was that I spent a lot of time learning about the technology and thinking about business models, but also got to talk to a lot of these individuals and understand how they got involved, the struggles they faced, the battles they fought.

Speaker C: And that was really the fun part of writing, was telling their stories.

Speaker B: And it was the fun part of reading.

Speaker B: Chris Miller.

Speaker B: The book is Chip war.

Speaker B: Thank you so much for your time and for writing the book.

Speaker C: Thank you for having me.

Speaker A: That’s it for this month’s edition of Gabfest Reads.

Speaker A: Our producer is Shayna Roth.

Speaker A: Ben Richmond is senior director of operations of Podcasts.

Speaker A: Alicia Montgomery is vice president of audio at Slate.

Speaker A: We’ll be back next month with another edition of Gabfest Reads.

Speaker A: Until then, all three of us will be back in your feed on Thursday with a new episode of The Sweet Political Gabfest.