We chat with Alex Williams and Hassan Khan, a pair of techie policy wonks who have done deep research on the past and present of the semi-conductor industry. The duo also also brings some strong opinions on the current supply chain crunch and what it would take to build a more resilient and adaptive semiconductor industry in the future.
You can find Alex's writing for Employ America here. You can find him on Twitter here
You can find Hassan's blog here and his Twitter here.
You can find their writing on the semiconductor industry and shortages here and here.
Our lifeboat badge winner of the week is jasme, who helped someone figure out how to fix email validation with Laravel.
Hassan Khan Beginning in the early 2000s, Moore's law essentially slowed down. So the way we made chips got smaller, got harder. And the way firms got around the physical limits they were hitting, was they started to introduce new materials. And as a result of that, you know, cost went up, you had to do a lot more material science and physics research. And that's also driving, sort of, the new complexity in semiconductor manufacturing today.
[intro music]
Ben Popper Couchbase Capella database as a service is flexible, full featured, and fully managed with built-in access, key-value SQL, full-text search. It’s blazing fast yet surprisingly affordable. Visit couchbase.com/stackoverflow to try Capella today for free. No credit card required.
BP Hello everybody, welcome back to the Stack Overflow Podcast, a place to talk about all things software and technology. I am Ben Popper, the Director of Content here at Stack Overflow. And today we're going to be talking about something that all of you have to deal with probably in your daily lives in one way or another, which is semiconductors. And we're gonna be talking about the semiconductor shortage, which is global, ongoing and related to a lot of the supply chain issues you might have found in your own life. I saw a very funny tweet, or no, it was in one of our Slacks, somebody was asking me if it's normal for a graphics card today to cost as much as the computer they bought two years ago. So I thought, this is the kind of thing that developers who are listening to this can probably relate to, in one way or another. We have two great guests on the show today, Hassan Khan and Alex Williams. I'm gonna let them both introduce themselves. And they'll tell us a little bit about the semiconductor shortage and what makes them such experts on it. Let's put it that way. Hassan, Alex, welcome to the show.
HK Thanks for having us, Ben.
Alex Wiliams How's it going?
BP Good. So Hassan, let's start with you. Who are you? What do you do? And what does it have to do with the semiconductor shortage?
So I currently work in product operations in the technology sector. And then prior to joining my current employer, I actually did my doctoral work at Carnegie Mellon on the semiconductor industry at the end of Moore's law, where my focus was on r&d policy. And that's where I've been involved with the semiconductor industry.
AW And as for myself, I'm Alex Williams, I am a research analyst at Employ America. As far as the semiconductor industry is concerned, I actually was not that much of a specialist in it until doing this research series with Hassan. I am a macro economist by training. So a lot of my insights are sort of into this question of, you know, who is producing what, where, why? What does that do for prices? Who's working on it? Yeah, I do policy work.
BP This can go out to either one of you, whoever prefers to take it, or you can split track it. But let's give people like a step back here for a second, you mentioned Moore's law. Can you give me a capsule history of semiconductors and the semiconductor industry? Like, when did these things become, you know, a critical component that we were relying on for the tech industry? And how do they turn to this thing that we're relying on for everything from my car to my toaster, and down the line?
HK Yeah, so I guess a lot of folks would put the genesis of the modern semiconductor industry back to 1965, when Gordon Moore, then a Fairchild and of fame with Intel, published his first paper on the future of integrated circuits. And you know, since that time, what we've seen is the ability of semiconductor firms to build more advanced chips has continued on an exponential scale, that was essentially his observation. And because it's continued pretty consistently over that time, it's also made more computational power cheaper. And as computational power got cheaper, the thing is, humans found more and more uses for it. So at first, it was so expensive, you would only put it on a missile that you specifically designed for the first uses of integrated circuits. And the DoD played a huge role in that development. And now, you know, computational power from integrated circuits is so cheap that we'll put it in your toaster, in the key fob for your car. And as a result, when we have a surplus of demand, and a widespread usage of these components, we get hit with shortages that impact, you know, every facet of modern life.
BP And Alex, what about you? What's your perspective on the industry? And if you had to give us like, a little bit of a capsule history that follows that, you know, where are semiconductors sort of principally made these days? And why are we experiencing this pinch? Is it solely pandemic related? Or does it have multiple sort of causes and effects?
AW So I would say it is solely pandemic related, but not necessarily in the way that you might think. One of the big points that we make in the first piece in the research areas that we put out that, you know, Skanda Amarnath, who's also at Employ America, co-authored is basically that, you know, the pandemic saw this big surge in demand for semiconductors at the same time as it saw sort of smaller, you know, but still reasonable disruptions in supply. So it's a pandemic story, insofar as you have this rotation and consumer demand towards buying things to, you know, make it enjoyable to stay at home 24/7.
BP Goods and services.
AW Yeah, but the thing that is tricky about that is part of the reason that it created so much of a pinch is that following the 2001 .com crash, there was a huge sort of collapse in capex in semiconductors in the US. And at that point, you know, because the US sort of declined to provide enough, you know, fiscal support to head off that recession, the industry in terms of CapEx domestically never really recovered.
BP So you're saying blame the Fed? Because I do that everyday.
AW No, not the Fed. No, this is fiscal policy. Yeah, no. So at this point, a lot of it is a really complex global value chain, like anything that you look at will pass through, you know, 30 or 40 countries, you know, on the on the path from being sand to being you know, a computer.
HK If I were to build on what Alex was saying, I think the acute reasons that we're having these shortages today are due to the pandemic, right, because when the pandemic first happened, no one really knew how it would play out, everyone thought we were headed for a massive recession. Instead, because of fiscal policy interventions, we avoided a recession, and people had a ton of money to spend on goods. And they decided they wanted to buy new laptops, and they needed webcams at home, and so on, and so forth. So there was a massive shift in demand. But then the shortages really became politically salient when they started to hit sectors like automobiles. And there the story is not just one of like whiplash and demand. It's actually the semiconductor firms whose supply automobile producers said, well, you canceled your orders. And here, the reason is automobile firms made this the classic recession scorecard, which is okay, there's a massive recession coming, we're not going to have orders for new cars. So we're going to cancel all of our orders and just whittle down our inventory. Well, turns out, we're driving a lot more places. Also, because of fiscal policy, people do have money on hand, and they wanted to buy new cars. And so then automobile firms went back to semiconductor manufacturers and said, hey, remember how we wanted all those chips? Turns out, we still want them.
BP Right, actually, we want more than we wanted before.
HK We now want more than we wanted before. Yes. And to the point that you're making earlier, now cars have higher semiconductor content than they ever did in the past, because new models, including electric vehicles, and like new autonomy features, like you know, the supercruise, or whatever Tesla's pumping out, those have new chips, right. But now you get into the point about you have these massively global supply chains where you can't just turn off, okay, I'm making chips for key fobs, or I'm making chips for your toaster. And now I'm going to flip my line over and make chips for cars, it actually takes like six weeks for a wafer to work through a specific semiconductor fab. And then that wafer itself has to be sent to what's called outsourced assembly and test. So they'll send it to like, usually the the chips will be manufactured in, say, China or Taiwan, the wafer, sorry, then the wafers will get sent to outsource assembly and test those, those are much more labor intensive and usually happen in Malaysia or Vietnam. Then those chips will be sent to get integrated into the PCB, or the final electronic assembly either at another vendor or ship straight to the auto manufacturer. So you basically, you stop that whole supply chain, and you said, hey, let's start it back up. There's lead and lag time and getting that going. And then you factor in, well, because of the pandemic, our shipping lanes are backed up, as everyone is aware of.
BP Port of LA is very constipated.
HK Right, exactly. Now, that doesn't necessarily impact semiconductors, but it might because they may have raw materials that are backed up, or the other one that disrupted it is a lot of transit happens via the air. And because no one's flying anywhere, there's fewer air cargo holes available to ship things to. So it's like layers of complexity that go down the supply chain that make this one so hard to detangle.
BP So let's back up for a second. One thing I'm pretty interested in is right, the idea that maybe you know where a fab is, and maybe there's some shipping lanes, but actually, you're going to need a bunch of sand and a bunch of chemicals and a bunch of rare earth minerals to even get to that stage. So in a few weeks, I won't be able to buy blue paint from my house, I'll have to painted a different color because they're missing one out of you know, 16 chemical components. What goes into a semiconductor like Alex, I think you're saying there's there might be 30 or 40 different things you have to source. You know, just to make this thing, forget the manufacturing logistics, like before you can manufacture you need X, what are the things that go into a semiconductor?
HK So the obvious answer is sand. Right? It's silicon. And what's interesting is for all the talk of American supply chains have been offshored. One of the main sources of the exact type of quartz that you need for this—
AW It's about sand, baby!
HK Well, sand we have a tons of but you need like a very specific grade. And my recollection is that it comes from like North Carolina, there's like a quarry. But the semiconductor industry loves to brag about how—don't quote me on this number—but there's like almost every element in the periodic table is in a modern semiconductor, like it's a massive number that I that I should know, off top my head that's really in part exploded in recent years because as beginning in the early 2000s, Moore's law essentially slowed down. So the way we made chips smaller, got harder, and the way firms got around the physical limits they were hitting was they started to introduce new materials. And as a result of that, you know, cost went up, you had to do a lot more material science and physics research. And that's also driving sort of the new complexity in semiconductor manufacturing.
BP Is the best solution is for all of us as a shift to quantum computing tomorrow? Could we just like get off—forget semiconductors and just go a different way? Whatever those freeze cool light chandelier things are? Can we just do that stuff?
HK So I think now, with the caveat here being I'm not super familiar with, like all the recent advances in quantum computing, I think even the biggest boasters of quantum computing would tell you, it's applications for the next decade are very limited, they're gonna be able to solve some very nifty problems that you can't solve with conventional computing. But it's a long time from you or I carrying a quantum computer in our pockets or our backpack.
BP Quantum toasters.
HK Yeah, yeah. There's no quantum toaster. There's no quantum toaster. So that sounds pretty cool, that's pretty like Rick and Morty. But I don't think it's in our—maybe our kids have quantum toasters.
BP Okay, cool.
HK But probably not us.
BP So Alex, I know we mentioned this earlier. But you both have been working on research. And there's this organization that's sort of focused on the supply chain in the United States. Can you tell me a little bit about that side of it? Like, where does your work intersect with fiscal policy or National Industrial sort of scientific policy? And I guess, to the degree that it makes sense, like, what have you been advocating for? Like, are you interested in seeing more of this production brought back, you know, onshore, to the US, for example?
AW Sure. Yeah. So I think a major thing that animated you know, our starting this research project and citing it within Employ America as a—I mean, Employ America is a think tank whose mission is full employment and tighter labor markets. Fundamentally, it's a labor market think tank, which makes it seem a little bit odd that we would get really interested in what exactly is happening in these supply chains.
BP Whole entire labor market? This is a golden era for you. We're there. We got it.
AW I mean, we're still—there's so many metrics that have like a ways to go. We've been very pleased with the strength of the recovery so far. And we've been pleased with the commitment to fiscal policy so far, but the idea behind getting interested in these supply chains was really sort of twofold, which is that, you know, on the one hand, in order to maintain like a full employment, high output, high pressure economy, you kind of have to plan a little bit what it is that you're going to put out. If you're going to have everybody employed, you need to use a little bit of industrial policy techniques in order to make sure that what's getting produced is going to be, you know, actually useful. And that, you know, by doing that you can kind of head off riskier inflationary.
BP When you say planning, I think central planning, and that's a dirty word in the US. Right?
AW Right, which is why I'm saying, you know, this is sort of an industrial policy thing, because a big part of this is I mean, we don't even really have sufficient data collected in the public domain to look at how any of these supply chains really work. I mean, we have the input output tables that are put out every five years by national income and product accounts, like we have these two sources of data. Neither of these tell us all that much about all of these intermediate steps or give us an ability to assess the degree of resiliency of sort of critical economic supply chains.
HK I think Alex is maybe too nice to say this. But the truth is, we do have an industrial policy of sorts in the United States. It's just that it's made ad hoc by lobbyists of firms who get tax giveaways. You can open up the newspaper almost every day and see, you know, some major firm has received billions in subsidies to build a new manufacturing plant.
BP Sure. Airline or car companies or—
HK Across all industries, right. I mean, if we focus just on semiconductors, the chips act is currently in Congress, it has $52 billion in funding, 39 billion of which is allocated for subsidizing semiconductor manufacturing facilities. And if you look at some analysis out there, I think over 18 billion of that is already spoken for by plants that have been announced by firms like TSMC, Intel and Samsung. That ends up being our industrial policy, it's just that it happens kind of at the state and local level. And then it trickles up to the federal level every so often when there's kind of a panic.
BP Yeah, we're still waiting on that beautiful Foxconn factory in Wisconsin to start producing.
HK I mean, I think that's a classic example of, you know, the state of Wisconsin, and the President Trump boasted about what that would bring. And we invested millions of taxpayer dollars and got nothing to show for it in return. And I think what Alex and the folks at Employ America and myself have been advocating for, is you can do this with a little bit of upfront thought and planning, and it doesn't have to be this like giveaway in a panic situation at the end. Right? And so I think the money gets spent, it's a question of, like, are we ensuring a return on investment? Or are we just, you know, giving money away to the person who speak loudest?
AW It's also if you imagine, you know, the the government as having the goal of having sort of resilient supply chains for key, you know, even national defense reasons. You want to go into the negotiations with these private companies, knowing what the state of play in the industry is, if your lobbyists are informing, you know, what it is that should be invested in as opposed to something more coherent and intentional on the side of the government, it's, it's a lot easier to get sort of rolled into a policy that is designed not with the interest of the American people, but with the interests of Intel.
BP So let me ask two questions. One would be, you know, let me sort of put this into two frames of mind we'll discuss one would be sort of like, okay, what's your vision, like, lay out the vision for how we solve this and how we think they picture in long term? And what kind of policies could be put in place? But another question would be, has anybody ever done it? Like, is there some version of this elsewhere in the world or in our past in the United States where we're doing it right? Like, in the 1950s, 60s, 70s, if we were building semiconductors, and then the internet and DARPA was running full steam? Were we doing central planning and scientific and industrial investment in a good way? Or was K Street and lobbyists just like, they weren't as powerful back then and so we could make better decisions? Like, point me in the right direction? Like, where would you like to see us go?
AW I think central planning is a strange thing to counterpose here, because I think what really is important, and what really characterized earlier eras of fiscal intervention into the space is an attempt to provide funding that opened a wider distribution of outcomes. So if you think of central planning is saying, oh, we're going to target one particular point and that is the point output that is going to happen, you're going to do thing X, scientific discovery doesn't work that way. But if you offer, you know, funding that says something in this side of the distribution is going to happen, we don't know who it's going to come from, or what form exactly it's going to take. But we're going to incentivize research in this way. And we're going to have people build out the capital stock to produce these things. Because in so doing, they'll be able to capture process improvements. You don't know exactly what it is that's going to come through, but you pursue a strategy that sort of actively tries to create and capitalize on opportunities. I'll let Hassan speak a little bit more about the American experience of this. But we even see this to a certain extent with the experience of Taiwan Semiconductor, which were was founded by an American who spent years working at Texas Instruments. And you know, in collaboration with the Taiwanese government, more or less set up a domestic chip sector for them.
HK I think that's a question of like, what would a good setup look like, a framework that I think is helpful is in the US, we have a very competitive and successful scientific funding enterprise between organizations like NIH, NSF, and DARPA. If you look, we want to invest in certain areas, we want to invest in certain research fields, and we want to make sure that we're making progress—
BP Not gain a function, but all the others, yes.
HK Sure. I mean, we can nitpick on specific things. But you know, by and large, they make sure that our research ecosystem is seated, and has funding available for, you know, critical areas. And then obviously, the military has its own funding agency of research funding agencies, and they help to see that as well. But if you go to a more industrial focus, we don't really have, like agencies that kind of help provide guidance on like, where do we want to go? So if you go back to the science example, the NSF and NIH will say, hey, we want to see progress in XYZ field, because we want to achieve these—
BP Vaccines, cancer research, Alzheimer's, show us something good.
HK DARPA comes out with programs saying, hey, we want to be able to improve capabilities for specific military use cases, XYZ. The difference is that we don't have a similar type of function at the industrial level. And to Alex's point, we don't even know where the gaps are. So I might say, as a policymaker, what I want is I would like our domestic semiconductor or domestic automobile industry to be competitive in EVs or next generation semiconductors. What might it take to do that? We don't have that view internally. What are our goals? Instead, what we have is, hey, we're not making enough EVs in the US. Let's give $5 billion to Ford to build a plant in Tennessee.
BP I can think of one example, I think where it rolled down. And I don't know if this was intentional, but DARPA had to like autonomous vehicle challenge. And they ran that and ended up leading to like, Lidar and Evie and autonomous vehicles in the US. And I forget—Velodyne. I think that was the name of the company, the guy had made speakers, he entered the challenge, because that was fun. And now he has a LiDAR company, right. So that's like an example of how you would like it to work, right? In that case, they weren't specifically saying why they wanted autonomous vehicles, maybe it was industrial, maybe it was military, maybe scientific, but like, they set out sort of an open ended challenge that they helped invest in. The person who won that, you know, was able to either capture funding from them or general attention. And that led to not creative destruction, that led to some innovation and, you know, catalyzed sort of some some industry creation in the US in a good way.
HK I think that's an example of, we have the scientific apparatus, right between DOD, NSF, NIH, we can lay the seeds for new science and new technologies, but to get the industry really going and and to make sure that the industry serves our policy goals in the US and those policy goals could be Hey, we want a resilient domestic supply. We want to make sure that our defense agencies have sufficient supply for critical applications. We have labor goals, we want to have, you know, manufacturing and service facilities in the United States that provide good high paying jobs for skilled labor and help develop that ecosystem. So we're really good at saying we want new technologies to exist, we're really bad at saying, and then we want industries to exist that address these goals. And in fact, I think one of the points that Alex and I have made in some of the pieces we've written is, I don't know that policymakers have really articulated concretely what those goals should be from an industrial policy standpoint, which makes it I think, more nebulous and easier to call central planning.
AW There's also sort of a well understood problem in a lot of manufacturing policy have something that is referred to as the valley of death, right. So you have your entrepreneur who wins this competition and, you know, incidentally, develops LiDAR, right, they own a bunch of intellectual property about how to implement LiDAR, they are sort of at the forefront of this, once you have that, the thing that you have to fund is you have to fund the build out of CapEx in order to manufacture that stuff at scale. When something is a newer technology, it remains risky, people are very happy to fund intellectual property, because you can continue to litigate that forever and draw value forever. And it doesn't take up very much space on a balance sheet for a firm referred investor. But if you need $800 million to build out to, you know, plants that you're trying to raise, as someone who says, you know, I have this new technology that sort of proven that the commercial applications are sort of there—
BP That's what venture capitalists for. I mean, isn't that what Silicon Valley does best?
AW There's a gap between VC and general commercialization for things that are capital intensive, which is very difficult. It's very easy to use VC when what you're doing is, you know, sort of designing a website, designing network architecture for a Snapchat or for Facebook, where your capital expenditures is something like servers. If you are, you know, building entire full size factories that require you to commit to a 10 year payback schedule, you know, on investment and physical capacity, it's a lot harder to get VCs excited about something like that.
HK I think a great example here on the symbiosis between private capital and government capital, actually, is electric vehicles and Tesla. So the early history of Tesla is they actually utilize the DOE loan guarantee program for some of their CapEx.
BP They've been farming—they still farm tax credits, right? I mean, they still often that when they're profitable, it's because they got these energy efficiency credits.
HK Exactly. They basically trade their energy efficiency credits to firms. But today VCs will take the risk on building a new EV company, right? So Rivian and other, you know, new firms have VC money. But back in the day, no one was willing to take that risk. But the DOE he said, hey, we'll give you this loan that we're guaranteeing because this is important for decarbonisation, and there that's like—
BP That's environmental policy.
HK The government has different policy goals than like VCs where they're focused on returns. And that helped to stimulate an industry. And you know, the US is better off for it, because we have Tesla here and Rivian. And Ford is building an EV plant in Tennessee. And we may have a very competitive EV sector.
BP Tesla's, you know, yeah, the most well known and probably popular EV brand in the world, and they're very popular in China and other ones and also other places. I guess one question that comes to mind is how to avoid I don't know, regulatory captures the right way. But you know, I read all kinds of horror stories about your mentioning the military, you know, they say our goals are to create XYZ, and then they spend a trillion dollars on like a plane that never flies, and they invest in so much, you know, cutting edge technology, supposedly, but like, it's just a total boondoggle for the taxpayer. So how do you as you as you sort of set it out, you know, create the incentives to try new things at not just a scientific level, but an industrial level, but ensure that there's, you know, the right incentives and the competition in the marketplace is at play, so that you don't just have, you know, a few defense contractors, essentially, hiring folks out of government and then capturing a trillion dollars for an airplane that never flies? The cutting edge, cutting edge technology that we need from a national standpoint, but which actually, we don't even get for the trillion dollars we spend.
HK I don't want to speak for Alex. But I think we're all very aligned on we do not want to see that sort of industrial capture of government spending. And I think, actually, the semiconductor industry's history has a great point on we knew how to do this. Well, if you look at early examples of chips that were developed for the military, like for the Minuteman missile, the government through the DoD actually had second sourcing requirements. So they would have cost plus contracts with manufacturers saying, hey, we'll pay you whatever it costs to produce it plus an additional fee so that you have a guaranteed profit so that you'll take the risk of developing this novel technology for us, but we'll also have a second source requirement in there. So we'll need you, we'll require you to make sure that there's another vendor capable of producing this chip. Why because the military understood what resiliency means. That means you can rely on a single supplier, right? Those are like procurement 101 practices that you would use, right? I would argue that the current state that we're in, when we don't have a proactive policy, where we don't have policymakers saying, like, we need these things and planning it out, it actually lends itself to the policymaking by emergency where firms come and say, oh, we noticed you need this, guess what, we're the only ones who have it and you enter into bad contracts. And I know Alex is waiting to jump in.
AW Yeah, it's also very uniquely linked to where we are at this point now, because if you—in terms of shortages, because part of the problem for a lot of these especially lagging edge chips, is that these are commodity products, you don't, you know, command a lot of profit by building the micro controller that moves the driver's seat forwards and backwards in your new car. Which, you know, in a system that is doing sort of just in time manufacturing, and you know, everyone is managing their global supply chains, the needs of efficiency eventually whittle you down to a single supplier for that. And if you, you know, say cancel your orders with them, or say, as was the example, with multiple Renaissance facilities in Japan, there's a fire or an earthquake, then everything that's downstream of that only has that one source to go to, and then is disrupted for reasons that are, you know, sort of completely unrelated. And also, you know, can sort of have these knock on effects. For example, I mean, people are extremely concerned about inflation. But then when you look in the inflation indices, substantial chunks of this are movements in the price of vehicles and used vehicles over the last like handful of months.
BP Right, you're saying we don't have hyperinflation? [Alex laughs]
AW Oh, man.
BP You know what that really means? That's like, you know, the sandwich in Vienna cost 4000 times what it cost last week.
AW Yeah, hyperinflation is the kind of thing you get when you've you've been bombed flat.
BP Yeah, exactly.
AW But yeah, so you wind up with these situations where, you know, certain economic decisions within very opaque and relatively profit oriented, efficiency oriented systems can create these sort of external narratives that stymie you know, other important programs simply by virtue of not managing the economic system well.
BP You know, this is obviously we're having a discussion about a global supply chain and global competition and national policy. To what degree do you think we can accomplish good things nationally in the US to make our supply chain more robust? Without it being a negative for our competitors internationally? Like, to what degree can you agree that we should build more chip plants here, and that we should be investing from our, you know, our own government to create new industries here? And maybe that's still a net positive for our competitors in China or Japan or Germany? Is that possible? Or is it you know, in a certain sense, we become, you know, this global game of like, who's going to be the best at this? And who's going to win the most business? And who's going to be the leader in technology?
AW Yeah, I think it's a little bit tricky when you start thinking about trade problems, because you, you know, if you take these seriously, for long enough, you start to ask the question is what is important in the current arrangement of global trade, the demand side or the supply side? Because you can kind of imagine this picture of the world today, as you know, the US is a giant demand sink for everybody. Like and you know, people in the US get very het up about this, they're, oh, we have a big trade deficit, oh, we do all this other stuff. Without the US pursuing policies that made that possible, namely, you know, actually running some fiscal policy during the pandemic and looking to, you know, sort of provide a demand side stimulus for sort of the world over, you wind up in the situation with, you know, China and Germany, both looking to do export oriented sort of produce, which doesn't make sense unless you have someone who is there to provide the demand for them. So insofar as there is a trade story about, you know, competition as part of a better industrial policy proposal, the ideal outcome would really be to not only increase resiliency by sort of spreading the individual number of facilities that can produce particular chips, you know, sort of more widely, so that they're less subject to geopolitical risk and such, but that insofar as you know, one country is pursuing a specifically export oriented policy, that isn't necessarily and the efficiency gains from that are not necessarily a net positive for everyone else. So the idea would be, you know, we want to run an industrial policy and also encourage these other countries to run stronger demand side policies on their own front, things like redistribution, and providing a better social safety net, and things like that. So it's a little tricky when you start thinking about these trade questions in sort of a macro policy sense.
HK One other comment I'd add is, I hear sort of Ben, you're concerned that there's like a, there's a zero sum element to some of this competition, right? Like, if everyone in the world has been building semiconductor factories, suddenly, we're going to have too many semiconductor factories and like, who's actually going to make money?
BP I worked for a year and a half for DJI and at that time, I was acutely aware of like what you know, the government's power was in terms of its plans, its ambitious 2050 plans, you know, the amount of money that they were throwing, as you said, in a way you would like to see behind the development of new industry and encouraging internal competition at a national level to decide, you know, where the investment should flow down the line. But right, if China does that, and if US does that, and if India does that, does some other area that we could be investing in lose out? And do we all just end up in this endless arms race, you know, like to get to the peak?
HK I think semiconductors are acutely an area where the global geopolitical arms race has become so heated that the dynamics are a little bit wonky. But one comment I would add is, I think there's real value in having domestic manufacturing and domestic supply chains. Because when you do that, actually, research that I did with my advisor shows, when you have a facility located in a higher manufacturing cost area, it actually changes the nexus of innovation. And by that what I mean is, when you move it to a lower cost, country where labor is cheaper, you tend to use more labor intensive processes, and you don't invest in like the latest technology, because labor is cheap. And we can do it with more people. There's actually a lot of value from pushing the technological frontier and maintaining a domestic manufacturing base, because then firms have to like work within the constraints of manufacturing locally, to maintain competitiveness. And that means they invest in new production techniques, they invest in new technologies, they invest in labor saving technology, and that has spillover effects for the global economy. Because it means like, now that you know, once I've invented it at Intel in the United States in California, where it's very expensive to manufacture, eventually that trickles through my supply network and through knowledge bases to the entire world, and actually help to create more wealth for everybody. So, you know, part of it is there is a zero sum competitive aspect to it. There isn't like unlimited demand for semiconductors eventually, like some of these fabs won't, you know, if we everyone builds a fab like some fabs will sit sit empty, but there is also a demand creation aspect to it, which is, as we continue to push the technological frontier, people will find new ways to utilize the technologies we're using. And it won't just be in your toaster, it'll be 50 years ago, no one thought we were going to put chips in a toaster, right? But because we push the frontier, we found new use cases, and the market continued to grow. And that employs more people. And, you know, benefits flow from that.
BP This is a fascinating conversation. I want to thank both of you for your time. Let's wrap up maybe at a slightly more sort of like micro human level. I can't go out and buy a new car or a used car I want to get is more expensive than the new car, but I'm having some semiconductor problems at home. What can I do? Can I build one myself? Can I get a Raspberry Pi breadboard? Can I go harvest them out of a, you know, an E-waste recycling plant? Like I need more semiconductors. I need to get creative? Where am I gonna find them?
HK Remember Radio Shack? It's funny because we used to have those and I'm just at the age where I'm old enough to remember going there as a kid with my dad for weekend projects and picking up electronic components. And I don't know where you do that. Yeah, I don't even know where you do that anymore. On a more serious note, I think the news that we've seen and you've seen it in earnings reports from you know, like auto manufacturers is that the shortage is hopefully improving. And I think Christmas season and the port backup at LA have made things very crazy. But once Christmas season passes, I think you'd start to see more normalcy return.
BP I hope you've gotten Christmas presents for all your kids, because otherwise if you haven't bought them now they're gonna realize, hey, it's too late.
HK You got to buy them gift cards and promised they get to January, February, right. But I do think that's when you'll start to see some normalcy. So hopefully you don't have any like immediate needs for a bespoke chip. And you can wait until after Christmas.
[music]
BP Alright, awesome. I am going to shout out the winner of a lifeboat badge. So at the end of every episode, we shout out the winner of a lifeboat badge on Stack Overflow. They came and found a question with a score of negative three or less they gave it an answer. They got up to a score of 20 more and now the question has a score of three or more. So they saved some knowledge from the dustbin of history. Thanks to jasme, who explained "Laravel Validation Email address." They create an edit form and Laraval but they have to validate that email. They're getting some errors on the server side. So if you've had similar problems, you can check out this question. It's got 35,000 views. We'll put it in the show notes. Maybe they can help you out. I am Ben Popper. I'm the Director of Content here at Stack Overflow. You can always find me on Twitter @BenPopper, email us podcast@stackoverflow.com. And if you liked the show, leave a rating and review. Really helps. Hassan, who are you and to the degree that you want to where can people find you on the internet?
HK My name is Hassan Khan. I work in Product Operations in technology and you can find me on Twitter and Instagram if you choose at my full name Hassan Khan.
AW And I'm Alex Williams. I'm a research analyst for Employ America and you can find me on my Twitter handle which is at vibe account @Vebaccount.
BP Alright, thanks for listening everybody. We will talk to you soon.
[outro music]