Sharp Tech with Ben Thompson - Making Advanced Microchips in Arizona and Imagining a Remote Work Future

Episode Date: December 13, 2022

Various questions after a new round of TSMC investments in the U.S., reactions to TSMC founder Morris Chang saying that “globalization is almost dead,” beginning the remote work conversation with Ben's memories of life at Automattic, and thoughts on the competing factors that make the work-from-home question so difficult.

Hello and welcome back to another episode of Sharp Tech. I'm Andrew Sharp and on the other line, Ben Thompson. Ben, how you doing? Doing okay. Do it okay. Two weeks to go until Christmas and powering through. Well, that's a perfect segue. Indeed, the holiday is right around the corner and as of this week, you can give the gift of Strateree. So look out for that. Do you want to tell the people anything more about the gift option they're going to have this week?

You just go to the link in your show notes. You can give a gift of Stratory Plus, which gives people full access to this podcast, access to all episodes of Sharp China, Dithering, and Strategery, both in written for more podcast form. You can, I mean, nothing better than giving people something that they'll actually use and make their lives better. Yeah. I'm imagining someone sending an email with one of those big red bows that Lexus has in all their commercials. That would be a good idea. You can go in, you can specify when it will be delivered. So you could put, you know, December 25th and then we'll send out an email on that day.

And, and yeah, it'll be a happy day for everyone. Well, listen, the holidays are not here yet. We're still in grind mode, hustle culture, hardcore employees only. And today we have a lot to cover. Is that a foreshadowing of some of the topics today? No, we're actually, we're zagging. We're not going to talk about Twitter and Musk today. I want to talk about TSM.

There is a hardcore angle to both these topics. But yes, there's plenty of discussion about Twitter, to say the least. For now, though, TSM and the future of remote work, those are two topics that could probably be their own standalone episodes. So we'll have to be efficient here. But let's start with TSMC because there was some news last week. President Biden appeared in Arizona to announce that TSMC is building a second factory in Arizona and will be increasing its investment in the U.S. from $12 billion to $40 billion.

And you wrote about the event for Stratory. We'll put a link to that analysis in the show notes. But before we get to why this news might be a little bit more complicated than it sounds, I want to start with the basics. Like for someone who's not really familiar with what's going on here, can you explain to people how this news fits into the larger Chips Act narrative? Well, I mean, it all depends how you look at it. By definition, anything chip related is going to fit into the Chips Act's narrative because, you know, why wouldn't it? The TSM announcement, so the original announcement was under President Trump. It was three or four years ago.

And at the time that TSMC announced their investment in Arizona, they bought enough land and the, like basically for multiple fab. So this actually, I think it's fair to assume won't be the last announcement about this location either. So that initial $12 billion announcement was pre-chips Act. The, you know, certainly, if you take Morris Chang at his word, who gave a very sort of blunt interview to the Brookings Institute last spring. There's another link we should put in the show notes because it's worth checking out. I also wrote about it, didn't check it at the time. It was pretty clear that, look, we're doing it for political reasons. The U.S. is basically demanded we build a foundry in the U.S.

So we're doing that. And that plant was at 12 billion one. It was 5 nanometer at the time it was announced. And this is important because 5 nanometer at the time of the announcement was cutting edge for TSM. But this wasn't a plant that wasn't going to open until I think December 20203 is the target date with real real scale production in 2024. by which point five nanometers would no longer be on the cutting edge. TSM is sort of ramping three nanometer production right now. It's sort of next step down the chain.

And so it's always been a case sort of from day one that what was being built was not necessarily cutting edge and was relatively small. I mean, just the, and I mean, those numbers are really big, but that's because building boundaries are really, really expensive. So in the broader Chips Act narrative, certainly it's more advantageous to build in the U.S. these days. I think some of the tax incentives are more compelling in some respects than even sort of direct subsidies. But this is something that does predate it. And you sort of back up big picture. TSM is in a really interesting position. We can get into their business generally, but just from a political perspective, obviously,

that's a core asset for Taiwan. It's a reason for the U.S. to defend Taiwan. And at the same time, it's also like a real sort of gift that Taiwan could give to its potential allies. So I don't think it's an accident. The two places where TSMC is now committed to building new foundries is number one in the U.S.

And number two in Japan, which would be their other and, you know, most important ally in a potential conflict with China. So there's politics all over this. So to the extent that you want to say, you know, how does this fit in the Chipsack's narrative? Well, from a political perspective, it's all politics. I think that's that's where you have to start.

Yeah. Well, and I think when I say Chipsack narrative, there are people who want to bring this industry on shore and think that that's essential to our national security going forward. And TSM, if you're really new to this, it's Taiwan. semiconductor manufacturing co. And they are essential to a bunch of different industries. And so I'm wondering if the goal is bringing semiconductor manufacturing on shore for national security reasons, does this Arizona factory actually get us closer to that goal?

I mean, obviously, they're bringing some of it on shore, but like, how should I view this if the long-term idea is to sort of disperse our reliance and bring, a lot of it to America. I mean, I think in the short to medium term, it's, it's probably a lot less meaningful than people want to think it is. And I kind of feel bad. I mean, I was fairly negative overall about this in my piece last week. And I kind of felt like it as being a wet blanket.

I mean, I think the most positive spin you can put on this is, look, it's a start. And there's going to be some sort of development of ideally an ecosystem, particularly in the Arizona area. area. There's going to be more employees in the U.S. that can operate these sorts of things. You know, the importance of TSM is that they are a foundry. They're a place you go when you have a chip design. You want to get the chip made. Whereas Intel, which is the sort of leading producer on the on the leading edge of U.S. companies, they've traditionally always just built their own chips. And so they have a lot. There are like all the most advanced factories in the U.S. are

specifically made to build Intel chips. Now, Intel's trying to shift its business to also be a foundry, to also sort of build chips on contract, but that's a difficult process for lots and lots of reasons. And so, yeah, this stuff is now going to be onshore. And, you know, that's, you know, all things consider it a good thing. What I'm skeptical of is if there were actually some sort of conflict, how quickly do these factories become worthless? Like, you know, if TSMC is heavily disres,

in Taiwan, which would happen in a conflict. Like, these aren't factories that are just sort of standalone sort of entities. That's what I was going to say. They're still dependent on Taiwan, right? Well, they're not just, they're dependent on an entire global supply chain. I mean, like, we've talked about ASML, like these high-end UV scanners and things on those lines. Those are piped into ASML.

And ASML is tuned in and running them. And there's, you know, a whole ecosystem that is very heavily integrated. And it's definitely an analogy for the question of decoupling with China and sort of in general. We've had 20 to 30 years building up capabilities and supply chains that are so deeply integrated and interconnected. It's going to be very hard to build something that doesn't have any sort of dependency on either China or Taiwan. And number two, it's very difficult to manufacture an economic incentive to do so absent a war. Like if we were truly cut off from Taiwan and cut off from China, we would have to figure out all these pieces that are impacted by that. But until that happens, this stuff is so complex.

And, you know, the, you know, not to get all like philosophical, but the invisible hand of the market guides all of these tiny decisions on the margins. and you're going to have a very difficult time unwinding those sort of decisions that have been made over 30 years sort of by Dictat or bypassing an act with some sort of incentives. And I think the real challenge with this is absent the geopolitical threat, there's just, this doesn't really make sense for anyone. And so, you know, is the threat specific enough and pressing enough and what needs to be done, discreet enough that we can address it. I don't know.

It's easy to sort of be skeptical about it. Well, yeah. And there's another aspect of all this that I find pretty interesting. And that's are we focusing on the right areas of manufacturing capacity? Like Lisa Sue, CEO of AMD, said, quote, the entire semiconductor ecosystem is ready to step up and work together. The industry has been through so much in the past few years. years, having more geographically diversified capacity is so important. At the end of the day,

what we want to do is ensure that our most important chips have a resilient supply chain. And what I'm wondering is, are these advanced chips the most important chips to our security? Like, they're talking about making chips for Apple and AMD and Nvidia and all of these cutting edge companies, but there are also what you call trailing edge chips that touch all sorts of other industries that are much broader than like the latest iPhone 14 max. And so I'm wondering how you would balance it if you were in charge. Well, there's a few things going on here. So number one is what I think the this new, so this new fab, the new increase was for a three nanometer fab, which again, three nanometer production is starting in Taiwan now, or it's happening

now in Taiwan. So when this new fab opens in 2025 or 2026 or whatever it is, again, it's going to be at least sort of a generation behind. So just to put, be clear about sort of what's going on here, my suspicion is, like this is, it gets so complicated, it's very easy to sort of deep into a wormhole. But one of the big advantages that TSM has relative to, say, in Intel and move with this found remodel is TSM is built to be very, very flexible and sort of responsive, both to customers, new equipment, et cetera. And one of the reasons that, you know, because they, they're contract manufacturer, they're not designing the chips. They're using other people to design. Obviously, they have a huge sort of tool set and toolbox of, you know, different IP that you can put the chips. That's all a big part of it.

But a big factor why they can pull this off is because, they're so centered in Taiwan. You kind of have a tradeoff between somewhere in your supply chain, you're going to have some sort of rigidity, just because you have to do this very high precision sort of manufacturing. So the rigidity in the TSM system is that everything's in Taiwan. And that allows them more flexibility in their interface with customers, in the designs that they do, and the equipment they incorporate, all these sorts of things. This is in contrast to an Intel, which Intel has long had worldwide manufacturing. But what Intel would do is as Intel, Because Intel, these fabs are mostly making the same chip, right?

Like, Intel's making, you know, their CPUs, they're making them at scale and for themselves. And so they would get a fab set up on the newest sort of process. They would design that fab, and then they would perfectly replicate that fab around the world. Like, all the way down to, like, the location of the urinals in the bathroom. Like, you would walk into one facility in, like, Israel or in Ireland or whatever. It'd be identical to the facility in Oregon where they, like, developed it or whatever. It's called copy exactly. It's like an actual thing.

And so that's the rigidity in the Intel model, right? So they can have flexibility in geographic location because they have the rigidity elsewhere. You need that rigidity in the model somewhere because you're doing this exceptionally high precision manufacturing where you have to get it right. And so how is TSM going to build these in Arizona? Now, this is my speculation. I don't know this for a fact. But my suspicion is what they're going to be. to do in Arizona is actually a little more like the Intel thing. I don't think these are going to be

foundries where they expect to service a thousand different customers with a high degree of flexibility and small order sizes and things on those lines. They have three really big customers. Those three really big customers, actually they have like four or five really big customers, but the three that are U.S. companies are Apple, the biggest by a significant margin, Nvidia and AMD. Those three companies, of course, would love to say, just for a pure PR reason. Yeah, our chips are made in America. I mean, Apple in particular, right?

Like, this is the first thing Tim Cook did on Tuesday last week, came out and tweeted, Made in America. That's right? A nice little distraction from the fact that their phones are utterly dependent on China, right? Right. And not shipping in part this holiday season because of the China situation. It's notable that, you know, when push comes to shove, it's Taiwan that has to pay the price here, not China.

But regardless, so I would bet these factories are going to be. really just set up to make Apple, AMD, and Nvidia chips. They're not going to be super flexible. Again, this is speculation on my part, just sort of thinking about what makes sense from all these sort of companies' perspectives. And so you have this situation where, yes, you have more cutting edge sort of manufacturing. Ideally, Intel will be caught up by that. Again, these are a generation behind. So, you know, Intel could certainly be caught up at this point. But you're getting advanced manufacturing that's still sort of dependent. dependent on Taiwan and tied into the global semiconductor ecosystem.

And it's good, but I'm not sure it's the best allocation of resources. Like, what would I do if I were in charge? If you were president, what would you want to do? So one of the big weaknesses that, we have a huge dependency on Taiwan. We have a dependency on Taiwan, not just for leading edge chips, where Intel is doing its best to catch up. but we also have a huge weakness in trailing edge chips.

These are the old small chips where TSM would have built a fab and say 2005 or 2010 that makes these chips. That fab is long since depreciated. So they're making, they're still operate. They're still making chips. But those chips are very cheap and they can, but they're still cash flow positive because these fabs are fully paid off. And these are the chips that go into things like cars, that go into things like appliances. Basically, there's a lot of chips that aren't necessarily on the absolute cutting edge that are super important. Now, there is zero economic incentive to make these new trailing edge fabs.

Because, again, the reason why they make sense economically today is because those fabs are fully paid off. In other words, those fabs have been built and are sitting there in Taiwan. So if they're just going to be sitting there, let's at least continue to produce these chip low margin, these cheap low margin chips. They're high margin because they're fully paid off, right? Okay, yeah. But they're low revenue. That's the key thing.

They're low revenue high margin. And so TSMC dominates this market. Now, the big increase in this market is in China. Because, again, China can make the old chips, right? It's not where they're sort of limited is in the most advanced chips. And so there's a lot of increase in Chinese capacity in this area. And a concern I would have if Taiwan were. cut off from the U.S.

is we would be completely cut off from these trailing edge chips. And the problem is we couldn't get toasters or whatever might be. Right. Like you thought that car problem was bad during the pandemic. It would be a million times worse, right? And so you have a situation where you have a real potential risk in these trailing edge chips. You also have zero economic incentive to address that risk.

that seems like an excellent use for government subsidies where we're going to provide money to build foundries that provide needed diversification that are not economically viable. And so that's like a definitional description of when subsidies are sort of important. So that would have been my number one sort of application. Now, when it comes to leading edge, like Intel was out there like threatening people. Like, look, you better pass this chip stack and give us money. we're going to stop building. Intel has to build leading edge foundries, like, or they're going to cease to exist as a business. Like, they don't need, sure, they will take

all the money they can get, but they already have the economic incentive to figure it out and to build these sort of chips. And so that's one of my first criticisms of the Chips Act is I would have much preferred to have seen a much more focus on this trailing edge thing. Again, just because it's a place where subsidies are kind of the only way to solve this problem. Yeah, well, and we should be clear. It's not baseless speculation on your part when we start to imagine who the biggest customers are going to be. Like Biden made his remarks last week and then met with a group of executives, including Morris Chang from TSM, Tim Cook from Apple, Jensen Wong from Nvidia, Sanjay Marota of Micron, Lisa Sue of AMD. like it's sort of the five or six families of like leading edge tech. So it doesn't take like a big, it's not a big leap to say, all right, well, this is who this factory is going to be serving, at least initially.

Right. And there is an argument where, look, in the very long run, these one generation behind fabs are going to be trailing edge fabs at some point in the future. So there is a question of like, what's your timeline for risk that you're trying to ameliorate here? Is it in the next five years? years or is the next 15 years? Because in the next 15 years, well, oh, if we think nothing's going to happen to Taiwan for the next X number of years,

then sure, maybe, you know, we'll have moved past the trailing edge by that point. Like the auto companies will have shifted from 90 nanometer chips to 28 nanometer chips, wherever it be. But again, we're not building 28 nanometer chips. We're building three nanometer chips, which are very expensive. Like, one of the underappreciated things that have happened in chip manufacturing is around 14 nanometer or so. It used to be the case that building faster chips were also cheaper. And so it was actually better to get the faster chip because, you know, when you could just fit more stuff on a wafer,

then, you know, you could actually make it cheaper. So it was amazing. It was like you got, you got a faster chip, a more efficient chip, and a cheaper chip all at the same time. Several generations ago, we crossed a line where that was no longer the case where it costs so much to build these foundries that the chips were markedly more expensive. And the reality is, is there's a lot of applications that don't need chips that are as fast as what's going to be produced in these factories. Like, like there, now, there are things that do, but there are a lot of things that don't. And so in the long run, are these going to be the sort of like bog standard, stick it in a car, sort of chips? I mean, for self-driving, sure, but for, you know, operating your window, maybe not.

The one thing that I worry about reading all of this is it sounds like this is great for Apple, this is great for AMD, this is great for all these companies who have their own incentives to diversify their supply chains. But if the goal is like long-term security of America's economic interests, then this is not the right way to do it. Yeah, I mean, again, it's not necessarily bad, but I don't know that it's really the most efficient use of resources. The other thing I would have done with the Chips Act is I would have much preferred to see it structured instead of just giving money to like Intel and say, you know, oh, you know, we really hope you succeed in building more, you know, these cutting edge chips. I would have structured it as being a guaranteed buyer. What Intel needs are customers. At this state, like there's no incentive for a cutting edge like an AMD or an NVIDIA or an Apple to buy Intel's. newest chips or to be a customer of their foundries. But you need a customer to get the volume to sort of

keep moving down the learning curve and to make these these viable. I would rather the U.S. government be basically a guaranteed buyer of these sorts of things. And that also it's like you have to actually succeed in building this and then we will buy your chips, right? There is a pot of gold, but you have to actually cross the bridge to get the pot of gold as opposed to we're going to give you the pot of gold and hope you cross our fingers that you cross the bridge. Those are the two ways I would have structured it. It is guaranteed buyer for cutting edge stuff and then definitely subsidizing the trailing edge stuff.

And the second prong there, that's what we did with development of the vaccine, right? That was part of Operation Warp Speed where we like guaranteed to be a buyer. I think that sounds right. I didn't. It sounds familiar. Yeah. Don't hold me to that. But I mean, the logic there does make sense.

So another basic question coming out of last week's event. Do you think this could actually come online by 2024 and then the other Fab come online by 2026? Sure. TSM is still projecting those dates. And I think they have been more reliable in their date productions than say in Intel. So yeah, but there's definitely challenges for sure. I mean, you know, I kind of mentioned it in my update, but there's a fair bit of scuttle butt, you know, here in Taiwan, just about cultural clashes and challenges. I mean, the reality is, you know, TSM's here in Taiwan is obviously a very prestigious company.

It's a great place to work. If you're super smart and super capable, it's a dream job. And you go there and you'll get, you know, subsidized housing and, and you'll get to say, you know, tell your in-laws that you work for TSMC. And it's sort of a great thing. But it's a brutal job, right? Like there's an athlete, like you're working long hours. This stuff is really, really hard. You're on call all the time.

And from what I've heard, there has been some tension both in terms of the Arizona stuff and also in terms of like they're working with Intel more, you know, because Intel is manufacturing some of their chips with TSM because they're so far behind that AMD is kicking their red because they are with TSM. And there's a lot of frustration about like, look, you're taking how many weeks of vacation? Like what do you do during the summer? Like your ski trip where? And again, I'm not defending or saying making a normative statement. Just a fact. There's no value judgment on one side or the other. Right.

But yeah. There is sort of a tension here. And I think that's, you know, that's come up in reporting in a letter that they sent to the Congress department saying, like, look, we're having a hard time finding, you know, qualified people. And that qualified people isn't just sort of like a the right education sort of thing. It's like a, look, there's a way we work. And, you know, they're actually shipping a bunch of Taiwanese people there. Like, they're building like an entire, like, village. for folks to live there.

And that's, you know, again, maybe that's a, maybe that is actually one of the positives where a development of talent and an ecosystem that can really, you know, work in this area is something that the U.S. needs. And I think that's actually probably the most optimistic take on this is, yeah, okay, sure. They're plugged into TSMCC now. If TSM's headquarters are destroyed, they may not be very useful because they've lost their connection to the mothership, as it were. And yeah, it's a generation behind, but we got to start somewhere.

And that is, I think, probably the most optimistic take on this. It's an analogy for lots of stuff, right? Like, this stuff went abroad for a reason. There just aren't those kind of jobs in the U.S. generally anymore in all kinds of areas that are like this. And there's a broader sort of question about globalization or culture or lots of stuff that goes into what it will take to bring this stuff back on shore. But this is definitely sort of the tip of the spear in that regard.

Yeah. Well, and it is pretty interesting. It wasn't just jobs that went offshore. It's like after doing this around the world for 20 or 30 years, the skill set necessary to do it effectively here just doesn't really exist. And like the processes don't really exist. And so training a workforce and establishing a culture, it's not surprising that it doesn't happen overnight. And, you know, Morris Chang was pretty honest about that when he spoke at the

event on Tuesday. He, like, he, his remarks were a little bit dower, I would say. I don't know if you saw them, but I'll read from Nikai Asia. Chang had always dreamed of building a chip plant or fab in the U.S. because of his own background. He was educated and worked in the U.S. for several decades, but his first experience did not go smoothly. It was, I thought, A dream fulfilled, Chang said. But it ran into cost problems. We ran into people problems. We ran into cultural problems.

The dream fulfilled became a nightmare fulfilled. It took us several years to untangle ourselves from my nightmare. And I decided that I needed to postpone the dream. And obviously, I think everybody involved in TSM says this time it's going to be different. But a lot of those challenges are still front and center as they try to bridge the gap there. Yeah, just for a broader context, what Chang's referring to is TSM has a plant in Oregon. Right. And it was opened around 2000.

And it's never been very profitable. Now, the way this stuff works is because almost the largest portion of the cost is in the initial buildout. They still run it because once you've built it, you might as well sort of keep running it. But there are ongoing unit costs as far as your workforce and things like that. and they've you know Morris cheng in particular has been very honest about this i mean you thought he's down or this again go go back and look at the brookings talk is even more the brookings talk is very straightforward uh it's it's just it's the unit costs how much it costs to produce one chip are just not remotely competitive with what it costs them in taiwan and now there's there's other

reasons beyond just culture and workforce stuff uh Taiwan in general the salaries are very low here for lots of complicated cultural and political reasons. Taiwan also has a real currency advantage. Taiwan, I think all evidence suggests, is one of the worst currency manipulators. But, you know, because they're- Can you explain that a little bit? Every time we talk chips,

you reference their currency manipulation. And without going too far, it sounds like you don't want to make any, like, explicit accusations, but I'm just curious what you're alluding to. The Taiwan dollar is, I think, it's much lower in value than it probably would be if it were allowed to float freely. And the, you know, there's kind of a, it's kind of a running joke among some of the expats here.

There's always like a daily article that's generated about the, you know, what the exchange trading was in that day. And it's like, oh, yeah, the, the NT dollar dropped to XYZ. And in the last hour, it suddenly increased back to XYZ, which is, everyone knows it's like someone coming in and changing the price. And basically the long and short of it is that chips made by TSM in Taiwan are likely much cheaper than they would be if the exchange rate were to floatically. And so they sell in dollars and so that that's a real advantage. So for TSM. So they sell in U.S. dollars. Yeah.

And so their cost are in low Taiwanese dollars, both because from from a currency perspective and also because it just labor is much cheaper here. Again, for other reasons, not just sort of currency reasons. And so there's a real cost advantage in general in Taiwan. So this isn't just a U.S. sort of having more expensive, lazier workers. Like I think that there is some aspect to that, but it's not the only story here for sure. But the long and short of it is they've had this plant in Oregon that is not worked out. Again, that was way back in like 2000. And so they never did it again.

And when it came to talks to doing again, Morris Jane was opposed to it. Because again, she's like, we try this. It doesn't work. This makes no sense. The only way to justify this from a TSM perspective is politics, where the U.S. was insistent that they build in the U.S. Again, it's a tricky decision for TSMC and the town government because to the extent there is diversification outside of Taiwan that reduces the incentive to defend Taiwan.

infested there. But at the same time, Taiwan has to play very nicely with the U.S. because their only real hope against Chinese aggression is the U.S. coming to their defense. So it's a very difficult spot to be in. And so that also ties into why I think these plants in Arizona are probably going to be relatively focused plants that are not necessarily just because, like, that's the most efficient thing for them to do. And so, yeah, you get into these really fine-grained details of, chip manufacturing and flexibility and capabilities and all that sort of stuff. And none of that actually, I don't think, resonates necessarily with politicians, right? Like, just like, hey, we have advanced chips in the U.S.

We can have a press conference. It is, you know, I found it kind of amusing that all of these announcements came from the White House, not from TSMC. Yeah. And I'd say that's where I start to ask questions about like, A, who is driving this and B, who should we be criticizing about the misaligned incentives and like ambitions here? Because like I understand. To be fair, I mean, if they are doing something that does not make economic sense, that is in some respects a justification and defense of government action.

Because the reason you would have the government step in is precisely to achieve a result that you feel is important for reasons beyond just the economics. Right. Again, the economic incentive is just to keep it all in Taiwan. And that's the incentive for TSMC. But if you're going to go through all of it, the big dog and pony show here, like, why would you do all of that just to help Apple and Nvidia? Like, I just think that's short-sighted. And it's a little disappointing.

It's not what I was imagining when I heard the chipsack was past this past summer. And I know the TSM stuff predates some of it. But it's all sort of the same long-term goal. and security interests. And so I'm not sure why we're just narrowing the focus. Like I understand even from TSM's standpoint, it's so much easier to just have a narrow focus where you can make your operations as efficient as possible

and try to make up some of the lost profits that way. But again, America should want more variety than it seems like they're going to be getting from these fabs. Yeah, I mean, again, it's tough because it's very easy to sit here and poke all kinds of holes in this. And I do think it's a legitimate defense. Say, look, we have to start somewhere. And also, it's a legitimate defense to say,

I don't know if we can depend on Intel on the leading edge, right? Like, and Intel is like Intel right now, the crap's really hitting the fan, like all the reasons why people like me have been writing about how Intel's in trouble for years. And everyone, you know, sort of people on Wall Street, like, look, their profits seem fine. It's like, yeah, the problem is that it's going to show up

like there's a huge lead time for this stuff and all those profits are going to. Wait, guess what? The profits that everyone was pointing to that are mostly from the data center almost completely gone last quarter, right? Like, and now TSM has a new server chip out that's just so much faster than what Intel has. And Intel has a new one coming, but it's probably going to be slower and it's also late. And there is an aspect of like, look, we can't be only dependent on Intel as far as leading at. So getting TSM into the U.S., again, even with all these sort of limitations,

and the fact that even this 40 billion investment, it's not sufficient to apply just Apple, right? So these things are so expensive to build that, you know, it's a relative drop in the bucket. And that's TSM's CAPX for like one year. And but this is an investment over many, many years, right? Because they're going to be investing tons elsewhere. So look, I don't feel it's quite right on my side to just be a complete debby downer about this.

Like, like the, to the extent you think it's important to diversify out. of Taiwan. This is absolutely a start. We're in a better place than we were before. But it's also, I did feel a need to push back on this sort of, there's a bit of triumphalism. Like, problems fixed now. It's like, no, like we're not even, not even close. Yeah. Well, and if you talk to people in Washington, they're like, look, we pass the Chips Act and we're bringing this stuff onshore, and we're going to be more secure long term as a result. And I want to believe that, but I can also see some of the holes in the logic. Let me ask you this.

Is it easier to build and operate trailing edge fabs than it is to build leading edge fabs? Yeah, for sure. Yeah, there are trailing edge fabs in the U.S., mostly run by global foundries, which was, you know, used to be AMD's manufacturing arm.

It was spun out like 12, 13 years ago. And so we do have those sort of fabs in the U.S. But yeah, for sure, it's a known technology. That's why China's doing it, right? Because it's relatively, It's relatively simple. It's relatively, you know, sort of inexpensive. And yeah, I do think that should be more focus. Again, it's a start, right? It's a start. And look, I'm in Washington, D.C. I'll get a note over to President Biden and we can get the wheels in motion for the

trailing edge. Final note here, Morris Chang, and this is the particularly dower part of his remarks last week. 27 years have passed and we've witnessed a big change in the world. a big geopolitical situation change in the world. Globalization is almost dead and free trade is almost dead. A lot of people still wish they would come back, but I don't think they will be back. Do you have any reaction to that quote? Like what comes to mind? Well, for sure, it's a forward looking statement.

I think more than a, the stuff that's built is not going to, like the integration with China is not stopping, right? the integration, the dependency on Taiwan is not going away. The big question is sort of investment going forward, right? So if you're, you know, how much investment in new capacity or new ability are companies going to make in China looking ahead? And for Taiwan, how much they're making Taiwan? Taiwan is one of the biggest, I sort of, you know, mentioned in passing with Apple, but, you know, Apple is, hey, if we can reduce our dependency on Taiwan, let's do it, right?

Taiwan is sort of stuck in the same boat because of these sort of race. risk factors. And that's a bummer. It's a bummer, I think, for, for, it's certainly a bummer for Taiwan. But again, it's going to be so hard to unwind what has already been done. And maybe that's a good thing, right? Like the, you know, maybe that remains the biggest reason why my percentage view of a chances of a conflict are very low, and I think lower than the attitude in Washington, DC just because this stuff over 30 years, this stuff gets so tied. And note like who can actually trace through all of the different component pieces that go into some of this stuff, right? Like you have suppliers like multiple levels down that go into making these chips or making an iPhone or making a stroller or whatever you can think of that all of those pieces need to be reworked and redone in a world where you're not going to be dependent on China or.

or you can afford the risk of conflict in Taiwan. And also, again, Taiwan is not motivated to get away from that world where there's no longer a dependence on Taiwan. Yeah. Yeah. Well, I mean, there's no question that the mutual dependence has been a bulwark against conflict and I think still acts as a bit of a security blanket and insurance policy against, like, true standoff between the U.S. and China because they would be mutually assured economic calamity

and probably a recession for the entire world in that scenario. Right. I mean, you see an aspect of this with Russia where dependence on Russian energy is obviously causing massive problems in Europe. Now, that's for a pure commodity, right? Oil from Saudi Arabia or from the U.S. is, you know, the same as oil from, from from Russia, broadly speaking, right? When it comes to this specialized sort of things, there's, you know,

there's no replacement. There's no replacement, right? And so the specialty things go very, very deep in all sorts of ways that probably no one in the world is probably impossible for any one person to hold all that sort of in their head and to sort of be aware of it. Well, speaking of unknowable futures, a lighter topic, but no less relevant in the United States. Every few weeks, it seems like there's a new announcement from a billion dollar tech company that's resetting its remote work policy. A few weeks ago, it was Snap

and Evan Spiegel who said he wanted to see employees in the office 80% of the time, and that's the new policy starting, I think, in February. I want to get your thoughts on where things stand now with remote work and that whole conversation. But first and foremost, you're a fun person to ask about this because a lot of your life changed because of remote work. So can you explain that part of your story? Yeah, well, it makes me a terribly unreliable narrator about this topic. So put that up front. This is very much speculative on my part. I did move back to Taiwan once I started working for automatic, which is a fully remote company. And it was, it was really interesting. I mean, there's a, I think one thing that is underappreciated is there's a lot of

investment in sort of infrastructure and processes that goes into making this work well. So one thing with Automatic, for example, is it's a very sort of textual written culture, and everything's asynchronous. And so there were like three levels of communication that you would have. So number one, you would like once a week, your team would have a meeting. Back then we used Skype. There was pre-teams or pre-Zoom. And, you know, just sort of touch base.

Well, actually, let's start the highest level. So how do you get to know your teammates? Well, twice a year your team would get together somewhere in the world. So we had to get together in New Orleans. We had to get together in Rome, where the whole team would get a big Airbnb, everybody would go in. Theoretically, you were there to, like, plan for the next year. Really, you went there to sort of, like, just eat a bunch of food and see some tourist sites and get wasted every night together. And the idea was this is really just, like, building camaraderie and culture and that sort of bit.

because the actual work was sort of done at home. And so, you know, Matt Moldwegg would always talk about, look, our, we're not, this isn't really a cost saving thing because we spend so much on these trips that it makes up for our office space that we don't have. And then there would be. Can I just say, look, I really enjoyed the Stratacri team meetup at Vegas Summer League, but can we do it in Rome next year? Yeah.

It was a good time. Lots of good food for sure. There was always a company-wide meetup once a year. So I think we did it in like Santa Barber. but I think right now, because automatic's got much bigger, what they do is they go to like a ski resort in like, and just in like the fall. Right,

because there's like no one's there. And no one wants to go there, then they get lots of capacity. And the advantage of automatic size, look, this isn't necessarily a money saving thing. It's a talent acquisition advantage where we can get people that want to live, you know,

live and work remotely. And it's a working on a sort of invertebracons. So number one, you have that that small in-person component. Then you would have like weekly meetings where everyone would get on a video call at the same time. It always tough schedule because you would have like our team was literally scattered all over the world.

Then you would have actually automatic has this internal WordPress theme. It's called, called P2, where it's like a blog and you write, you know, every day, every team is deputized to write a summary of everything the team sort of accomplished and did that day. And what's interesting about this theme is like all the comments are sort of right there on the front. And it's meant to be sort of like almost like a message board sort of thing. And then you could have a you could every morning people would start their day. They'd have an internal feed of like all the teams in the company. They could just go through and read and know everything that's happening in the company. And so they're just sort of like that was sort of

of a real discipline of every day sort of writing that posting that. Then there'd be discussion and you could always be aware of what was going on. Then the actual conversation all happened in IRC, which I mean, I think basically Slack before Slack where there's, and all this was all saved in archives. You could always go back and find it and search it. And then again, everything was sort of asynchronous where there was no assumption that someone would respond to you immediately. And so you had to like think about what you needed from someone. You had to like get all the details. And it was, it's a very different culture than sort of the in person running to some of the water cooler or stop by their desk or, you know, that sort of thing.

And there was a lot of discipline around having alternative tools and means to sort of accomplish that. And I think that was something that, number one, it worked, it worked well. But also that was something that a lot of companies, I think, struggled with when the pandemic came because they just tried to shift everything that they did that was in person to suddenly being remote instead of in. We'll have Zoom meetings now. We'll do X, Y, Z. And it didn't quite translate right because so much non-explicit. like knowledge sharing that happens sort of in person, there wasn't processes and habits around writing it down such that a way it's visible and accessible to anyone at any time. And so I think

to the extent companies succeed remotely, it will require and depend on really changing ways that you operate to support that in sort of a meaningful way. In reality, I think the way it actually worked is most companies, to the extent they support remote work, it was still mostly live. You had to be in a time zone where you could be in the same time zone as your folks because it's hard to shift from being, you know, working synchronously to asynchronously. But, but yeah, it was a great experience. But, yeah, so I did have that experience. That was also, you know, a decade ago. And since then, I've worked by myself. So, you know, world's worst authority on, on workplace sort of culture and interaction and how stuff has changed. Again, when I was doing this, it was before most of the

that people used to say even existed. Right. Do you think you would have started Stratory if you weren't working remotely? Oh, yeah. I started Schetectary at while I was still at Microsoft. So it was my goal for it to be my job. You know, automatic was great for me because it was hard for them to, you know, not allow you try to blog on the side, given that it was literally a company about blogging.

And then also, you know, it was a chance, you know, I wanted to move back to Taiwan. but my long-term plan all along was that shrekery would be my job sort of in the long run. So, yeah, so it was, it was actually made it more difficult in some ways because I didn't have the shortcut of showing up at work every day looking like I was working. Instead, like, I felt a lot of tension because there was a strict policy you couldn't have a second job for understandable reasons, right? And so I wasn't monetized as shattery. And so when I did want to make shattery a paid product, I had to quit at automatic. But even while I was still there, strategy was kind of blowing up.

And there was, I could feel the tension like, why is, what is this guy doing? Is he doing anything? It seems to be like spending a lot of time running. He's really productive over here. He can't possibly be doing anything over here.

Well, the reality was I was actually pretty lucky because I was hired to be like a growth engineer, like, uh, which really I, I didn't know anything about. I was not really qualified for for a team that a few weeks after I started changed what they were working on to.

to work on something that would come out many months or years in the future. And so I literally had nothing to drive growth for. So I had like no job. And so actually I owe Matt a lot, Matt Moldweg, because I think he did kind of run interference for me a bit. And we did talk a lot, me and him actually kind of worked with him directly where, you know, he would ask, you know, they're working different things. I wrote a couple like strategy documents for the company. He claims that I was very, it was helpful and impactful. maybe he's just making me feel better.

But honestly, it was incredibly stressful. I felt really bad because I didn't actually have a clear thing to do in the company that was measurable and seeable. And I also had this very visible blog that was clearly posting on twice a week and was drawing a lot of attention. And it was really, really, I felt really bad and stressful at the same time. This thing was blowing up. And so I, of course, wanted to keep working on it. So I was in a real rush to actually try to get out, not because I didn't like automatic. but because I just personally felt so conflicted about these sort of two competing things.

You and Matt are in a great place. You had a terrific interview on March 12th, 2020. I'll put that in the show notes as well. It's a pretty harrowing reread. I read it before we came on to record here. The intro, you're talking about Tom Hanks getting COVID, the NBA shutting down the night before, really took me back to March 2020, which is not. not something I necessarily wanted to do. But while I have you, world's least reliable narrator on this topic. So what comes to mind when you start to imagine like what the future of remote work might look like, what it should look like?

Do you think we're moving to a hybrid model or, you know, there's a commercial real estate aspect to this? And there's also like broader social concerns that will probably factor into at least the impact this has on society, I don't necessarily know that any individual business has a responsibility to think about broader society as they make these choices. But just with that general question as a starter, what are you most interested in over the next couple years here as we all figure this out? I mean, that's a wide-ranging question. I think from a micro perspective, there is a situation right now where tech is going through a lot of layoffs. I think there is

going to be increased pressure on tech companies, in part because of Twitter, being very visible, cutting a huge portion of the workforce and still functioning. And there's going to be a lot of questions like, wow, do you really need all these workers? And so I think there's going to be,

there's a real shift in sort of bargaining power to companies away from workers. And from a management perspective, of course you'd rather have people in the office, right? And so, you know, the, and so I think in general, that's why you see the news like the snap news that you reference and you're probably going to see it from other companies a real pushback sort of in this direction. So that's sort of number one. Number two, I do worry on behalf of workers that remote work is one of those things that is great for a couple years. Then you look back in a couple decades, you're like, oh, maybe that wasn't so great after all. It's, it's, you know, if you're going to actually implement the processes and the coal, of effectively managing remote work, then it's like, well, why should I hire the engineer

somewhere else in the U.S.? I could just hire an engineer in India? Are I could hire an engineer in some other sort of lower cost markets? And, you know, I think that's an underappreciated long-term risk for a lot of these folks that are very agitated in favor of sort of remote work now. And I think the analogy here is globalization that hit the blue. collar class, and we sort of just talked about now, it's a risk for the white collar class. So it is happening in conjunction with this AI bit that is sort of automating arguably some these functions.

And if you're getting to this more processes, like at what point is it not just hire or worker in India, but just have AI do it, right? Like there's, so that that's definitely a macro risk. I do think there is a potential real upside to the extent that we're shifting. talent and high paying jobs and ecosystems out of just a few core areas like New York and San Francisco, Los Angeles, whatever, into broader areas. That's probably a good thing, you know, just sort of socially and, you know, diversifying the economy internally to the U.S. to an extent.

But all these are, there's some really big questions that are raised here. I mean, what do you think? I mean, you have more recent experience. in the workplace that I do. Yeah, I find it interesting because it's an intersection of two big themes that I think will probably recur throughout the decade. Like, I think number one, there's this collective realization that reliance on technology as a substitute for in-person interaction can be a mistake.

And, I mean, we've talked about it a bunch on Sharp Tech, even just the last couple months. A bunch of different sectors in society, you can see that, like, we, can stay connected to professional and social networks simply by sitting at home and sending emails. But like, the efficiencies gain there aren't coming without a cost. And so I think even as a young lawyer, like, it helped to be in the same conference room as a partner, as opposed to just slacking back and forth and making one or two phone calls to him every day. Like, if you could just walk up to somebody and ask them, you just learn a lot more. And I think if anything, it's more important for young people than the more experienced people in a business to be in the office.

And the other thing that I find interesting about all of it is like another theme we've seen, if you want to understand the impact of certain policy choices, like you have to look at them from a holistic perspective and consider some of the secondary effects. And obviously some of that's always been true. But like with remote work, it may be more. affordable for individual businesses to shift to remote work and reduce some of their commercial real estate footprint. But what does that do to like an economy in a big city? And what does it mean for society if a bunch of privileged people are working from home and everyone else is running

around outside in the service economy? Like I don't really like the idea of that stratified version of society. It's sort of what we were living with for a couple years there. Yeah, I was going to I think it's really interesting to consider would sort of lock down policies or school closures or whatever, you know, there was an extent where the people that sort of dominate the narrative in the media were all able to work at home. And there's a lot of folks that were, you know, quote unquote essential workers that were out and about and, you know, it had to leave their kids to fend for themselves and figure out Zoom because they had no choice in the matter. better, but also their voice was not very prominent in sort of the debate about this sort of stuff. And on one hand, you can look at tech and say, look, tech was the only industry like tech kept the economy running during the pandemic. You could also back up and say, if we didn't have this tech, maybe we would have moved forward

much more quickly. I mean, maybe it wouldn't have you been a debate because we wouldn't have had any choice. I mean, that's what's happened in. There's been pandemics, right? There was a pandemic of probably similar withality in like 1957 or so, which no one remembers a lot of people died, but we just kept trucking. There was a recession, but sort of just kept trucking through because we had no choice. That was, that was the way sort of stuff worked. Yeah, well, and I just think about my own life. I like interacting with the outside world.

Like, I liked going to the office every day. As soon as we were allowed to go back into the office during the pandemic, I was there every day because it was just better than sitting around in my house. Now, flip side of this, I'm a giant hypocrite as well because I also, over the last year or so, have taken advantage of the work from home model and have learned how to incorporate exercise into my day. And I'm just like, in general, a lot healthier this way. And I think a lot of people, a lot of young people feel better about their professional lives when they have a little bit more flexibility and don't have to spend like 90 minutes commuting back and forth to a job every day. And that's important to factor in as well.

I mean, almost everyone my age prefers to be working from home than working from the office. So I was sort of the weirdo outlier, like among the associates of my firm. I like how you are still classifying yourself as a young person. I'm not sure how long you can carry that. It might be. Yeah, I'm going to cling to that as long as I can. but it's a point well taken. The young versus old thing, I think, is applicable to companies as well, right?

Like the older a company is, and if it has an established business model, like, say, a Google or a Facebook, I mean, theoretically snap, but, you know, we'll throw them in there. You know, the easier it is to have people, because there's sort of like, no, there's jobs that need to be done. You know what those jobs are, and there's measurable outcomes as to whether the jobs are done or not. Like, that lends itself to sort of remote work. if you're a startup, like there are examples of startups that are fully distributed. I think GitLab is one of the most popular ones. But there's a reason everyone goes to the same few examples of successful remote startups because there is some aspect of when you're figuring stuff out,

when processes are not clearly defined because you can't afford to sort of become very structured and regular because you're still trying to achieve product market fit. Like when you're in that, you know, go back to the TSM example. When you have to be super flexible and figure stuff out, it's really beneficial to have everyone in the same spot. When you get to a spot where you know exactly what you need to do and you just need to reproduce it, then you could be Intel and have fabs all over the world that are the exact same.

And that tension and tradeoff, I think, you know, that young versus old isn't just an employee thing, which I think is a very correct and astute observation by you. I think it's probably also a company thing. Yeah. Well, I think what's most interesting to me is how unresolved all of these questions are going forward. And we're all sort of going to figure it out together.

And I don't know what the right answers are because I look at my own life. And I am happier with my current setup. I just have to build in certain, like, I have to be very proactive in the same way that automatic was. I have to build in opportunities to interact with the outside world. Because if you're just sitting around and this. like there's the loneliness epidemic that has been in the news

recently where people just aren't seeing other humans as often as they once did and that's a big problem and if you're just sitting around you may feel connected but you often look around and you're just like moderately depressed and like it and I have to

be very careful to not fall into that trap and I am I don't know whether you've changed your work from home habits over the last 10 years have you developed any like skills and tips? Well, for sure, one thing that I did for a while have like a separate place to go work because I think it is, there's just a lot, particularly when my kids were younger,

it's just challenging to be there but not be there, you know, to an extent. And so that was, you know, I'm back at home now, but in part that's because everyone else is gone more often, right? So I think I'm sort of more more on my own in that regard. and that's definitely a real challenge. I think it's just you can, your partner and your kids can understand intellectually that dad's busy right now. But it, like there's an emotional aspect where he's just right over there.

I go, I can't just pop in and ask them. And, you know, at least for me, I get, when I'm writing, I'm so locked in. I'm like catatonic to everyone around me. And if I get knocked out of that, it takes me a couple of hours to get back into it. So now you know, when one of my updates. shows up like four hours late. I probably got knocked out of it. And it took me a while to sort of walk back in or I never could walk in in the first place

because I had too much going on. So that's definitely one thing that's going to learn. Another thing goes back to like the cigar club bit, right? Like you need some sort of, I agree, social interaction, I think in real life outside of just sort of online. And I didn't fully appreciate to your point how sort of, I don't know if depressed is the right word. but lacking, like what a hole I had in my overall psyche and happiness by having insufficient,

this is particularly the case when I came back to Taiwan. I was really starting to checkery. And, you know, like that, I had friends online. I had group chats and things on those lines, but I didn't really have any interaction outside of my family and extended family. And that's not, that's not healthy at all. And I didn't appreciate how unhealthy it was until I sort of got out of it. But yeah.

So those would be the two big things. You might need your own place that's separate, physically separate. And I completely agree you need something regular to give you physical interaction with folks independent of your family. Your family's great. Not saying your family is bad, but it's just for your sort of own mental health. Yeah. I mean, depression aside, it can also just put a ceiling on how happy and fulfilled you might be if you're not getting out. and, you know, talking to other humans beyond your little pod there.

And as far as the family stuff, that has been a struggle. My entire career outside of the two years I was working in a law office because I was a writer. And now I'm like podcasting. And so I'll be sitting at my computer. And it's really hard for my wife or sometimes if I'm working by my parents. Like, nobody can really accept that I'm working and like just give. me space because they see me on like Twitter every now and then and they're like you're not you're just like sitting there hanging out and which might be um it yeah I know that's the that's the other

problem but look um we need email questions for Wednesday night's show so email us at email at sharp tech.fm and I would make sure on that email to us it's not email at sharp psychology. FM so this is armchair psychology ends with this episode but I would love to hear what other questions people have about remote work. And we can keep this conversation going in the next episode and hit on anything else. There's a lot going on out there. We've got a question about Amazon recently. I want to talk about them.

For now, though, Ben, listeners can give the gift of Stratectory and we will keep this going later in the week. Yeah, you can call it give the gift of Sharp Tech. Remember, it's all it's all bundle. We go to your show and there's a link there. Yeah, happy holidays to in advance to all the people that you bless with this sharp psychology content. Santa Claus was the original bundler. So we're just following it his footsteps. We'll be back later in the week.

Talk to you later.