s02 e10 Joel Mintun – AI and Global Supply Chain

Welcome to the Big Cheese AI podcast. I am your host and my name is Sean. I think I said that twice. Yeah. I'm here with Brandon. Hello everybody. Jacob, as usual, the CTO. And we have a special guest today, Joel Minton. Hey guys. He is in the semiconductor industry in the supply chain space, which that is today's topic, which is a little bit exciting, less nerdy, more businessy, but kind of all the more interesting because in order to do all this stuff that now we are trying to do, it takes a lot more CPU and computing. And people don't really think about that, right? Like you have the new refresh cycle happening with the iPhones and all that they're advertising different chips and the M4 and this the first in AI, I haven't my AI enabled M4 chip into my iPad here. Does it really take more to do all this stuff than what we had before? Let's start with that.

• So yes, obviously, right? It's a lot of horsepower to do all of this AI stuff. But what's kind of funny is like when we, and if you just kind of take it on face value and you look at kind of like the market, everybody pretty much just thinks it's all Nvidia. That Nvidia is pretty much the ruler. That's where all your stocks should be, or all your money and your stocks should be like,
• Wait, that's not-- (laughing)
• But the reality is, and we were just talking about it, is that people might be surprised to know that NVIDIA doesn't actually manufacture any of these chips. Now they design 'em, they design really great ones, they have a lot of things, but there's actually one specific company that I think we'll be talking about them, and I'm not gonna say the acronym 'cause I might just like, "The guy's gonna butcher it, I'll let you say it." That actually is coming out to be the five billion pound gorilla in the industry that pretty much rules everything right now.
• Yeah, so if you look at just in general, what's going on, you have to manufacture chips. And this came up during COVID, right? Like a lot of the companies started winding down a lot of their output and their manufacturing and it created a lot of bottlenecks throughout the world because everything needs chips now. Like you need a chip to start your car, right? Like it was, I don't even know how many chips are in modern vehicles, but they're just sticking over a thousand over a thousand over a thousand. Yeah, like 10 years ago, there was like my wife's car 2012 40 chips, maybe like windows. Now like a Tesla probably has 2000 plus. That's why my car dealer when I bought a car a couple years ago, they were like really grilling me on the insurance or the warranties. Yeah, like now you're your mirror has 30 chips.
• I'm not buying that shit.
• And they can't ship a car if it's missing a single, one of those, you know, 2,000 ships. So that's-
• One of our clients is United Auto Workers Union, and they have the Louisville Assembly Plan, the Kentucky Truck Plan down in Kentucky, and during COVID, they had stops on a lot of work just because they couldn't get, they could get steel, you know, all day long, but they couldn't get chips. And so their manufacturing lines were stuck. And obviously, I don't know how you think about that from a product design standpoint. We basically bottlenecked ourselves and we have become dependent on these chips. And I don't even know how to explain it, but like--
• Well, why did that happen though? I just had a kind of curiosity, like what with COVID actually stopped our ability to make and produce these chips?
• It was really just supplying demand. So no one anticipated the surge in demand, right? you had all these people going home, like my company sent entire call centers home, gave them a laptop, accessories. So you had more people buying consumer electronics, the data centers had to have more compute to support all the teams and Zoom meetings. People were buying more video game consoles. Everyone was just sitting at home, you know.
• Computing, right. (laughing)
• And the chip manufacturers actually were assuming a drop off and to me end.
• Right, they had the people. That's what happened with a lot of the economics of COVID is that people reduced their output, but the actual demand in certain places such as chips, such as bandwidth, such as whatever, dramatically increased go-jaker.
• So the demand side went up, they thought it was gonna go down. And also the supply side went down because they cut it, but did they also have issues just like, like getting people to go to work?
• Yeah.
• Was they big shut down?
• Yeah, they shut down fabs.
• I mean, in China, you couldn't even leave your house.
• Yeah.
• Yeah, all right.
• In China is where they do a lot of the assembly, like the final assembly.
• Yeah.
• And Taiwan's where they do a lot of the actual, like, our manufacturing.
• Chip manufacturing, yeah.
• You said fab.
• Yep, what could you--
• Yeah, so if think of the chip industry and the semiconductor industry into two kind of like sides, there's the chip design, which is Nvidia. They don't actually manufacture anything. That's AMD, that's Qualcomm, Apple, and Microsoft actually designed their own chips now and just set them straight to manufacturers. So you have chip designers, then you have chip manufacturers, That's TSMC, that's Samsung, that's the company I work for. Intel is like one of the few companies that still does both. They design and they manufacture, but they have to kind of split that in the way you think of it. Because I think in popular media, they just sort of use them interchangeably.
• I googled it to prep of the show. I googled top chip manufacturers. And BDA was always number one. And what you're saying is they're a designer, not a manufacturer.
• Correct, yeah. higher margins.
• Yeah.
• You know, their campus is very nice compared to like a manufacturing office and that's what they're doing really well too.
• And most of the design of chips is operated within the United States, is that correct?
• Right.
• Okay.
• I mean, it's kind of wild though, if you think about, if you really think about what NVIDIA has done in terms of the value that they've created for their company and they don't actually manufacture.
• Yeah.
• They're like a design studio.
• Well, well, it, so I think it's important to understand the paradigm between other countries and the United States of America. And it doesn't really have to do with, it doesn't really have to do with anything besides just the climate and the culture of our country. Our culture is creative. It's everything we do is we think up stuff. And then people from the entire world, they're like, I wanna go to America because I'm gonna be able to be creative and think up stuff.
• All right.
• And remind me later, 'cause I wanna touch on the Arizona plants that the Taiwanese chip manufacturing company, they're having issues in Arizona because of the culture shock and culture change. Yeah. So it's a really good story, the supply chain of the semiconductor industry, right? Because you have Americans designing and thinking up all this stuff. And then you have the suppliers, Japan, Germany, Netherlands, right? these raw materials. You have you have Taiwan, which is, which is manufacturing, all this stuff. You have assembly in China. You have this, you just, it reflects kind of like all these different things about, but, but in video, Apple, they're just really, really smart. And they, but all they're doing is sending us spec. Yeah. Right. Yeah. Yeah. I'll do one to one to one supply. Right. So that, that kind of sets a stage for, for the story that we have to tell, which is, and you guys, say with me and ask and let's just kind of go through it together. But there's this, the, what do you call it, the 400 pound gorilla, the elephant in the room, the company and that's TSMC. So there's a company in Taiwan that- Taiwan, some eye conductor manufacturing company. Yes, correct. And Taiwan, the reason why it's named Taiwan is because initially in 1987, when the company was created, 50% of the money was put up by the Taiwan government. The Taiwan government decided, "Hey, we are going to try to raise the standard of living for people." And so they took billions of dollars and tried to figure out how to do that. And there's this guy. Go ahead, Joe. It was an agricultural island. How it shifted from that to the largest chip manufacturer in the world in 40 years is just wild. It's super wild. And obviously that's a story of government investing. And, you know.
• Well, this kind of, I mean, isn't that what we're trying to do with our chips and science act, right? Is that we're trying to basically say, "Hey, listen, we need to get some of this stuff happening "on our land here." Especially if we're putting these chips in our defense and our missiles and everything else that we've got.
• Yeah, and just don't give stuff to China, it's basically that too. But anyways, so you have this classic story of this guy who changed everything. Okay, so his name, this guy's name is Morris Chang. He was born in China. He grew up, he lived through three different wars. He bounced from China to Hong Kong to China to Hong Kong, and then eventually moved somehow during all that, lived through World War I or World War II, which they call the Sino-Japanese War. It's basically World War II. The dude somehow gets to Harvard. Okay, somehow gets to Harvard, gets to Harvard and goes, they don't even have an engineering program. What am I doing here? transfers to MIT, gets a mechanical engineering degree, goes and works for a company.
• Exist instruments.
• Yeah, that, well, he worked, yes, he eventually, he worked at the company before that, but yeah, those works at Texas Instruments, kind of a normal entry level, ends up rising pretty quickly through their chip manufacturing his first client's IBM, and they're building chips for mainframes.
• Oh, good Lord. So this guy decides he's going to change the game by introducing the concept of basically inverting the price curve. So what manufacturers used to do is they used to, and this is when computers were super expensive, I don't even know the context of what IBM was selling these main fans for. Right, and for the size of rooms, they were just ginormous. Right, and I don't know what their refresh cycle was, but they were designing… Probably 30 years. (laughing) I don't know. But the way that they would do that is they would basically tax their customers for new products. So they would go in, they would say, okay, you gotta buy this new thing. It's gonna be super expensive and we're just gonna raise our prices. So what cost $1 today now costs $10? And so they would do that and that's how they would basically justify the investment on their end too. Anyways, he's basically like, you know what? Every single chip that comes out of this company, 10% of them make it to market. Their failure rate on their chip manufacturing, and they weren't even chips at that point. They weren't even trim.
• It was Germanium and not silicon, but yeah.
• It was different technology.
• This big college.
• This big college.
• The punk market.
• 90% of the stuff that they built failed. And he was able to basically get it up to 20%, which is what made him famous at Texas Instruments. But what they did was he said, the only way for us to get better is to iterate and improve over time. And in order for us to iterate, we need more volume. And so what we have to do is lower our prices. That one idea has now cascaded through time. Go ahead.
• Yeah, well, and I think, you know, I think I'll jump to the end a little bit, sorry, but that idea is where the moat is in manufacturing where if you iterate and figure out how to increase yield on these wafers, you make a lot more money, right? It's super expensive. You got this way for if only 10% of them are working at the end of it and now you have 20% and then you get to 30 and then that's really hard to just reproduce for another company 'cause it takes, there's no substitution for the learnings when just going through the process, right? Like you can hire the engineer that did it but at the end of the day, like that just takes time and there's no real shortcuts to get to that.
• And it's such an exact process.
• Yeah, like literally there's probably fewer processes that need to be as exact when you're dealing that three nanometer.
• This is made of-- - So what the hell is that? So when we talk about nanometers, you're talking about on these wafers or these chips or whatever, you have to install some sort of electrical device.
• This is a transistor, right? So yeah, and the smaller the better, 'cause the more you can jam into the wafer itself. And then you have all sorts of issues with that, 'cause the smaller they go, the closer they get together and they have heat involved in that, so they can be firing off their little neighbors. So like, yeah, so like, it's all home. Go ahead, Joel.
• Well, that means, so just to put in context, we're talking to three nanometer transistors. It's kind of like feature size. Some people say it's more of a marketing term, like 5G6G now, but it originally did. It's basically like how small stuff is. So like a little, three millimeter ant, like a little ant on your watermelon, 300, sorry, three million nanometers.
• Oh, wow.
• And there's a, and the guess how many are in ship nowadays, like an NVIDIA chip, there's like 40, 50 billion easily. Human strain of DNA is two and a half nanometers. And we're making chip parts, chip features that are that small. And it's just like so crazy. So people think it's literally the pinnacle of like human engineering is the microchip and where it's going. So I can't, it's hard to even wrap your head around. So this stuff is the hardest stuff in the world to build. So you go back to Cenk, all right? So Cenk gets, he's at Texas Instruments and he's kicking ass. Well, they put him over. Guess what's happening at the same time? They're building the frickin TI-83 calculator.
• Still got mine.
• They still have those, right?
• They still have.
• They used to play, I used to play that.
• Drug wars.
• Drug wars.
• Oh yeah, I was trying to explain drug wars to someone the other day and they were like, so you were like a video game about dealing drugs. I'm like, yeah. (laughing) And you're in middle school? Yeah.
• I don't know.
• How did we load those games onto that thing?
• It was a whole process. Like, you had a lot of nerd cred when you actually got it running. And everybody was like, whoa.
• Well, you're probably to do load it though.
• You plugged it into someone else's calculator? I forget.
• It was like a network effect. Like it was like someone had to pass down the data to you.
• Dude, no, it was just like, how did any memes pass back before the internet?
• How did any of that work?
• No, the hamster and Richard Gere, right?
• Oh yeah.
• The hamster and the hamster and the hamster.
• No, no, no, no, no, no, we're actually, I mean, I guess it's kind of a hamster dance. (laughing)
• No, no, no, no, no.
• I'm not a hamster.
• But what the reality was is, I mean, that thing spread across the entire United States, and we had no internet and, you know, we're in Indiana, so we were always like five or six years behind the stories, but like it landed here. How that stuff--
• And that was like objectively false, right?
• It was, well, maybe, I don't know, I still don't know. (laughs)
• Anyway, so they move Chang over to this consumer products division, and he absolutely takes a shit because he's not a consumer products guy. He's a chip manufacturing process expert. And so he fails, he ends up just, he's 55 years old. And he's actually, he goes to a couple of different companies and he's just kind of washed up at this point. Goes, basically loses his job. Okay.
• At 55.
• And he puts, he puts open to work on his LinkedIn. (laughing) Okay, right? And the time on his government approaches him. And they're like, we heard you're like, pretty kickass at building chips. They're like, we're starting this initiative. And he's like, I'm in, but they're like, hey, you gotta go raise 50% of the money. You got it, we'll give you 50%. They get the, just FYI, the total initial investment to build TSMC was $228 billion.
• And what year was it?
• 1987 was when they started.
• So that's like a couple trillion billion.
• We're not talking about, we're not talking about some seed fund.
• No.
• Okay, we're talking about these guys knew the scale that they, that's why they--
• So the government put in 100 million and basically said you gotta go raise another 100 million.
• Billion, right?
• Billion.
• Billion.
• Yeah.
• If you fact check me bitches. (laughing)
• And this isn't a software company that has like a five year and it was like in five years it'll be a 10x or 100x or this is million fact check.
• By the way it was 228 million, not billion. (laughing) It was not billion, but still, that's still a ton of money.
• That's still, yeah, yeah.
• So fact check me, I fact check myself.
• What? Anyways, they, the Phillips, Phillips of all people, Phillips, the company Phillips, I'm sure it's European, puts 28% and then they get the rest. So he gets the rest, he starts the company. Anyways, they essentially they start growing, they start growing their market share, but what they did is instead of designing and being an integrated manufacturing company, they just said. Pure play. What, and explain the, the foundry concept. So at the time, like Intel and AMD were designing chips and also building them. And it wasn't until the 90s when everyone started out sourcing manufacturing in pretty much every industry that they started outsourcing the manufacturing of the chips to Asia. But yeah, I'm going to just do pure play foundry and build chips. I'm not going to design. I'm not going to try. I'm going to be available to any designer and I'll do it cheaper than what Intel and AMD are doing it for the US. Do you think in the 90s, do you think when we did decide to start outsourcing a lot of stuff, if Intel and AMD, is AMD an American company? Yeah. If both of them decided, no, we're not going to outsource it, and we're going to maintain it in house, that would have radically changed the entire future. Because outsourcing it is ultimately what gave Taiwan all the power. Well, yeah, I mean, you would have also been requesting the US government to subsidize that. Well, but at the same time, we're now subsidizing it with the chips. You're assuming that they would have been successful.
• Yeah, yeah, yeah.
• And as one company didn't do that. Which one company didn't split fab from, or manufacturing from design, it's Intel. They still do it.
• Okay.
• And they're, I mean, they--
• What's be honest, they're getting, they're, they're, they're shitting the bed right now.
• But do they not also have a ginormous opportunity to kind of reinvent, if they can, if they can give past some of the dogma that probably exists within an organization that's as old as that and as entrenched as that, that it seems like, especially with the chips act and the amount of money that the government's starting to willing to give to get companies to do it, Seems like Intel is one of our best options as a country to be able to start producing these things in-house. So maybe there's some other players that are out there.
• So let's just diverge then off that topic. So the reason why we're saying that is because essentially what's gone on, and we'll get back to the TSMC store 'cause it's great and there's more to it. But essentially you have like seven really big-ass manufacturing facilities on the West coast of Taiwan. Okay, well, guess who's over there about 200 miles away, China. Okay. So, so there is a huge geopolitical aspect of this, which is if, if the manufacturing capacity was dwindled down to zero because of geopolitical issue, the entire world would stop on a, on a, on a topper or whatever the phrase is, right? Because the, all of that manufacturing is going on in this, in this area. People think that Taiwan is claimed by China as a territory. U.S. recognizes Taiwan as an independent entity, right? China and Taiwan have this relationship. And it's basically the reason why Warren Buffett sold his shares of TSMC. He's like, "I love the company. Unbeatable. No one can compete with this company that does the most important thing in manufacturing in the entire world. But I sold my shares because I don't trust China." right? Right. He's rather see that happening in the United States. And so, you know, we talked to one of our history buffs and essentially you had like the biggest concentration of troops on the ground from a US perspective anywhere in the world is all around there to try to secure this, these trade routes to secure Taiwan to secure basically what happened after World War II, right? They stayed. Yeah. Just to put into perspective how important Taiwan is from a manufacturing perspective, I think it was 92% of all leading edge or what do you call that? Edge technology chips are produced in Taiwan that the US buys. And then you said how many of the worldwide chips? 60% of the global market share of chips, which is huge. So it would be a not a good situation for either side if that manufacturing went down. Chaos, like industry would fall. Yeah, and we know that because we created, because there's so much value in this technology, it's embedded into everything. It's embedded into your toaster oven. It's embedded into your cars, you know, building into... We would go back to toasters without a Wi-Fi. Right. I want to go and get my Pontiac Thunderbird, right? And then I'm going to store it in my garage. So in the morning, I've got this simple engine that's not... I can only drive that. It doesn't have a Wi-Fi connection. I don't have Apple CarPlay. You know, like I, I, I, uh, I, my, my car, my, my, you know, your Tesla first world problems. My Tesla stopped charging, right? Yeah, it broke. It's not really. Yeah. But it comes with an eight year battery warranty. Okay. But essentially the other day I'm like, well, my car is not charging. It's broken. Like, and there's one, there's one repair facility in this, in, in Indiana, no one for Tesla. And, um, you take it there and you, you go on the app. The only way to make an appointment with Tesla is through the app. It's actually a lot of parallels there because one, you have less moving parts in your Tesla, so there's less likelihood to break. Same difference in one manufacturer. They can do things much more efficiently, right? But if something goes wrong, if they stop being able to produce it, if your battery breaks and you have to go to a Tesla repair shop to fix it, you got to wait. Things are going to be bad for a little while. And so literally, the first appointment available was September 19th.
• Shut up.
• And so I'm like, well, I'm gonna go ahead. Yeah, it's currently August 16th. This was August 10th, probably. And so I literally, I mean, and this is just, you know, advice, you know, just show up. I just took it over there. And I was like, my car doesn't work. I need a car. So you're gonna have to go ahead and get me an appointment right now.
• Right.
• And you were like, yeah, we will. (laughing)
• Nice.
• Which, and they did a great job. Thank you, Tesla. But the, like, the, the, the, the parallels are rear up.
• Yeah, it's all right.
• 'Cause it's all sourced. The Bob's Auto Shop can't fix a Tesla. There's one company that's an integrated supply chain. So you're kind of, it's like, it is a great analogy. I don't know how to say that came up.
• So back to why is there one company doing all this work? Right, so anyways, you're back. So they start doing awesome. Chang retires in 2005, okay? Well, guess what happens in 2008?
• The whole crash?
• Yep.
• Yeah.
• But what else happens? iPhone.
• Uh, uh, 2007. - 2007.
• 2006, 2007. - Yeah.
• 2008, Crash. But at this point, guess who's making all the chips for iPhone?
• I guess you turned it down.
• Who? - Intel.
• Intel turned it down? - Yeah.
• Not enough margin.
• Jobs went to Intel and said, we're going to make this computer phone. This is like before it was even released.
• Oh my god.
• Do you want to be the chip manufacturer? And he's like, I don't know. He's like, we got Intel inside. They got, you know, it was AMD and Intel going back and forth still. They were turning it around on computers. Yeah. Yeah. So huge, huge mistake. So they just, so. But at the time to put that in context, the personal computer was everything. Yeah. No one had anything. The closet, the computer in your head didn't make sense to anybody yet. Yeah. Except for the guys building Furt apps, making millions of dollars. They're totally undiminated. I am a well, someone who went to my high school that made a bunch of money with that app that was just ringtones. It was early on, right?
• I remember my app actually made me some money.
• Did it?
• Yeah.
• What was it?
• It was just called pee. People that have pee shyness, and you could just drag a thing down and it would just sound like you're peeing. So you had the sound of peeing and it would make you pee.
• That's awesome.
• That's all it was.
• That's all it was. (all laughing)
• And all it was is because I was gonna be working on a native app and I'm like, I always have to kind of go through a first app just to figure out all the nuances. And so I was like, I'm just gonna throw it together. Yeah, how many sales did you get? I have no idea like literally I threw it up and then all of a sudden I started getting purchases and I'm like oh my god Cuz I could still have just my name. I do OP was the name of the app and Brandon Corgan was underneath it That was when I learned oh you need to set it up as a company and so you're dumbass is that your name associated It's funny that almost every week you just bring up something you built
• And like him generating revenue off.
• I'm like, what?
• I did, so then I'm like, "Ah, shut it down."
• All right.
• So Intel, I did not know that. Intel turns down building chips for Apple. Apple goes to Samsung, and Samsung's building, Samsung built the chips for the iPhones all the way through iPhone 5.
• Damn.
• So 2008, your boy Chang takes back over as CEO at like the age of like 78, okay? And I think he's still now the chairman at 91 or some shit like that. So he comes back and he's like, we got build these, we got to build these phones. We got to build these chips for these phones. So iPhone six, Apple comes out with their first chip. They design it. TSMC manufactures it. And that changes the trajectory of the entire company. They go to the point where they're doing $40 billion on 50% profit margins and making $20 billion of profit a year. Right. But now you have this situation where even to this day, All of Apple's chips are manufactured by TSMC. And Apple is 25% of their revenue. Damn. Yeah. I mean, and you do think like the the the the leaps that Apple made with their M series chips were just absolutely astronomical. And now you have not Intel. Well, who's who's doing another one? Is it the not as a firefly? One of the new chips? I think it's like the NVIDIA ARM chip that they're starting to use in the new Windows laptops. Have you guys seen those? Like were they really competing with the M? Why is this so bad? Well, we're going to clip that. That's going to be a good short. Well, but what's interesting is these new ones that they're really kind of competing with the M1s, I mean, those chips are looking really, really impressive. Like they're on point with the M series chips. And so I think that Microsoft is getting to that point where there's the new Surface Pro laptops are powered by these new chips and I think they are AMD based. They're not Intel or they're not. So when I click the start menu, it's actually going to open. Oh, yes. They're still full of ads. Well, one thing that Apple did, which was smart, is they started designing their own chips. They built an in-house chip design, so they didn't need to rely on AMD or any other designer. They could get the talent. So they go straight to TSMC. They don't have like a designer, they do them. Whereas Windows has to make, you're talking about a lot of multi-purpose, like general-purpose chips, right? They're not designed just for one piece of hardware. So it's hard to compare, you know, it's a little different.
• Right, and that's the reason, that's the fundamental reason why Apple was always so successful is Steve Jobs is like, "Hey man, we're making these products, "and I need to understand this product, "and I'm gonna design this product." I'm not a software company.
• It was always a product.
• Just like what we're seeing with the AI and the rollout of the AI stuff, the AI here is more of just, it's a feature based AI versus being like,
• It's a new chipboard application, right? Like we're incorporating the features of AI into the overall product experience.
• That goes back to the premise of why they may or may not be really successful and what they're doing. But for those who haven't experienced the leap from the Intel chip to the M1.
• Oh, it's, it's easy.
• If you bought an Apple device in the last three or four years, essentially, and you got one of the M series, I mean, there's a noticeable difference in performance battery life.
• Oh, it's ridiculous.
• How many times have your fan, remember our fans used to turn on?
• You get all the time.
• I mean, my battery life is way better.
• Oh yeah, you go all day with these, with the M2.
• Yeah, I'm even, like you used to have anxiety about not bringing your charger with you places. I haven't brought my charger with me. Now, part of it is it's a new battery and my other computer was maybe a few years old. But even then, compared new to new, if my memory is correct, it's way better. So basically we got to this point. And this is before the last couple of years, it's like you got to the point where those M chips, basically you could just do whatever you wanted and you were overpowered. Like you had so much computing power on your device that whatever you needed to do, you could do it. You didn't have to close out of applications. Like you're going to run Docker, you can run, You know, you do all your dev stuff, but now we get to this AI world and the demand for this technology now and the demand for improvement, like you said three nanometers or whatever, like that's the question now is like, okay, can Apple continue to design, can Taiwan continue to manufacture or do we actually need to consider the fact that the current architecture is probably may or may not be able to even support the what we're trying to accomplish?
• So I've actually got a question for you very specifically about this because I have a feeling of what happens in this space and it just shakes all the traditional manufacturers to the core and that's gonna be the inclusion of biology, right? That's gonna be including, there's the brain on a chip, stuff that people are working on where they're trying to graph like, you know, brain neurons literally on chips. There's ones where they're using brain cells as part of like a nano, you know, chip thing. And what they're finding though is that anytime that we're starting to incorporate the biology side of things, all of a sudden our efficiency goes through the fucking roof, right? Because our brain like runs on like 20 Watts of power, right? It's crazy how efficient Mother Nature has been. And I think the companies that are ultimately going to start leveraging that kind of stuff is ultimately going to win because all of the old brothers who are still just pushing electrons around are not prepared at all for the biological side of things. Are you guys seeing anything? are you hearing anything? Like what's happening?
• I think that's right. I mean, so there's, you've heard of Moore's Law, right? Gordon Moore, co-founder of Intel said every, about every year, the number of transistors will double. And that was true for decades. But I think that we are hitting like that, that like diminishing returns.
• More on them all, almost.
• Yeah, that's what they said. There's like, you'll get into quantum tunneling if you get past two nanometers, which is where we're almost at. So I think that, I think Moore's Law,
• Yeah, Moore's the hell is quantum tunnel.
• Oh, I don't know. All of a sudden you're starting to pull resources from other universes for your power, your computer and shit.
• So I think if Moore's law is slowed down, and it's, I think that's right, in existing architecture, existing like kind of format, it's probably, we're probably getting there, but they have to reinvent the manufacturing process to keep continuing innovating at the same pace. So I think there is gonna be some kind of disruption at some point.
• So there's this, and I keep trying to find the name of it, but there's specifically a company and they're this startup, but still well-funded, that they're literally using stem cells that are this kind of humanoid, not nodule cell thing. They have these two little black eyes on it. I mean, it's wild, man, and that they were able to go and process like a huge amount of data. They've got an API that's available that you can actually use. It's super crazy expensive. And they have to worry about these things like dying and replacing them, but they also get smarter because they're using a lot of the same brain synapses and stuff like that, the more that processes, the better it gets. And you can actually see it's training. And these things are actually getting better. I mean, that is wild.
• I think the issue with those, well, long term, yeah, maybe. But the issue is just like meeting the demand today. So like the H100 NVIDIA's chip to the H200 is like 40% more efficient at training a model and 20% less energy and Intel's got one. And they're going to keep re-architecting and re-designing these chips to be more efficient to do the task, these very specific tasks of training these LOMs, which is relatively, a relatively new task to do at this scale, right? So while I do think like, that's like a two or three decades down the road.
• Oh yeah, absolutely, yeah, yeah.
• Yeah, yeah, yeah.
• So we're trying to think in the near term, you're just gonna see--
• Yours is the signal.
• Manufacturing chips there. We, you give me an M3 chip and I can write code all day long.
• Sure. - Right.
• Right. - Give me, I can do Adobe Photoshop, right? I can send emails, I can have 70,000 tabs open and Chrome. I can push the edge of the envelope, but the computing demand that we're headed to is beyond any scale that we could ever comprehend. And so that you're gonna hit that quantum tunneling limit, right? And so what happens is that you just create a demand for something that's never had demand, right? And so what do humans do? they build shit and they create it.
• And what's the best thing to do with that too?
• They're not gonna be able to build a TSMC factory day one to manufacture that at scale, but the technology is coming that will disrupt that entire concept.
• So here's what I'd like to propose, is that what we could do is we could actually start just creating a bunch of humans. We could just keep them in a vat, keeping kind of just plugged in, right?
• I know where this is going.
• And we could use their energy. Basically, almost like a battery in a way. and we could use their brains as a processing unit. Perfect. Right?
• Is that what they were doing in the matrix? (all laughing)
• But I think that's a great point about the next generation of technology. And if you look at the ChIPs Act, it was $280 billion over the next 10 years, but only 50 of it was for like pure play manufacturing. Like Intel got 10 billion, we got a couple billion. 200 of that was for like R&D sciences, making sure that US students and engineers are inventing that next generation. I think they're already seeing like TSMC's got this generation of chips and it's yeah, we're never gonna catch up But like what's that next wave and how can we protect that? That's a good point And I wanted to talk about the issues in the Arizona factory for for TS. Yeah So so is so TSMC diverse all their stuff's in Taiwan. That's where they do alert, but they diversified to Arizona Yeah, well, I mean we paid them to The airship's funny. They got six billion in US funding.
• Yeah, yeah.
• Two basically go set up a plant in Arizona.
• Three plants.
• And they got there and they're like, this is really hot. (laughing)
• Yeah, they're like, it's too friggin' hot. Why are we here in the middle of the desert?
• Well, you're making it pretty hot.
• For manufacturing, you want to have no access to water and cool weather. So why they pick Arizona?
• I mean, what is it?
• I don't know.
• You got a last, you guys.
• So literally, just, wait, hold on. What are the issues in Arizona for TMC?
• Yes.
• TSMC's manufacturer. So one of the issues is culture change. So the Taiwanese are very hierarchical and basically they're like asking these engineers, these manufacturing engineers to come in like at the drop of a hat in the middle of the night and us Americans were like, (laughing) "I'm banging on my job description." So they're just having issues getting them to adhere to their processes and how they do things. And also they're asking a lot of Americans to go over to Taiwan for extended periods to watch how they do things. So I think the near-term solution is they are bringing over a bunch of their engineers to fill any gaps and to help smooth over the process and then train hopefully to sustain this 'cause you can't have an--
• Wasn't there a movie in the '80s with Michael Keaton? That was this exact thing. Oh, someone, we need to look this up. I swear they're-- - Trying to dangerously?
• No, no, but there was. And it was like, there was a Japanese company came over Michael Keaton was in this movie. I'm totally having a flashback to what you just described was exactly a story.
• Well, I'm gonna look at it.
• I'm gonna look at the Michael Cone jacket and make it a factoring movie.
• The unions don't like this 'cause they were like, "Hey, you said you were gonna employ Americans." And now you're bringing over a bunch of non-Americans.
• And they have funding for it from the Oscar for it.
• Yeah, so like, yeah, it's a little bit interesting to where that's--
• It's like the farmers. And have you seen these YouTube videos going around right now where it's like farmers in Montana and like other areas like out in Utah and like kind of like where you're like, are there farms out there? (laughs) And they're like, okay, so here's why we have to employ PEEP immigrants because Americans that we've offered them benefits, $20 an hour, 401K, insurance, retirement packages, and guess how many people show up for the job?
• Yeah. - Zero.
• Yeah. Ain't no American that's in a fourth generation sitting on the couch lifestyle want this job. 'Cause it's hard.
• You said it earlier, Americans, we are very like, we're creative. We like to be our own bosses. We're pretty level playing field. There's not a, I mean, there's hierarchy, but not a ton of it, right? It's a pretty, pretty flat structure. Other countries are not so much, right? Like Japan and Taiwan, and they have pretty hierarchical culture. So it's like, yeah, that makes sense. You go to work, your boss tells you what to do, you do it, you need to come in the middle of the night, you do it. The other issue that they're having is these clean rooms. Like we just don't have a lot of experience.
• We're dirty.
• We do a lot of dirty manufacturing. We used to, that's where our roots came from. They're like, as far as like clean room manufacturing, we just don't have a ton of experience with that.
• What does that mean? What does a clean room?
• That just means you literally are, you've seen the--
• The ground road.
• Yeah, the head, that you just absolutely no dust dirt. Right, nothing can come into that room because these wafers are in there, I'm assuming.
• Yeah, I'm getting a little piece of dust.
• Where do you think, where do you think that? So that's part of what's kind of wrong with America. Like, where do you think the grit and determination from people and these cultures comes from? Is it from their parents?
• I think it's very much cultural for them. For us, it's like that individual cowboy, you know, like, we're American, we can go do this on our own and blah, blah, blah, blah.
• But also, that's how it always was, right?
• Yeah, well, they were in World War II, we did a bunch of crazy shit because we were like, we're gonna die if we don't do this. And they probably have, we just don't have the driver underneath of like, you're gonna die if you don't figure out how to do this tough job.
• China's having that issue right now where I think that they're like second third generation, factory worker family, like their kids are out on TikTok and wrapping and they're just not feeling like, I'm not doing this. I mean, my friend went to college in Kentucky and loved it and probably and came back and was like, man, you don't have to live like this.
• Like when, I don't know if you, do you guys notice it too though? Like during the Olympics, any time that you had some of these like Asian countries that were going out and they were like in a group, those people were in lockstep. It was insane. Like just how well they were, they were a fine well oil machine. They would all come out, they'd be like going in like robots. And I mean, they were just completely in. then you had Americans coming out over and just like, "Pee, I like their own individual rock star." You know, all this stuff. It was, it was kind of wild. And well, there's pros and cons to that, right? Yeah, yeah, yeah. When you're talking about manufacturing, where it's like, you need to be robotic. Actually, robots are better at this. That's why they've replaced a lot of those jobs, but you have to be robotic. You have to do exactly what you're told all the time. Yeah. So the every reason that they move manufacturing dares on it is because it to diversify their geopolitical risk. Yeah, it had to be.
• Yeah, it's not a financial, and it's a long-term stability move, not a, it makes no sense.
• It's not gonna work.
• No, well it might, Gung Ho was the movie by the way, but Michael came 1986.
• So that's gonna tell you I was 10 years old.
• Shawn, it'll work in the sense that it's gonna take a while for them to acclimate, and they're gonna have to learn more about how we work, we're gonna have to learn more about how they work, prices may have to come up a little bit. There's gonna be compromises to me, because it's worth it, because why? We don't want to be 100% dependent on Taiwan. So if-
• But we have to. I mean, like as a country, we have to not be
• Yeah, 100% a lot.
• And that's correct.
• You say 200 billion of that money is going towards research development. And I'm sure there's a clean room labs getting set up in the high schools and colleges across, you know, that's just gotta be a part of it.
• But what you guys are just saying brought something up. Like who do you think ultimately is responsible for the situation we're in today? Is it, 'cause Apple knows they're getting all of their chips from one single supplier in a geopolitical hot zone on the ring of fire. They know there's risk. Are they actively trying to diversify their supply chain? I don't know, it doesn't seem like right now.
• Whoa, whoa, whoa.
• They have so much money.
• Warren Buffett sold all their shares in TSMC and Apple.
• He did sell all of his Apple.
• Oh, sorry, 50% of their Apple, which was like they owned more shares than any insider Apple.
• I will say what Apple has done an impressive job and being able to do some pretty dramatic shifts, right? Like just being like, all right, we're killing Intel, we're gonna go with AMD, right? And we're just gonna make that move over the next two years. So they do seem to have a pretty viable ability to make those types of changes. So if they had to, I think, you know, where like we were talking about with Windows, you know, they have to serve so many different masters where Apple really is just serving a bunch of hipsters, right?
• I wanna learn more though about like Apple another big-text role in disrupting the supply chain, because you're right, for pure profit play, yeah, of course they should go to the biggest manufacturer in the world who can do it at the lowest cost and they don't have to worry about it. But other than it'd be good for America and it'd be good for global stability, could they make a argument that they could make more money if they controlled go the Intel route?
• I just think it's more like, they're clearly sourcing all this stuff from TSMC and from Taiwan because there's demand, they're making tons of money, they're trying to push product and ship product couldn't advance and innovate. But no, I don't think, I don't know if they've stopped and thought like, I don't know, it's just like all those questions. Like who's the simple answer to this question? There's a simple answer to this question. No one can do what TSMC can do. No one, not even close, can't do it, can't do it, can't do it. Like if Apple wanted to build the M4 chip and ship all these iPads, and they just couldn't do it without it.
• Is it a similar?
• So they have to take the risk.
• It's a similar concept to what Tesla did with the car industry, right? Like, but only maybe even more entrenched, harder to get into. 'Cause like car manufacturing Toyota has owned car manufacturing. Like the big car makers, they've been doing it for a long time. And they have worked out a lot of the kinks and their precision is unrivaled.
• Good point.
• Tesla's precision.
• It's not really good.
• And questionable.
• And they're not diversified. You're right, and Toyota, Toyota diversified their supply chain. They got factories in America.
• Well, let's go back to their TSM season stop a little. For now, yes. But I think they're, you're seeing, so ASMR we talked about makes those extreme ultraviolet lithography machines to make the AI chips. They make five of those machines a year. That's all they can put out, and they cost like $400 million. But guess which company bought all five of the 2024 machines?
• Intel. - Intel.
• Yeah.
• So what we're saying is, is everyone just came home from the, from 20 years of just absolutely stacking money and Lux, Lux, I go, okay, wait, what's going on?
• Exactly. - $40 billion of property year. All right, let's re-think this, let's re-invest.
• It was the chip shortage that woke everyone up.
• Yeah, everyone, it's an awakening. There's an awakening happening.
• Your house gets on fire, should I get interest?
• Yeah.
• No one even probably thought about some of my conductor supply chain until they're like, I can't get a car for how long?
• Yeah, because I can't, but the radio's not ready.
• I need chip. I still, I still, I think my biggest struggle with this entire thing is thinking that you're going to grow American manufacturing. I just don't think it can happen. I think we're literally France. I think we're Spain. I think we're a fourth, fifth, sixth generation, working from nine to two PM, you know, take a Siesta culture at this point. I think that the Americans have, that are called, I think we're all fucking mailin'. I think everyone's fucking mailin' it in. I'll counter that argument with. (laughing) If we really wanted to invest the money into doing the thing, what you're saying is we don't have the manpower at the cost required to do it the way they are currently doing it.
• I don't like it. - We innovated.
• That's what I'm saying.
• Yeah, and I agree with you. I think it would go ahead.
• I was gonna say, if we lost the power grid, we lost the power grid. I think all of a sudden American would step up and our cowboy roots would actually pan out for us and we'd be willing to get dirty.
• They did something to galvanize.
• We're so covered, right?
• But what is it?
• What is it?
• It's oil.
• What is it, is it oil?
• You're right though, that's, so what it all comes down to, the whole chip sex really, probably it's economic and being more competitive, but it's ultimately about national security.
• Yeah.
• Military weapons, power grid, being ahead, so that if anything does go bad, we're not relying on Asia, China, South Korea. That's, I mean, literally, that's probably the White House and the capital they're thinking about.
• I really do think though that as a country, we are one catastrophe away from being united, right? And that it's going to suck. It's going to be painful. People are going to hate it because again, we like our, we like our overpriced, you know, big league coffees or your venti, whatever the hell that thing is. But we like our cushiness, we like our comfort, we like all that. However, we have shown that Americans are absolutely willing to not be comfortable, willing to build some amazing shit and willing to go forth.
• I agree. And also all the political shit that goes on, all the noise. - All the noise. I look at it and I laugh sometimes, 'cause I'm like, if one thing, if we had to unite, you know, when the tower's got hit,
• Right.
• All of a sudden everyone's just like, same page.
• Yeah, totally.
• Let's get, I mean, we maybe made the wrong choice, but we still like immediately aligned.
• I hope that that's true.
• Right, I do, I think. Either that or we just--
• It sure was in 2001.
• It was. We just need, maybe what we need to do is we need to allow, we need to basically airdrop walls and lions across all major cities, right? Just cover them and basically make everybody have a fear for their life and their safety. And then all of a sudden we're all unified.
• Did you see that?
• If we're all running down the damn street from a line, guess what we're not yelling about? Oh, it's Trump or Kamala, right? We're just running from the line.
• Did you guys, so what was, I sent you that video on Instagram. So there was this guy that does free, what's it called when you climb, you climb?
• Pre-Solo, he's a capital capital.
• He's got the time.
• So he's on his podcast. So free solo, he basically climbs the hardest. Alex Honal, but yes. Yes, yeah. The guy's got the biggest eyes. He's like an owl. He's crazy. But he's on this podcast, and I send a joke because it's a motivational thing. So he's on this podcast and he's like, people need to fear for their lives. They need to fear something. Because if you don't have visceral fear in your day-to-day life, or maybe week-to-week or month-to-month or year-to-year life, what happens is you start sweating the small things.
• Yeah, totally.
• And he's like, "I think that fear, danger style fear "is one of the more important emotions "that you can experience."
• Yes.
• And he's like, "I'm at an airport "and people are worried about the delay in the three hours." He's like, "You have abundance running, "you can go to McDonald's, Burger King, "you have a bar, you have your computer, "you have, it's air conditioned, "and you're sitting here worried about a three hour, "like who gives a shit if you're three hours like?" He's like, "Go experience the fear of if I lose your grip, I'm going to die. Yeah. Everyone freaked out about the whole crowd strike thing. Like, I couldn't order my Starbucks ahead of time. I was like, this is every store. I was like, something's wrong. Like, so mad. So, you know, and maybe governments understand that. And there's obviously something that needs to happen. But there is this paradigm right now where our culture is not necessarily prepared to kind of get their cans dirty and build the things that we need to build because we realized that we're building it in a place where all these people have made generational sacrifices to do these things and build these things. And--
• But he's the catch-22, right? Because they did that so that the next generation did have that.
• It's a hard time.
• We are the next generation.
• Hard times make hard, man.
• Yeah, but whatever that whole kid uses.
• You would never wish a bad thing to happen to someone unless you're bringing in any one.
• I don't want them to run, I don't want them to get eaten.
• Okay, okay, so, but like, yeah, manufacturing fear, you know, which we already kind of do, but for the wrong kind of shit, right? Like that's not really, but I hear what the point is, is like, we do need to figure out like, all right, what can we rally around? What is, what actually matters? I think the other side of that, aside from getting yourself out there and doing things that challenge you, maybe it's, maybe it's dangerous, maybe it's just something that's outside of your comfort zone, go experience new things. Like people are so stuck in, and we're all guilty of it. 100% we have our comfort zones. Like that is what, if we're gonna get unstuck in this world of like, I'm worried about the, the transgender taking a piss in the backlight. (laughing) Like, oh, that's really follows me. Oh, no, we better get that on a show. And I'm like, we gotta go, we gotta go learn a little bit more. We gotta experience this shit.
• 100% dude. - 100% yeah.
• Well, I think that sums it up pretty well.
• Yeah. So, you know, we learned that today that Brandon built a P app.
• No, no P anxiety. There's a new store too apparently.
• Oh, I'm sure there is.
• I thought it was gonna be an app about finding the nearest toilet.
• Listen, that's actually not a bad idea. Like the best clean toilet that you can find.
• Someone in New York did it. It was like for pooping or whatever.
• Right.
• Yeah, but yeah.
• That's my favorite thing in sign felt like, George and you were all the greatest spots to do number two was, I think it might have been Kramer. I'm like that definitely came from real life experience. But if you didn't understand how brittle this entire universe is, I hope that today you learned something. And there's opportunities galore. And if you're young and you're in college, Go get a little degree in electric engineering.
• No, dude, totally. Go learn about how, what, learn about AI.
• So learn about neural networks. Learn about, go and let like learn.
• Right, so let me just throw this out there, that there is right now a, the National Electricians portion of America, whatever, the NECA. They basically have to hire 500,000 electricians over the next five years, right? And they're having a real hard time finding electricians. So if you are a college kid right now, or if you're just not exactly sure what you want to do, consider doing an apprenticeship at an electrician place. You're going to get paid. They're going to pay for your education. You're going to learn a valuable skill that's not going to be replaced by AI anytime soon. And that's going to be, I mean, you're like electricians right now, like when you become like a German or even a master, I mean, you're making like 150, 160, 170 and you're making good money. There's huge opportunities to like, I think like the trades were looked down upon for a long time. have to go to college or else you're just not good enough. But think about all the things that are dependent on electricians. All of our infrastructure. You want a new building? You want a substation? You want to connect this shit? Got a call in our electrician? And they might even eat the green stuff. It still requires solar and wind and all of it. Solar panel, you want to put solar panels in? Exactly. And to tie that back to the semiconductor industry, all these companies are building these fabs because of this government funding. And there's a huge labor shortage. There's not a lot of people that really are experts in building high-end manufacturing plants. So you're bringing people in from Asia, you're paying them double the rate. They're sort of just moving around and it's just--
• I got bad news for everybody, but it's also good news, which is this is a reflection of reality. There are so many people chasing blue collar entry-level jobs right now that went to college to go be a blue collar finance marketing person, even even people like us, lowly developers, right? And honestly, it might be the wrong move. Like if you're in high school right now, go to Mackenzie, go to the Trade Center and get a skill because I think that that might be a better found a use of your time to get foundational skills than you can see.
• You're gonna consult team.
• Because blue college jobs are gonna be done by AI in 10 years.
• Yeah, yeah. And if you're gonna do it and you have a lot of drive, go get a degree in entrepreneurship too, then add what you're at.
• Exactly, they start a plumbing business.
• If you could get the skills of being electrician and then five or 10 years after you've learned the trade, go start a business, you'll probably be a millionaire by the time you're 40.
• How many unhappy carpenters have you ever met? (laughing)
• That's true.
• You know what I mean?
• That's a really good point.
• Building stuff is just that's fun as to-
• It's so good.
• My grandfather was a carpenter, my dad was a carpenter, but he doesn't account. like building stuff and sitting and now doing all that. And it's just the most fulfilling thing in the world.
• People always make fun of me because I don't like to hire people to do DIY projects.
• He literally will hire no one. He will be about that.
• I like to build everything. It takes me twice as long and it's normally not as good.
• We just justify what, exit everything out. We just just do it. (laughing)
• It's just gonna find all those contractors.
• No, but I did, I have a ton of projects and I did hire a bunch of contractors 'cause even some products are too big for me.
• But it is a lot of fun to work with your hands. I'm like, I think the issue becomes like if you're doing the labor stuff and it's hard on your body and that's not a long term. There are some long term issues there.
• All right, America, step up. It's not gonna be me, but hopefully I do.
• Next, Jen, please.
• This has been the Big Cheese AI podcast. We really appreciate Joel for hanging out today.
• Yeah, thanks for having me.
• And just side note, Joel is the drummer in our band Prestige Worldwide. So look us up on Instagram Prestige Worldwide Music.
• That's right.
• We're trying to get to 100 followers before we apply the tonic ball. So hopefully you can see us.
• What are you guys?
• We're in 67.
• Yeah.
• Jill is the head of, what is it?
• That's a week.
• That's a week.
• That's good.
• White leather gloves.
• Black leather gloves. You guys obviously haven't seen stepbrothers. (laughing) And yeah, we need to get to 100. So we got rejected last year. But yeah, we appreciate you guys coming out. Check us out, like us, like, subscribe on YouTube. We need to get to 500 subscribers as soon as possible. Thank you guys.