What We’re Reading (Week Ending 25 September 2022) - 25 Sep 2022
Reading helps us learn about the world and it is a really important aspect of investing. The legendary Charlie Munger even goes so far as to say that “I don’t think you can get to be a really good investor over a broad range without doing a massive amount of reading.” We (the co-founders of Compounder Fund) read widely across a range of topics, including investing, business, technology, and the world in general. We want to regularly share the best articles we’ve come across recently. Here they are (for the week ending 25 September 2022):
1. How Much Do Interest Rates Matter to the Stock Market? – Ben Carlson
The 3-month T-bill was just over 1% in 1954 but ended the decade at more than 4%. During this time frame, the S&P 500 was up 21% per year or more than 210% in total.
Short-term rates nearly doubled in the 1960s, going from a little more than 4% to 8%. The 1960s weren’t a great decade for the stock market but the S&P 500 was up a respectable 7.7% annually. Close to 8% per year is not bad during a time when interest rates doubled.
In the 1970s, short-term yields went from 8% to 12%. Nominally the U.S. stock market did okay in the 1970s. Stocks were up 5.9% per year even as interest rates were breaking through double-digit levels.
The problem is inflation was 7.1% so stocks were down on a real basis. And that’s the biggest difference between the 1950s, 1960s and 1970s. While inflation was more than 7% per year in the 70s, it was just 2.0% and 2.3%, respectively, in the 50s and 60s. So while the real returns were spectacular in the 1950s and pretty good in the 1960s, they were awful in the 1970s.
You can never gauge the markets using any single variable but if I had to rank them in terms of importance, inflation would get more first place votes than interest rates…
…The stock market doesn’t do nearly as well when inflation is rising and it does really well when inflation is falling (on average).
But when it comes to interest rates, there isn’t much of a discernible pattern. I know a lot of people would like to believe falling interest rates were the sole cause of the entire bull market in stocks from the early-1980s but my contention would be disinflation was a bigger catalyst.
2. Frameworks v0.2 – Chris Paik
All successful consumer-facing companies appeal to one or more of the seven deadly sins. They are time-tested core motivators that incentivize people to do things (the fact that they have survived for all of time without any edits is proof of their power). There are no successful consumer companies that do not appeal to any of the seven deadly sins.
Different motivators can apply to different constituents within each company, even different behaviors from the same constituent.
Examples:
Sloth: Uber, Amazon
Pride: Instagram, Tik Tok
Gluttony: DoorDash, Netflix
Lust: Tinder, OnlyFans
Envy: Pinterest
Wrath: Twitter
Greed: Bitcoin, Robinhood
Sloth tends to be the easiest to monetize because the end user places a fairly consistent price on the trade off between money and time (convenience). Pride is also easy to monetize—see framework on elasticity of demand and Maslow’s Hierarchy of Needs. Gluttony is straightforward in its monetization as it is rooted in consumption. Lust, while easy to monetize, has historically been confused with long-term mate finding, a seemingly impossible atomic value swap that has plagued dating websites. Envy is slippery to monetize, as the challenge is how to own the point of purchasing decision (the path from inspiration/envy to a monetizable transaction can be long and convoluted). Wrath is a very difficult sin to monetize and often manifests through in-group/out-group dynamics. Greed is the hardest of all sins to monetize, naturally so as the user is loath to engage in a sub-optimal transaction, and will prefer to be monetized via any other sin (i.e. sloth in the form of performance fees)…
…Market risk is where the demand for the product is unknown. Execution risk is where the demand is well understood, but the hard part is in the delivery of value against existing competition. Any company that is pure execution risk without any market risk is not a suitable venture investment.
Examples:
Instagram, Snapchat, and most of what we consider “consumer” companies neatly fall into the market-risk bucket. No amount of money spent on a customer survey or consulting project would yield the conclusion that there is an opportunity. It requires an explorer to basically set sail with conviction and strike land.
An example of execution risk would be becoming a franchisee. There is well understood demand for the product, but delivery of that product to demand is not so simple. Opening a new location in a new city or country would be some small amount of market risk, so near pure execution risk would be opening a franchise location a few blocks away from another one…
…The reason why Charli D’Amelio is the #1 Tik Toker is because she is a dancer. Tik Tok is the first platform to incorporate audio as a native part of the core product consumption experience, and thus the first opportunity for dance to be properly appreciated as content.
JFK would have been less advantaged in his presidential campaign without the invention of the television. Similarly, FDR would have struggled in a post-TV era.
Another way to interpret this framework is that a new content network that aims to poach top users from a pre-existing one will fail. Success stories in every new network will be homegrown—the opportunity/switching cost for would-be emigrants established on other platforms is too high…
…Treating supply as a commodity is the core philosophy of all marketplaces. This leads to supply competing to serve the firehose of demand with that competitive dynamic translating into a consumer surplus. Treating supply as unique is the classic “arm the rebels” approach which requires the supply to think of themselves as non-fungible. Supply that believes itself not to be a commodity will invest in products that allow themselves to differentiate from competitors, capture more margin from their customers, and avoid being platform dependent.
Examples:
Uber is an excellent example of supply as commodity. Interestingly enough, I explicitly wouldn’t want the same driver as I had last time because in all likelihood, they are very far away when I need them. Shopify is a great example of supply as unique. No company using Shopify thinks of themselves as a commodity, which is exactly why they invest the time and energy into setting up their own storefront rather than plugging into a marketplace with pre-existing demand…
…A Swiss Army knife is very useful when you are space constrained. It is less useful when you need a dedicated screwdriver to assemble a room full of furniture. Similarly, products with a generalized value proposition will inevitably be cannibalized by more specialized competitors. Convenience is the only defense generalized tools have against erosion by specialized tools.
Examples:
Very famously, Craigslist as a generalized tool has been competed against by more specialized tools in each of the classifieds categories. Similarly, over time, eBay has been cannibalized by competitors who are focused on a specific vertical within eBay (i.e. Poshmark, Bring a Trailer)…
…A very effective strategy to unlock potential energy in what may seem to be a calcified ecosystem is to do something that the existing, entrenched players deem to be completely irrational. The conceit in this strategy is that while the behavior may seem irrational at the first order level, it is rational at the second order level and often leads to a market leading position if not monopoly.
Examples:
Credit Karma is a perfect example of this strategy. When it was founded, the credit bureaus all made very good money by charging for credit reports. Whether by corroboration or complacency, Equifax, Transunion, and Experian neglected to rock the boat and were content in their business model of clipping coupons from customers paying to access their credit reports. By offering credit reports for free, Credit Karma employed a strategy that on face value seemed entirely irrational to the credit bureaus—why ruin a good thing? But Credit Karma was after a much more lucrative pot of gold—financial referrals. Banks and credit cards would pay hand over fist for new customers and Credit Karma had them in spades because of their strategy. This was the second order optimization they were playing for and it worked beautifully.
3. Jay Goldberg – AMD: How Chips Are Changing – Patrick O’Shaughnessy and Jay Goldberg
[00:08:37] Patrick: Let’s zoom all the way to today, maybe describe the industry map, if you will, and we’ll come back to some of the history too to figure out how we got to here, but just to set the stage on how big this has all become. The semiconductor shortage over the last couple of years was one of the big geopolitical headlines, it’s become something that’s an essential thing that fuels a lot of what the world does and how we spend our time. Maybe just talk about that industry map today, what the major functions are and who the major players are in the production of the semiconductors that fuel all this stuff.
[00:09:10] Jay: To get at that question, we need to do some level setting, because there’s a few basics of the industry that are going to play through the rest of the story. The first is this idea we’ll call Moore’s Law named for Gordon Moore, the original CEO of Intel. And the idea here is that every 18 months, the semiconductors that we can manufacture double in performance or, more formally, transistor density per chip doubles every 18 months. And it’s almost a miracle of productivity growth, not just for electronics and semis, but for the whole economy. We’re at the point now where the smartphones we have in our pockets have more compute capacity than all the computers on the planet in 1950, something crazy like that. We have super computers in our pockets that we can literally take for granted because they’re not even the most advanced things out there. There are even more incredible systems in the cloud. That’s all because of Moore’s Law. For a long time, every 18 months, chips are just getting better, and better, and better.
Unfortunately, that’s slowed down. I said 18 months a few minutes ago, now it’s more like three years, four years. That pace of innovation has slowed and that leads to the second part of this is there’s a corollary to Moore’s Law that people don’t always talk about, which is the cost of manufacturing those chips increases at almost the same rate as the performance gains. It’s a little bit slower but the cost of building a manufacturing plant for semiconductors, we call it a fab, fabrication plant, a fab. Cost of building a fab today at the advanced manufacturing process is $7 billion for a building and a bunch of equipment. $7 billion is a lot of money. Over time, that led to another branch in the industry. I talked about in the ’70s, we saw a split between people who made electronics, people who made chips. In the ’90s then, we started to see a split between people who designed chips and people who manufactured chips.
If you think about a chip company today, the names that come to mind are AMD, Nvidia, Qualcomm, Broadcom. Those are all what we call fabless chip companies. They don’t do the manufacturing themself. They design the chip, an architect would drop a blueprint, only a few trillion times more complicated. They take that design and then hand it off to a third party who does the physical manufacture. And we call those third parties who own fabs, we call them foundries. Most chip companies today are fabless and they work with foundries. Now, there are a few that still make chips on their own, Intel, we call them IDMs, but most of the industry today is around this fabless foundry ecosystem. Those are the two pillars of the industry, this idea of Moore’s Law and this fabless foundry ecosystem. With that baseline, the market for fabless semis today is about 400, $500 billion, which is a big market. Then you throw in the foundries, and the equipment companies, and the software it takes to do the manufacture, plus other parts of test and packagings at the end of the process, that altogether is 8, $900 billion global market.
[00:12:04] Patrick: How does that shake out between what I’ll call sources of power? A big thing that’s been in the news obviously has been TSMC. It’s important, it’s the big foundry, it’s in Taiwan, there’s geopolitical concerns around that. Intel is building a big plant, I think, in Arizona. There’s geopolitical underpinnings of the foundry part specifically. Feels like the fabless designers have been pretty Western dominated, but what can you tell us about the state of play there and, in your view, on the importance of foundries being more localized? Maybe there’s going to be a reintegration so it’s not fabless anymore, it’s design and manufacturer. What do you think about the state of things? Because it’s really been a key news item in the last two years.
[00:12:45] Jay: And rightly so. The cost of building a fab has gone up incredibly high and the R&D required to not just build the building, but the R&D to be able to actually use the building and take advantage of it is immense. As a result, fewer and fewer number of companies have been able to produce at what we call the leading edge, the most advanced manufacturing process. To the point today where there are really only two companies that can produce chips at the most advanced process nodes, seven nanometer, that’s TSMC in Taiwan and Samsung in Korea. And Samsung, at this exact moment, August 26th, 2022, Samsung’s looking a little shaky. Nobody else can do that.
That is an immense choke point strategically for the supply chain and geopolitically because, again, all of TSMC’s plants are in Taiwan, and with all the tensions around that, that’s a big concern. The big question for the industry is can Intel catch back up? And we’ll talk about this later, but Intel hit a wall a few years ago and stopped being able to stay at the leading edge. And this is a company that had defined the leading edge for 40 years, they named the law after him, named Moore’s law after an Intel guy. They hit a wall 2015, 2016, somewhere around there, that had a really, really hard time advancing. And so this then becomes a big geopolitical question of can the US ever do leading edge manufacturing in semis again? Unfortunately it all comes down to Intel. At this point, no one knows whether or not they’ll be able to catch a backup again.
[00:14:14] Patrick: Why can’t we throw unlimited money at this problem and make it so that Intel or some other US based company, for political interests and national security reasons, can produce at seven and maybe five nanometer and beyond?
[00:14:30] Jay: If it were a money problem, we wouldn’t necessarily be in this place right now. Intel had plenty of money, and they’ve been generating huge cash flow for years. I think what happened inside Intel was partly organizational and structural. They made some bad management decisions. They made some bad capital allocation decisions that got them stuck off, that got them off track. Ultimately, yeah, I think if we threw a lot of money at it, we could solve it, given enough money and time and motivation. But it’s big sums of money. Just sort of put it in context, the chips act, which just passed in the US to subsidize the US semiconductor industry is $52 billion to be allocated over five years, to be spent over five years. $52 billion TSMs CapEx in 2022 alone is $44 billion. So yes, we can catch up theoretically, but we were talking really, really big sums of money here…
…[00:17:37] Patrick: Let’s talk about the difference between CPU and GPU and sort of general purpose semiconductors, which have dominated the history like a single architecture of the X86. Intel architecture has been dominant through history, NVIDIA is very famous for its dominant position and GPU’s. Maybe explain those two big horses, what they do, and why so much of the history and everything we take for granted that’s driven by compute rides on the top of basically two general purpose types of chips, so that we can then talk about how AMD has been at the leading edge. I think of that changing to some degree and how that might change in the future.
[00:18:12] Jay: Let me start off by saying general compute has been the pattern for the last 40 years. We’re now at a transition point. And that’s going to change. I’ll get into that. But in terms of where we are, to start with, CPU’s Central Processing Units are the main key chip inside of a computer, to have them in our laptops. That market for a very long time was dominated by Intel. They had 70%, 80% of the market. The rest was AMD. They had about 20% of the market. That grew really, really well until the dawn of the smartphone era when plateaued seeing GDP growth in PCs. In the early 2000s, we started to see the rise of GPU and they kind of do what says on the label, is they process graphics. A CPU can do graphics, but a GPU is dedicated special purpose to just do graphics. And that’s made it a lot easier to do video calls and play games on our computers. It’s become more important over time, and it’s becoming even more important as we go forward. The GPU market is split roughly between AMD and NVIDIA. NVIDIA is 70-ish percent of the market, depending on which category you’re looking at, AMD is the other 30%, right? That’s sort of starting point steady state where we were in the 2000, teens. That’s all changing. And I’m mostly going to talk about from the perspective of CPUs that we put in our PCs, our computers. There is another important category use of CPUs, which is in the data center for servers. And the server is very big, very powerful computer. And you line up racks of those. And that’s how you build a data center. You have a million CPUs inside of a data center, and those CPUs are much, much, much more powerful than what you have in your laptop.
That’s another fantastic market. But that has been dominated by Intel for 40 years, 30 years. If you think of servers in the data center, it’s an Intel x86 chip. It’s like a hundred percent. It used to be that we would say data center is basically just a giant room full of CPUs. That has started to change. And a big cause of that change is what we call AI, which I think is a misleading term. Really what it is linear algebra. And for a variety of reasons, you don’t want to use a CPU to do linear algebra. You can, it works perfectly well, since it’s fairly straightforward, fairly streamlined, it’s better to do it on a GPU. And as a result, everyone’s talking about AI today, chance are, if you are using some software online and it says, oh, our tax software is powered by AI. What that really means is that software is being run on NVIDIA, GPUs in some data center. And so we’re starting to see this shift away from a hundred percent CPUs in the data center to a lot more GPU. And then over time what’s going to happen or what’s already starting to happen is we’re not even use GPU to do AI. We’re going to use special purpose Asics that just do AI math. Google is the first one to do this very famously but they have another chip that just does AI math. It’s even better at doing AI math than the GPU. And if you want to play around with AI at home, you’re taking some course and you’re doing AI, your laptop at home will run that just fine. But if you’re doing it at Google scale where every search has to go through multiple iterations of AI algorithms, and it’s billions of people using it at once, you want your data center to have the most efficient way of performing that AI math, and that often is going to mean a special purpose chip.
So this is important because for years, we’ve just been making do general purpose compute, and it used to be if you had a computing problem and you’d sit there and go, wow, I have this weird corner case problem. And I really wish I could have a special purpose chip to just do this and be more efficient at it. I’m the acne box company. And I want to make a box processing unit. 10 years ago, I could have sat down and gone, all right, let design that my own chip. And by the time I could have actually gotten that into production, there’d be a new CPU out. And no matter how good my special purpose chip was, CPU was twice as good as the last generation. And I just throw more CPU at it. My problem solve, I don’t need my own chip. But if it’s not taking four or five years to see those performance gains on CPUs, suddenly making my own chip makes a lot more sense. And that is a big change coming to the industry. And we’re starting to see these data centers go from a hundred percent CPU to 50%, 60% CPU and 40% everything else. We call it heterogeneous compute, fancy term for mixing and matching different types of custom semi custom chips in with CPU…
...[00:31:53] Patrick: Can you say a bit more about this transition point? If we’ve gone 40 years with two general purpose workhorses that do just about everything and some of the things that made that possible, like the speed of Moore’s Law is slowing down and we just need ever more specialization, how far down does that go? And what are the most interesting ones that are happening now? You already mentioned the data center and the transition to GPU and ASICs and ASICs for specific AI workloads or something. How far can you extend this story, do you think, in a way that makes this industry look very different than a couple of major dominant companies, and instead becomes a much more fragmented set of designers, all designing around specific use cases?
[00:32:35] Jay: I think the best way to view that is to think of this industry as moving with a pendulum. I talk about, in the fifties, how everybody was vertically integrated, and then we swung all the way to the point where everything is completely abstracted. We’re starting to move back. Not necessarily for manufacturing, that’s still going to be specialized. But we’re starting to see a lot more non-chip companies make chips. Some of that’s electronics companies, Cisco makes their own chips, Western Digital makes their own chips for electronics. But mostly what I’m talking about is internet companies. The seven biggest data center companies, the Super 7: Google, Facebook, Amazon, Microsoft, Baidu, Alibaba, Tencent. Those companies are all designing their own chips. And it’s very interesting. You asked what’s one way to look at it. To me, the most interesting chip on the market now is something that Google has created called the VCU, the Video Coding Unit. You could almost just call it the YouTube chip because it was designed by the team at YouTube to do video compression and decompression, video and coding. Very, very special purpose chip. You don’t see it much in the press because there was no competitor. Usually, Google releases their chip and there’s some competitor who’s going to publish a white paper and talk to all the press about how “This is not as good as what we make.” There was a whole new category of chips that nobody really had thought to create before. But Google had a very, very specific need to build this chip for their own benefit. And so they went out and designed it and got TSMC to manufacture it for them. And it saves them, they haven’t said numbers, but it’s got to save them hundreds of millions of dollars a year in OpEx and CapEx. It’s a pretty interesting story there.
And then that percolates through all of these companies. It’s important to remember, when we’re talking about data center electronics, the CPUs, the GPUs that go in data centers, they’re very, very powerful niche product and they’re very, very profitable for all the chip companies. And those seven companies, maybe throw in Apple on a couple others, consume probably 70% of the server CPU category. Data center, electronics is a discrete market in its own right. It’s about $20 billion market and you have seven, eight companies buying 70% of the output of the most profitable chips that any of these companies make. And now you’re starting to see your customer has become your competitor. And that’s just this one area. And we’re starting to see this percolate through lots of other companies. The automotive companies are going to have to do something like this at some point to get to autonomy. We don’t know what they’re going to do yet, but there’s going to be a big change when that happens. My favorite example is John Deere. I think most people, where I live in the Valley, think of John Deere as very low tech. But in reality, they are a very high tech company. They’re actually designing their own chips now to do autonomous tractor driving essentially. I think we’re going to start to see many more companies like that, industrial automotive, aerospace companies design their own chips. Everything’s going to become very much more custom purpose specific.
4. An Interview With Nvidia CEO Jensen Huang About Building the Omniverse Cloud – Ben Thompson and Jensen Huang
We’re on the same path here, because one of the striking lines in your keynote was that “Future games will not have pre-baked worlds, it will be simulations”. What’s interesting about that is it’s getting into the economics of gaming, where it used to be that all the cost of a game was in the actual game development, in developing the engine and all those sorts of pieces, but at some point the cost of creating graphics for immersive games has surged in line with the capability of chips like yours, and resulted in this world of super expensive AAA games, to your point, of all these artists laboring away at all these scenes. Then there’s little independent games, and there’s not much in the middle. Is a world of pure simulation, as fantastical as that sounds, perhaps counterintuitively a more accessible one, because you can code the environment, or you can describe the environment, instead of laboriously drawing it?
JH: Well, there’s a perfect analogy for what you’ve just said. Artificial intelligence, people thought, would aggregate technology powers in fewer companies, but in fact artificial intelligence democratizes computer science. It makes it possible for anybody to write software, it’s democratizing and makes everybody a creator, it’s going to make everybody a game developer. If you can go to the extreme and say that a user could generate and create all kinds of really interesting games, why can’t small studios create interesting games? Why can’t larger studios than that create even larger-scale games, and so and so forth? In the future, game developers will be less about the hard and the repetitive engineering of creating the world physics, which includes computer graphics and physics and everything around it, and more about the game. So it’s going to be more about creating interesting games and fun games, and gameplay, and interesting assets and things like that, and less about making physics work, because hopefully physics just works. In our world, physics just works…
...I did want to go back to, you mentioned this democratization of AI, and that’s been a real mind shift and wake up call for me, and definitely, I was on the other side of assuming it would lead to more centralization. Stable Diffusion is like the preeminent example of this, where it’s open source, you can run it locally, but it’s also already being modified in all these super interesting ways. At the same time, it’s still pretty expensive to run this stuff, and I’m curious, 1) did you always have a view that AI would be more democratic than people thought it would be, or has that shifted for you over the last little bit as well and then 2) to what extent, if any, is some of this cloud offering, not just about, again, the sort of big industrial applications, but I’m particularly interested in using these GPUs on the edge, making them actually accessible for consumer-grade applications in a cost-effective way?
JH: First of all, if you think about the first principles of artificial intelligence, it’s about a computer that writes software by itself. The extent by which computing could help society is limited not by the cost of computers, it’s limited by the number of people who know how to program computers, and so the number of people who can now program computers has gone up by a factor of several orders of magnitude, and the reason for that is because just about everybody knows how to teach somebody else how to do something, or show somebody how to do something properly, but very few people know how to program C++. I think for artificial intelligence, the concept that you now have a computer scientist inside your computer, that’s the first part of democratization.
The second part of it has to do with the fact that you can write a piece of software, and this piece of software can now adapt to a whole bunch of other skills that you never intended to write it for, that’s the second layer of democratization. The first one already happened, right? You could argue that it even got a boost with GitHub’s Copilot that helps you write software. So the number of people who could write software, artificial intelligence software, the concept of zero code, the concept of low code, all of that concept of AI-assisted coding, all of that is really about democratizing programming, and so on. Next layer is what’s happening right now with large language models. It’s really quite an amazing thing that the scaling law caused not just only the recognition of patterns and relationships, but the scaling law made it possible for us to literally encode human knowledge into this neural network.
And everyone has access to the Internet to get all this data, to feed into it.
JH: It’s a little bit like this neural network now has memory, because it learned from everything that we’ve ever done and spoken about. What’s a really interesting application? Well, maybe there’s a rare form of cancer and it was described in doctor’s notes in a lot of different ways. Maybe it was described in research, but it’s rarely seen and the reason why it’s rarely seen is because it’s rare. But you know the characteristics of it, the multi-modality characteristics of it, and if you can now teach an AI how to now imagine a whole bunch of different iterations of that rare disease that you’ve never seen before, or you rarely seen from the descriptions of research. Now, this AI image generator can generate a whole bunch of different versions of it and you go, “Oh my goodness, this is what it looks like” and then you could take a computer vision algorithm that then now in the future, when you see something like this and a whole bunch of versions of something like this detected in your ultrasound or your CT, or whatever it happens to be.
Now you can imagine how, in fact, this large language model, which has embodied and encoded so much human knowledge could reduce the complexity and make it possible to solve problems that we’ve never solved before. Now this large language model says, “Well, four or five people can go and train four or five companies or institutions, or you could team up together just like we did with Hugging Face.” A whole bunch of people can come together and train one model, which then afterwards, we could adapt, fine tune, prompt learn into a whole bunch of other skills that it’s never been trained to do, and so now it’s been democratized.
5. Henry Ford: My Life and Work – David Senra
Henry Ford detests lazy people, and he’s going to bring this up over and over again, humans are made to work. I’m going to read this to you first and then I’ll tell you what I wrote down right after I read it. So he says, “The natural thing to do is work, to recognize that prosperity and happiness can be obtained only through honest effort. Human ills flow largely from attempting to escape from this natural course. So the way I think about that is humans are made to work. The sense of accomplishment from overcoming difficulty is satisfying in a way that a life of leisure and ease will never be.”…
…He’s giving us advice. He’s like, “Hey, if you want to build a business, you want to build a product, start with a product that already exists, study it, and then just find a way to get rid of the useless parts.” Well, he starts out building $6,000 luxury cars for other people at a time when the average worker was making $2 or $3 a day, and he realized, “We have all this stuff in the car that’s just not needed.” And so you can think about Ford’s career as like, “Okay, that’s the state of cars now. And so for the next 15 to 20 years, I’m going to find a way how to keep making this simple, keep making it lighter. And then the more simple I make something, the more of them I can make. The more of them I can make, the lower the price gets.”…
…He goes into a lot of detail. I’m going to pull out one because this is important. So at the point, at this time, there’s a good amount of wood that is being put into these cars. And so he says, “Take wood for example. For certain purposes, wood is now the best substance we know, but wood is extremely wasteful. The wood in a Ford car contains 30 pounds of water.” This is very basic. It’s simple, but not easy. His approach to building a business is simple but not easy. So it’s like, “Okay. Well, this is just one material and one part of a car that has hundreds, maybe thousands of different pieces to it.”
And so he is like, “Okay. Well, this seems to be very wasteful. The wooden afford car contains 30 pounds of water.” And this thought process is fantastic. So he goes, “There must be some way of doing better than that. There must be some be method by which we can gain the same strength and elasticity without having to lug around useless weight,” and this is the punchline. This is so important and why I’m reading it to you. “And so through a 1,000 processes.”
His point is like, “I didn’t just think about this for wood in the car. I thought about it for every single part of my product, every single part of my business.” That line of thinking takes time because you have to do it a 1,000 times to different areas of your business, but if you are willing to invest that time, by the time he starts mass producing, by the time the first Model T comes off of the assembly line, he has been experimenting and using this kind of thought process for 20 years. So they think, “Okay…” And why is that important? Because by the time that process starts, the game is already over, game, set, match because now he can make a product in 12 minutes, that takes his competitor two days.
[00:26:40] And so he talks about in kind of a weird, hard to understand way that, “Basically I only had one idea. I wanted to make the car for the every man. I was only interested in trying to get that car, the price as low as possible.” And so what he’s about to say here is, “Really, you start with the product,” and the product that he wanted to make was the inexpensive car for the everyday man. And then you try to work backwards on how to do that over many, many decades. The place to start manufacturing is with the product. The factory, the organization, and the selling and the financial plans will shoot themselves to the product. And that process is extremely long. He says, “I spent 12 years before I had a Model T, that suited me. We did not attempt to go into real production until we had a real product.”…
…And so he’s saying, “Hey, other entrepreneur, waste is largely due to not understanding what one does or being careless in the doing of it. Greed is merely nearsightedness. I have striven towards manufacturing with a minimum of waste, both of materials and of human effort. And then towards distribution at a minimum of profit.” That means profit per car. He’s not trying to artificially suppress the amount of money he makes. He just doesn’t want to make a lot per car.
“Depending for the total profit upon the volume of distribution.” So there you go. I kind of ran over his point there, and then he has this four-part philosophy of service. I’m going to save that for the end because he repeats that in the last chapter as well. So I’m going to skip to… He’s going to actually give us… Oh my goodness. So all we’ve been talking about his philosophy in his autobiography. Now he actually hears a little bit about his life and he says, “On May 31st, 1921, the Ford Motor Company turned out car number 5,000,000.” And he’s about to tell us, “Hey, I’ve been at this a long time. I’m happy I just made my 5,000,000th car, in the Model T alone.”
[00:29:32] He’s going to sell 15,000,000, just of the Model T. So he’s talking about number 5,000,000 of any Ford car that’s not… Model T alone is 15,000,000 but that’s going to happen over the next decade and a half from where we’re in the story. But the reason I’m bringing this to your attention is because like he said, he built a gasoline buggy, which is kind of the first primitive car, 30 years before he produced his 5,000,000th car. And why is he bringing this up? Because he’s listening, car number 5,000,000. He says, “It’s simpler than my first car, but almost every point in it may also be found in the first car.”…
…And what’s fascinating is the fact that he focused on… I don’t want to be too complicated because I don’t want to direct your attention away from what we’re talking about, but there is something interesting in setting the early days of the automobile industry that I think applies universally to other industries.
Ford took the lead, was the dominant player, had over 50% market share, but that’s how good his one idea was, but that one idea led to the giant gap because Billy Durant of GM’s like, “I’m not going to just make one car. I’m going to make a conglomerate. I’m going to have a cheap car, a medium car, a luxury car. I’m not going to do all this. I’m going to go out…” What he did is, “I’m going to go out and buy all these other existing…” He bought Buick and Cadillac, and I think he started Chevrolet. I don’t know if he bought it. I can’t remember, but the point being is just eventually people got tired of just having the one black Model T and once they started, then cars became like, “Hey, your car says something about you,” and there’s all this other things.
The entire time that Henry Ford was having success, Billy Durant and then really Alfred Sloan because Billy Durant lost control of his company and Alfred Sloan was the one that brought GM’s… Essentially, Alfred Sloan was the one that brought Billy Durant’s ideas to fruition in the marketplace. And because GM had better product offerings, once the Model T tapered out, that’s when GM had this huge rise. And I think by 19… I want to say like 1950, maybe, let’s say 1950, 1960s, GM’s one of the largest companies on the planet.
And really, the takeaway there is like, okay, one industry, you have two legendary founders. They both have completely different philosophies and sometimes the philosophies work depending on the time…
…So check this out though. He’s selling these cars. Obviously, if there’s only one car, people are like, “What is this saying? Who made it?” His boss at the electric company tries to get him to give up his experience. “During all this time, I kept my position with the electric company and gradually advanced to Chief Engineer. I was making $125 a month, but my gas experiments were no more popular with the president of the company than my first mechanical leanings were with my father. It was not that my employer objected to experiments, only to experiments with a gas engine. I can still hear him saying, ‘Electricity, yes. That is the coming thing. But gas, no.'”
And this is for its point, remember we’re in 1892, maybe 1893 at this point. No one had the remotest notion of the future of the internal combustion engine while we were just on the edge of the great electrical development. So everything, all the energy, the attention, the money, the entrepreneurial energy and spirit is going into electricity, and yet Ford was able to ignore that distraction and with great independence of mind, be like, “Ah, no. I’m going to focus on this thing because I like it. I’m interested in it. I think it has a future.”
[00:40:47] That’s easy to say, “Oh yeah. Think for yourself.” It is almost impossible to actually do, especially when everybody around you is focused on something else. And this is where Henry Ford at every… There’s always like a fork in the road in an entrepreneur or founder’s life, right? It’s like at some point, none of this is going to work if you don’t bet on yourself. And usually, you’re not in the best position when you have to make this decision, which is why so few people do it.
“The Edison company offered me the general superintendency of the company, but on a condition that I would give up my gas engine and devote myself to something useful. That’s their words. I had to choose between my job and my automobile and I chose my automobile.” This is a crazy thing, right? This is a calculation based on arrogance. “There was really nothing in the way of choice for already, I knew that the car was bound to be a success. I quit my job on August 15th, 1899 and went into the automobile business.”
[00:41:46] Why did I say that is a calculation based on arrogance? At this point, there is no such thing as a commercially successful car company. And he’s saying, “I already knew it was bound to be a success.” It gets even more unbelievable. “I had no personal money. What money was left over from living was all used in experimenting, but my wife agreed that the automobile could not be given up and that we had to make or break.” And this is so important. He’s talking about the fact that at the beginning, there is no demand.
He’s essentially describing the creation of a new industry and how humans are likely to react to a change. He says, “There was no demand for automobiles. There is never demand for a new product. At first, the horseless carriage…” Not even called an automobile, right? “The horseless carriage was considered merely a freak notion. And many wise people explained why it could never be more than a toy. No man of money even thought of it as a commercial possibility. In the beginning, there was hardly anyone who sensed that the automobile could be a large factor in industry. The most optimistic people at this time hoped for a development that was akin to the bicycle.”…
…So 1905, 1906, they’re selling models. And some of the models are $2,000 and $1,000 each. And he’s like, “It’s cheaper than a $6,000 car or a $4,000 car that was common at a time. But he’s like, “It’s still too expensive.” And so he is like, “At $2,000 and a $1,000, we only sold 1,500 cars.” And so he eventually goes like this continuous improvement. And this is still way before the Model T, keep in mind, but he gets the price down. Instead of $2,000 and $1,000, it’s $750 and $600.
[00:53:07] And this is where he’s like, “Oh, I’m right and everybody else is wrong,” because once he got the price low enough, he goes from selling 1,500 cars to almost 8,500. So what’s fascinating is even though they’re selling thousands and thousands of cars, he’s getting closer to his original idea. He’s not there yet. The other people like the shareholders and other people in the business are like, “Okay, this is good enough.” But again, Henry Ford’s like, “No, I just have one idea. I’m glad of the progress I’m making towards it. I think I’m getting closer, but I’m still not there yet.”
And so he says, “We were a prosperous company. We might have easily sat down and said, ‘Now we’ve arrived. Let us hold onto what we got.’ Indeed, there were some disposition to take this stand. Some of our stockholders were seriously alarmed when our production reached a 100 cars a day. They wanted to do something to stop me from ruining the company and when I replied to the effect that 100 cars a day was only a trifle and that I hope long before to make a 1,000 a day, they were shocked.”
And he says, “The temptation to stop and hang on to what one has is quite natural. I can entirely sympathize with the desire to quit a life of activity and retire to a life of ease. I have never felt the urge myself.” So again, there’s like a little bit of an ego there. He’s like, “Yeah. They thought a 100 cars a day was fine. They thought a 1,000 cars a day was crazy.” He’s like, “Listen, I understand it’s very natural. You don’t want to mess up what you have. I sympathize your desire to kind of take it easy, but that’s not why I’m here. I have never felt the urge myself. “
And it continues. “It was however, no part of my plan to do anything of that sort. I regarded our progress merely as an invitation to do more.” And why is that? Because he says, “I had been planning every day through these years towards a universal car and we are not there yet. So therefore we are not stopping.” And so now we get to Henry Ford’s iPhone. So we can think of it, “I designed eight models before the Model T. I made the following announcement. I will build…” This is him quoting himself. “I will build a motor car for the great multitude. It will be large enough for a family, but small enough for the individual to run and care for. It will be constructed of the best materials, by the best men to be hired after the simplest designs that modern engineering can devise, but it will be so low in price that no man making a good salary will be unable to own one.” That is Henry Ford’s one single idea.
“It will be so low in price that no man making a good salary will be unable to own one.” The important part is it’s easy to say, “Oh, just make my product cheaper.” But he’s like, “I don’t want to make a low quality cheap product. I want to make a high quality cheap product.” And so for there, he’s literally got to invent the ability to mass produce cars, which did not exist before Henry Ford. So he makes this statement, they runs this ad and it says, “The question was already being asked, how soon will Ford blow up? Nobody knows how many thousand times it has been asked since. It is asked only because of the failure to grasp that a principle rather than an individual is at work, and that the principle is so simple that it seems mysterious.”
[00:56:14] And so then he goes into all… There’s a million different decisions he has to make and a million different processes that he has to examine and just continuously improve over and over again, and you do that for decade after decade. And you get to be able to build a high quality product at a price nobody can match, but the way his brain works is very similar to the way Rockefeller’s brain works. And so, for example, he’s talking about the layout of the factory and the workflow. And he says, “Hey, if we can save 10 steps a day for each of the 12,000 employees that I have, you will save 50 miles of wasted motion and misspent energy every day.”
So that’s his idea. He’s like, “Ford’s focus on the continuous improvement as this company scaled.” It’s going to remind you of when I just reread Titan. It’s episode 248 and Henry Ford and Rockefeller understood this. He’s like, “These little things make a big difference because our organizations are going to be gigantic.” And so Rockefeller’s watching a machine that would like solder caps to cans.
And he’s like, “How many drops of solder do you use on each can?” And the guy’s like, “We use 40,” and he’s like, “Have you tried 38?” Rockefeller asked. And the guy’s like, “No, but I will.” And so he tries 38 drops and a small percentage of the cans leaked so that doesn’t work. But then they try 39 instead of 40 and none of them leaked. And so 39 drops of solder became the new standard oil, at all the standard oil refineries.
And why is that important? Rockefeller’s going to tell us right now, that one drop of solder, and I’m most likely pronouncing that word incorrectly by the way, that one drop saved $2,500 the first year. But that export business kept on increasing after that and doubled and quadrupled, and then became immensely greater than what it was. And the savings has steadily gone along one drop on each can and now has amounted since to many hundreds of thousand dollars every year. One drop on one process, on one tiny part of a gigantic enterprise. That is how Rockefeller and Ford thought.
6. 8 Dangerous Things People Say About the Stock Market – Chin Hui Leong
1. The economy looks bad, I will invest when things clear up
When confronted with uncertainty, you may feel safe sitting on the sidelines until the economic problems blow over. But in reality, the stock market often recovers before the good news arrives.
Take February 2020, when the S&P 500 fell by almost 34% before bottoming out in March 2020 at below 2,240 points as news spread about the COVID-19 virus. Then, without warning, the index proceeded to rebound and surpass its February high by August 2020, long before Pfizer (NYSE: PFE) or Moderna’s (NASDAQ: MRNA) vaccines were approved.
As of its close last Friday, the S&P 500 is 82% above its March 2020 lows.
Investors who sat out, waiting for better news, have been left behind.
2. I will sell my stocks now and buy back later
When the economic outlook becomes hazy, there is the temptation to sell your shares today with the intention to buy back when share prices become cheaper. While the idea sounds good in theory, it rarely works in practice.
Research from Ned Davis shows that half of the S&P 500’s best daily returns occur during a bear market while another 34% happen in the first two months of a bull market. Missing the best days can be detrimental to your returns. Data from Bank of America (NYSE: BAC) shows that if you missed the 10 best days from each decade between 1930 and 2020, your returns would be a mere 28%
Instead, if you have held through the entire period, you would have earned a handsome 17,715% gain.
3. This stock is down 75%, it can’t go any lower
The next lesson comes from the depths of the Great Financial Crisis.
On 5 May 2009, I bought shares of American Oriental Bioengineering after it had fallen by over 75% from its all-time high. Alas, the idea did not work out; today, my shares are worth zero.
Here’s the thing: in 2009, the company had US$296 million in revenue and US$41 million in profit. By 2012, sales were cut in half with a loss of almost US$60 million. Shares were subsequently delisted that same year.
As Warren Buffett once said, “If the business does well, the stock eventually follows.” Conversely, if the business does poorly — so too, will the stock.
4. My stock hit a new high, how can it go any higher?
Buffett’s saying above cannot be understated.
In particular, when your shares hit an all-time high, there is a strong temptation to take some money off the table, lest you lose your gains. Yet, decisions made based on share price movement alone rarely work out.
Case in point: I sold half of my holdings in Netflix (NASDAQ: NFLX) at around US$7 after the shares doubled. Today, Netflix is worth around US$233 per share.
Tellingly, this enormous gain is backed by its underlying business performance. Between 2007, the year that I bought my shares, and today, Netflix’s profits have risen by 76 times. Fittingly, the other half of the Netflix shares I own today are worth over 71 times my original investment.
7. Incentives: The Most Powerful Force In The World – Morgan Housel
True story about a guy I knew well: A pizza delivery man who became a subprime mortgage banker in 2005.
Virtually overnight he could earn more per day than the earned per month delivering pizza. It completely changed his life.
Put yourself in his shoes. His job was to make loans. Feeding his family relied on making loans. And if he didn’t make those loans someone else would, so protesting or quitting felt pointless.
Everyone knew the subprime mortgage game was a joke in the mid-2000s. Everyone knew it would end one day. But the bar for someone like my friend to say, “This is unsustainable so I’m going to quit and deliver pizza again” is unbelievably high. It would be high for most of us. I didn’t blame him then, and I don’t blame him now.
A lot of people screwed up during the financial crisis. But too many of us underestimate how we ourselves would have acted if someone dangled enormous rewards in our face.
This goes up the food chain, from the broker to the CEO, the investors, the real estate appraiser, the realtor, the house flipper, the politician, the central banker – incentives lean heavily towards not rocking the boat. So everyone keeps paddling long after the market becomes unsustainable.
Disclaimer: None of the information or analysis presented is intended to form the basis for any offer or recommendation. Of all the companies mentioned, we currently have a vested interest in Amazon, Meta Platforms (parent of Facebook and Instagram), Netflix, Shopify, and TSMC. Holdings are subject to change at any time.
