Podcast - The 6 Laws of Technology with Nathan Furr Part 2

Moz Afzal:
Welcome to Beyond the Benchmark, the EFG podcast with Moz Afzal.

Nathan Furr:
So we're continuing with our three-part series on the six laws of technology. Hopefully you listen to the first podcast of the first two here with Nathan first, so I still have Nathan here with me. So Nathan, let's talk about the third law of technology.

Great Moz, and be careful you get me started talking about technology. I'll get excited because we have all these decades of research around the patterns by which technology changes the world we live in, but it's usually not very accessible to folks or it's fragmented, it's all over the place. And so our work has really been to try to synthesise that and try to say, how do we help a leader know where and how to invest in technology?

Moz Afzal:
And I was going to say, isn't this like a chat GPT version of the laws of technology?

Yeah, it's really because what does chat GPT do? It's really a large language model is about creating correlations between language and using language or written text and then other forms of media, but essentially creating a correlation. And so what are those correlations that drive the world? That's really what it's about. So yeah, you're actually quite right. So the third law of technology, and again, the laws are supposed to be what do we see from research is compared to conventional wisdom. And I'd say the conventional wisdom, we tend to think about technologies as unitary holes. I'm going to invest in this, I'm going to invest in ai, I am going to invest in electric vehicles, I'm going to invest in blockchain. And I see really smart people invest in areas of technology that are really unfortunate because they lose their shirts.

And what we like to say, the third law of technology is to reinvent the pieces, not the whole. And if I could use a really classic example of this, when I teach MBA students, I teach them this case about a company called Media Tech. It started in the late 1990s in Taiwan. And they really laid the groundwork for the modern smartphone in the rest of the world, in Asia in particular. And they ended up making just boatloads of money, just boatloads of money, enabling a hoard of phone manufacturers and phone distributors and a whole ecosystem of phone people that made just millions and millions of smartphones. And some of those smartphone makers grew up to become the big smartphone names today. But it was always a puzzle. Why did media tech make so much money and why would they enable their competitors, so to speak, to make phones?

Why would you do that? And what they were, what many people miss is that most technologies, especially manufactured technologies, but we can think about other types of technologies as well. They have what we would call a hierarchical structure, meaning they actually can be broken down into smaller pieces that we nerdy academics call modules. And inside those modules we call components. And what an architecture does is it just defines the rules of how the parts and pieces go together. So if you think about, let's say an electric vehicle, take a Tesla for example. There's a flat panel battery in the bottom, there's two motors at the front and back. There's four tyres, there's doors. That's the architecture. And what you want to define is what are the interfaces or the rules by which things communicate and allow different modules to work together. And the idea is that this process, every technology starts out typically fairly integrated, meaning it's all built as one piece.

Look at the early computers, they're all built as one piece. Look at the automobiles, early automobiles, all built kind of couple workers like putting everything together and tying it all together. And then over time, technology naturally become more modular. And what that modularity does, it enables the independent pieces to innovate on their own because you can swap 'em in and out, you can move a lot faster, and the technology tends to develop much more quickly. It's much more scalable because you can just combine different modules in different ways. And the big challenge though is to ask where does the value end up? Where's the profit end up? Because as this integrated technology breaks down into modules, profit is not distributed equally, and the profit tends to get distributed into the pockets where either A, that module is a bottleneck to the performance of the rest of the system, or B, you've developed a smaller little architecture, a little maze so to speak, inside your module that other people can't break or imitate and they need you.

So when IBM introduced the first PC that everybody ended up buying, there were other PC makers out there, but they really created the first PC and they launched the PC market and grew it to a couple billion dollars within a few years. It looked like a massive success, but it was actually a total strategic failure because while they had enabled the market to take off and they'd kind of defined the architecture, they didn't own the modules where the value would fall and where were the modules where the value would fall. It was in the microprocessor, which Intel owned, and it was in the operating system. Those were the bottlenecks performance. And so I remember I've written about Tesla way early on in their journey. I got to interview Elon Musk when he seemed like a more reasonable human being. I don't know, he's a very polarising character.

But back in the day, people would say, this company is so stupid, why would they build battery factories? Everybody knows you want to outsource your components. It's a huge risk. But if you look at electric vehicles, where's the bottleneck to performance? It's batteries and they own that. And as the future's evolved, where's the next bottleneck to performance? It's in the self-driving. And what's the bottleneck in that? It's the ai. And so that's why you see Tesla making these gargantuan investment in chips to make their kind of algorithm a machine work. And so again, rule reinvent the pieces, not the whole.

Moz Afzal:
That's a really good example I think with I guess X AI in terms of what they're doing. Obviously it's bridging the gap, also using X or Twitter that used to be known as the inputs, but using that engine for self-driving as well. Again, they've got huge amount of data to pass out. So the strategy does actually make a lot of sense. Do you think then just the natural extension, if we use Tesla as the model here, the natural extension is as batteries no longer become the bottleneck, and we have quite a few, the Koreans make them, the Chinese make them. Obviously Tesla makes them. There's a few others as well. Will that be the bottleneck in the future? And is the self-driving now going to become where the value sits rather than in the battery?

Nathan Furr:
Yeah, so most your intuition, the right is right in the sense that here's the tricky thing about bottlenecks is they move, they migrate. So if the bottleneck and batteries is just about making cheap batteries, that's a bottleneck that will probably move either with scientific invention or manufacturing invention. One thing that might protect the Tesla battery is that it's not just about a great battery, it's also the architecture to manage the batteries themselves because it's actually quite challenging technical problem to optimise that. So that's one reason why somehow the Tesla range, even though people think they're cheating, but still somehow Tesla kind of seems to beat everybody on the range nonetheless, and it's because of that architecture. But you're right, how strong and durable is that architecture versus, I know you've talked to me a lot about Nvidia, and Nvidia is a really interesting case because through both chance and strategic choice ended up really owning a really important architecture because artificial intelligence really is about using matrix algebra to understand large amounts of data and graphic processing units, which Nvidia made are really kind of a specialised architecture for handling the kind of math you need to do ai.

And so they got kind of lucky we had the architecture, but I think then where Nvidia has been really smart is to say, how do we extend that architecture and extend it and make it deeper and make it even higher performance and harder to imitate? And so yeah, what's the nature of the bottleneck or architecture that's holding the profits and you either got to fight to keep it there or will eventually migrate away because what do bottlenecks do? There's a lot of profit, they attract people to try to do stuff there.

Well, certainly Nvidia is a bit of a poster child for the current environment that we're in in terms of how successful that's been. And it does remind me your earlier comment about IBM, and obviously you've got the processor, but Nvidia started making those same processes for graphic scars because the typical micro pressor wasn't strong enough to actually, or fast enough, should I say, to do those calculations to make those games and those videos. And that's just an extension of that technology. It's just moved on and speed has moved on and so, so forth to the point that now even companies like Qualcomm are now making PC laptop chips that are faster than Intel chips to specifically around AI and what's going on there. So yeah, it makes complete sense in terms of how fast that technology moves and how I guess strategic mistakes as Intel has made, for example, have really come to haunt them having been in such a strong position just 20 years ago.

Yeah, I absolutely agree. I have to add one more piece about this. Reinvent the pieces, not the whole, and we've been talking about specific technologies, but this is actually a way bigger deal than we give it credit for because it's not just about an integrated technology breaking down into little pieces. It's about the rules that allow those pieces to communicate. And one of the big shifts that happened in our digital world is through the integration and standardisation of APIs, application programme interfaces, these are just rules for how software and data sets will communicate. That's actually really a total revolution because what it means is it's easy to break things into modules and to connect things, and it breaks some of these familiar trade-offs around as something gets bigger or scales, it gets more complex and unwieldy. And I think the best example of this is a fairly overlooked email that Jeff Bezos sent to Amazon employees in 2006.

Moz Afzal:
And it basically said in simplest terms, everything inside of Amazon will happen via APIs. There will be no special exceptions, no back doors, nothing. Those APIs will be designed so that external parties can connect into them. And then he finishes the email saying, anybody who doesn't do this will be fired. And so pretty harsh, but I think that was like to me from an organisation design perspective, we're not even talking about tech here. That was a watershed moment because what it meant is Amazon became hyper modular and it could scale in a way that companies of the past could not scale because it got complex, because of all the rules to get things done, all the communication to get things done, all the coordination. You go back to the old multi-divisional form corporations, the General Motors of the world, they got bigger, but they just got way down.

With all that complexity, I think Amazon has, depending on how you count it, it has hundreds of thousands of employees and operates over 40 divisions doing things as diverse as e-commerce and robotics, AI and B2B and B2C. And it's just remarkable how fast it scaled and breaking that kind of fundamental trade off. So by the way, by some counts, Amazon has employed 1.6 million people. How is that possible for an organisation to employ 1.6 million people and not grind to a halt under the complexity and the bureaucracy? It's an unimaginable scale, and that's what I mean by there's this fundamental shift where every company is fundamentally becoming in some way where they recognise it or not a tech company. And again, to more or less degrees if you make steel rollers for grinding coffee less. But nonetheless, I could point out ways in which technology is entering the core. So that modularity idea is actually so much bigger than media tech or Intel or Nvidia or Tesla. It's actually entering the DNA of organisations.

Yeah, I quite like a nice API to connect with my different departments in EFG. It would make life a lot more easier. So let's move on to the fourth law, Nathan.

Nathan Furr:
Yeah, so the fourth law is probably more familiar to people, but I am surprised how often I run into this. This is a little bit different myth, and I see it kind of broken on two sides. There are some folks who think there's been a lot of talk about platforms and platform, this platform that, and platform's a fairly new term. So PayPal by the way, was one of the early big digital platforms back just early internet era, so late 1990s. And I have a video of Peter Thiel and Max Lian talking about what they do. And they don't even have the word platform. They're calling it the product, the product, the product, the product. But it was a platform. So on one side of the kind of aisle, I see folks kind of think that platforms and products are kind of the same thing, but platforms are just kind of better products or more powerful somehow.

And by platforming examples would be Amazon, Google, apple, meta and so forth. What does a platform do? It facilitates interaction. And then on the other side of the aisle, I see people saying the world is exponential. Everything's exponential, and platforms are everything. And I think both of those are kind of a little bit misguided because not everything is a platform and not everything is the same. So what we suggest is you want to, we called it watch out for the whirlwind. And what is the whirlwind? Well, the thing that makes a platform powerful or AI powerful is that it has these self-reinforcing cycles and the platform-based world, and we call it network effects. And a network effect, just so everybody's on the same page, is that every user who joins the platform makes it more valuable for every other user. And the classic example of this was the telephone network.

The more people who got telephones, the more valuable it was. And this is why you have only a few big platforms today because it tends to create a winner take all because if more people make the thing more valuable, then everybody to go to one or two. It's why you have one or two search engines, one or two ride sharing platforms, one or two food delivery. So that's like a self-reinforcing cycle that leads to a winner take all. But there are other self-reinforcing cycle. It's like a learning cycle. So the early Yahoo, and remember Yahoo was like the darling of the internet, and Google was this little nerdy startup formed by some CS graduates. And Yahoo had this vision to be the media portal for the internet, and they were going to offer all these services. And search was really, who cares? It was just like was a dirty task that had to be done.

So they outsourced it to these friends, these little startup around them called Google. But what they didn't realise is that as Google performed more searches, they got better at making searches, which drew more people to do searches, which made them better at making searches. And so it kickstarted this cycle of what we call a learning cycle, so that by the time Yahoo woke up and said, oh wait, search is really important, it was too late, they could just never catch up to Google. And I've talked to senior execs inside of Yahoo who just really bemoan it with their head hung low, like, oh, that was our biggest strategic mistake ever. But they weren't looking for this. We call it the world when this positive cycle that just makes it so you can't beat it. And AI has very similar effects. It's more data makes you more powerful, more processors, makes you more powerful, better.

It allows you to develop better algorithms. So fortunately, if those self-reinforcing cycles stay true, if they remain strong, then it'll probably be a winner take all big player kind of game at the end of the day. And I guess the other, by the way, I still debate about what to call this law sometimes I want to call it cut the sandwich. So I did a little work with, I'm doing a little work with Sanjit Chow. He is one of the big platform thinkers and he had an idea, I want to give him credit for it. He said, what's funny about the big platforms, it's not just that they're a platform in their industry. He said they start to extend horizontally across industries and then a couple of these mega platforms or sandwich layers is what he was calling them. They start to squeeze the players inside a vertical industry.

And the vertical players have very little power. So they create usually an access layer. So access to customers, they create an infrastructure layer, they may create another layer. And the simplest example here would be Amazon. I mean Amazon opened up, the minute they became a platform was when they opened up the third party sellers. Great, wonderful. But then they did some incredibly smart things. So they started to say, well, we can do logistics for you too. And by the way, if you want to have the prime button, then you need to do Amazon logistics. So they went from taking say 20, 30% of a third party seller's gross margin to taking maybe 40%, maybe 45, 50% of their gross margin. And then they said, and now we're going to introduce advertising. So if you want a chance of being sold, and if you really want to be the buy now button, which is like the crucial button, then you've got to buy a lot of advertising.

Moz Afzal:
So you have both volume and pricing, and now we're going to get 60% of your gross margin. And so they're squeezing those third party sellers. And so that's why I say cut the sandwich because you want to watch out for the whirlwind. If you happen to be a player who has a lot of money and a lot of power, you could create your own platform, great. But for the rest of us, we've got to be watching out for that whirlwind and say, how do I break it? How do I step outside that? And it is possible, it totally is. But it's why you see some of the big antitrust legislation going on now because I think legislators are starting to wake up and say, wait a minute, these companies have a lot of power. And I'd say, yeah, you're like 20 years late and realising it.

Nathan Furr:
No, that's exactly the point I was going to say is that in this environment, big has become beautiful, right? In that respect. And these guys unfortunately make so much cash. They're some of the richest companies we've ever had. So they can take the advantage that others can't. And I think what's interesting is that regulation has definitely got tighter because they now certainly in the last two or three years have stopped those simple acquisitions. Acquisitions stopped as a strategy because they know that they will have regulatory scrutiny and probably won't allow 'em to happen. But they're finding back doors, aren't they, in terms of Microsoft taking a stake in AI or some of the other sort of companies that are coming out in Europe, for example. And because they can't acquire them, the regulator won't allow them to. But let's take some stake and then become a strategic partner. And that seems to be the strategy they're using or they're just throwing money at it and just doing it themselves. So it's got quite a unique situation they're using. And I think we'll see where this goes because it's super interesting.

It's super interesting and it's a super challenging, thorny problem solve because these are really sharp people who tend to run these, and I can't say publicly some stories I know, but there's definitely some strategies that they employ that are pretty remarkable. And it's a rock and a hard place, right? Because you regulate the American platforms or you regulate the European platforms, great, but what about the Chinese platforms? Because they have even more people, even more data and more. And so that's a lot of people are wondering about sometimes geopolitical tensions. And another conversation you were having was about US elections and US politics and global politics and all that. Don't pretend that this technology battle is not part of it is absolutely part of it. And it is about this winner take all journey. And so it's not a problem with an easy answer other than you as the player in the market, how do you either cut the sandwich or if you're the investor I suppose buy the sandwich, the sandwich you mosey.

You got to know the field of strategy. Again, the study of why does one firm perform better than another? And competitive advantage comes from a really interesting place, and that is, it really exploded because there'd been like decades of economists working on antitrust legislation to break up the old monopolies, the steel and oil and railroad monopolies. They'd done all this work to break up the monopolies. And then this young guy, this young economist says, well, why don't I just take that logic of how to stop monopoly and flip it on its head and use it to describe how to create a monopoly. And that guy's name was Michael Porter and he developed the famous Porter's five forces. So these are two sides of attention. And so again, I don't want to destroy the world or create a one overlord meta is the overlord over all of us, and we're all these minions.

Moz Afzal:
I don't want that. But let's be honest about the forces that are driving this and about some of the trade-offs, geopolitical trade-offs, policy trade-offs, consumer surplus trade-offs, and make a wise choice. There's got to be a middle road technology thing about technology that can always be used in different ways. It can be used for good, it can be used for bad, it can be used to create incredible inequality, and it can be used to create immense equality and benefits for all. And those are all choices. It's not inherent in the technology itself usually. It's usually a choice. So that's just something to keep in mind.

Nathan Furr:
So wise words to end here. So thanks very much for that, Nathan. So we continue next time with the third part of the sixth laws of technology at the EFG podcast beyond the benchmark. So we'll speak again next time.

Moz Afzal:
This podcast is provided for general information only and assumes a certain level of knowledge of financial markets. It is provided for informational purposes only and should not be considered as an offer investment recommendation or solicitation to deal in any of the investments or products mentioned herein and does not constitute investment research. The views in this podcast are those of the contributors at the time of publication and do not necessarily reflect those of EFG, international or new capital. The companies discussed in this podcast have been selected for illustrative purposes only, or to demonstrate our investment management style and not as an investment recommendation or indication of their future performance. The value of investments and the income from them can go down as well as up, and investors may get back less than the amount invested. Past performance is not a guide to future returns, return projections or estimates, and provides no guarantee of future results.