Accelerating Returns in The Exponential Age
ED HARRISON: Hi. Ed Harrison here for Real Vision. I have the distinct pleasure of speaking to Azeem Azhar, who is a writer, technologist, and he's also the creator of the acclaimed "Exponential View" newsletter. Welcome to Real Vision, Azeem. AZEEM AZHAR: Thank you so much for having me, Ed. I appreciate it. ED HARRISON: You know, I'm really excited to talk to you. Because you do have an upcoming book called Exponential. We're going to get into that in a little bit. And this is part of a series of interviews that we're
doing here at Real Vision on the Exponential Age. So you're exactly the right person to talk to, who's been looking at this. You're looking at this from a number of different views. I was saying that you are a writer. You're also a technologist, and a entrepreneur. Maybe you can give our audience a little bit of a background into your technology and journalism stakes. AZEEM AZHAR: Yeah. Well, thank you. Thank you so much. In a way, I'm a child of
the microprocessor revolution. Because I was born the year after the Intel 4004 processor was released, which was back in '71. And I grew up with a computer in the home from the age of 9. So I still have that computer, by the way. And it still works, just about. So I was one of those lucky kids who just happened along at the right period of time, in the right place. I went to a junior school where there was one guy who bought us a-- one teacher who brought a computer in one day. And it stayed in the school. And when I got to University, I discovered the internet in about
1991. And that had me absolutely hooked. And so my journey-- and I talk a little bit about this-- my professional career has followed, in a way, the growth of the tech industry with the microprocessor, personal computing, and then the internet. And the funny thing is that in those early days, nobody was really on the internet. I mean, I remember sending emails back and forth to the people who defined the mail protocol that everyone's emails runs over, SMTP. And them replying to me while I was a University student, because you just did. I emailed a guy called Marc
Andreessen in '94, because there was a bug in a web browser he had written. And he replied to me, really politely, within 10 minutes. And of course, you don't know. You're just the lucky kid who's around at the same period of time. And that is where the bug for all of this bit me. In those days, it was really primordial. I don't think people can quite remember. But when I launched the websites for the Guardian newspaper in 1996, I was physically the switch for the internet cable between our servers. So when the traffic got too heavy, I would physically disconnect a cable from one server and plug it into another server. And that would reboot the other server. And in 20 minutes, I would do that again.
And I sat in this equipment room, doing this for the whole of the afternoon when load was too high. And that's what you had to do back then. And I had a chance in the first couple of chapters of my book to reflect on how rapidly exponential technologies change, and reflect on some of the funny experiences I had in my life. ED HARRISON: Wow. That is amazing. I think those are some great anecdotes. I mean, obviously, you've seen the Exponential Age, so to speak, from the beginning of the internet to where we are today. I mean, I think technologists in general have been talking about exponential things happening for a long time. How different is what's going on right now with new technologies converging, with crypto, in this space, versus, say, the internet bubble? --- RAOUL PAL: Hi, I’m Raoul Pal.
Sorry to interrupt your video - I know it’s a pain in the ass, but look, I want to tell you something important because I can tell that you really want to learn about what’s going in financial markets and understand the global economy in these complicated times. That’s what we do at Real Vision. So this YouTube channel is a small fraction of what we actually do. You should really come over to realvision.com and see the 20 or so videos a week that we produce of this kind of quality of content, the deep analysis and understanding of the world around us. So, if you click on the link
below or go to realvision.com, it costs you $1. I don’t think you can afford to be without it. ED HARRISON: is what's going on right now with new technologies converging, with crypto, in this space, versus, say, the internet bubble? I worked at Yahoo during the internet bubble. I know that you were in that space at the same time. AZEEM AZHAR: Right. ED HARRISON: How more exponential are we today? Is it just hype? AZEEM AZHAR: Well, I think, having lived through the dot com bubble like you did, and investing at exactly the wrong time like many other people did, I know the pain of bubbles. But I spent a lot of time trying to understand what's really going on
here. And there are a few reasons why things are different now, kind of fundamentally different. The first is that the computer industry was a harbinger. It was a precedent for industrial technologies that are very general purpose, that can improve at exponential rates. And it demonstrated to us what that exponential improvement does. What it does is it dramatically drives down prices. And as prices drive down, we use much more of the compute. Because we use much
more of it, complementary businesses emerge around that. And so the distinction between where we are today and where we were 20 years ago, let alone 30 years ago when I first got on the internet, is that there is an entire stack on which these innovations now expand and explode. Back at the rate with which Facebook grew in 2005 to 2006 was unprecedented. And now if you grew as fast as that with a new consumer app, you'd be a bit of a disappointment. Because you have to grow as fast as TikTok or faster.
So that has happened, that dynamic, which is that the core infrastructures of exponentiality have been laid down. We all have smartphones. There's unlimited compute in the cloud. There's multi- megabit and gigabit connections to people's homes. Means that businesses can grow, no questions asked. Back in 2000, I invested in an online real estate business in France called [FRENCH]. And not only did we discover that most of the buyers-- the consumers, not have PCs or internet connections, most of the realtors didn't. So to even get them started on this onramp, we had to buy them a PC and get them an internet connection, and teach them how to use it. So that was difficult. That problem has gone away. Now, we established that
idea within the computing industry. Prices come down, demand goes up, complementary businesses emerge. The infrastructure gets built. You now don't have to ask-- buy your realtor a PC or a mobile phone to deliver them a digital service. They have one. But now in the Exponential Age-- and that incidentally is the title of my book in the US, is The Exponential Age. So if your American viewers look for it on Amazon, that's what they have to look for. There are three other technology platforms that have similar dynamics. The platform of
synthetic biology, which is bringing genomics and protein engineering together with the tools of the technology industry. There is the general purpose technologies of renewable energy, like wind and solar, and lithium ion batteries, which are transforming the energy system. And then there are the manufacturing techniques of additive manufacturing, which are allowing us to print everything from meat to cement, to lightweight titanium, to wood. Each of those three general-purpose technologies are on a similar exponential march-- the one that we saw silicon go on between the 1960s, and it is still going on now. So that's why I say the computer industry is a harbinger for all of this. The next three platforms will do
similar things. And we sort of understand some of the dynamics. And that's why I think now, we are, I mean, I argue in the book, that it's between about 2013 and 2017 that we'd realized this-- entering a new socioeconomic industrial paradigm which I call the Exponential Age. ED HARRISON: Yes. I had to take notes as you were saying that. Because there is so much to unpack there. AZEEM AZHAR: Yes
ED HARRISON: But before we move forward with all of that, let me just go backwards for a second. Because I think I've seen something where you talked about technology becoming ever increasing. Maybe we can-- just to give people a sense of how different today is versus the past. There
was something where you were talking about, the Malthusian problem through to the Industrial Revolution. And then now it's sort of like a second, third, fourth Industrial Revolution. Maybe you can walk us through the history of technology from a longer-term historical perspective. AZEEM AZHAR: Yeah. No, I'd love to do that. And I want to apologize to viewers. Because Ed and I have known each other for a long time. And we haven't had a chance to speak much. And I've just finished this book, and I'm really excited. So I'm put up with the excitement. So if we think about humans' relationship with technology,
there are a couple of things that are important. One is that technology has a purpose. We have an itch to scratch. And it may be that we're feeling a bit lazy about having to go out and, you know, fish with our hands. Go back 50,000 years. So we kind of figure out how to jerry rig a
net, or a spear, or something. The second is that technologies tend to be a little bit ecological and evolutionary. They don't appear from nowhere. They're not like a gift from the gods. They happen through an interaction of what's gone before and what's important in our societies, and in our cultures. And they have fundamental impacts, then, in the ways we live. One of my favorite ideas is that if you look at the length of the digestive tract of a human, and you compare it to that of one of our hominid relations, we actually have a much shorter digestive tract. Because we outsourced part of our digestion to flint axes. So when we were on the veldt, in the plains,
and the rift valley 2 million years ago, whereas our chimp cousins were chewing bark and plants and having to digest it within their bodies, we chopped it up and macerated it with the flint. And as a consequence, our bodies needed to have a less robust digestive system. So for a long time-- and I'm not sure if this is what you meant by history-- technology and humans have had this relationship where it's connected to a purpose, a goal we're trying to achieve. It is influenced by what we need to do and what's around us. And it fundamentally influences us. And we can take that forward to technologies like the printing press, which people think of as a printing press coming around and increasing literacy. What the printing press did was it dropped the price of books. The salaries of professors in Italy rose in the 15th century. Literacy rose. The Reformation
came along. Science was born. The Catholic Church lost its power. But the other interesting thing about the printing press-- and I know viewers of Real Vision will appreciate this-- is that the group of men who came around the printing press represent one of the first for-profit capitalist enterprises, where people took risk shares in the success of this project. So the knock-on effects are very, very powerful in society more widely. And we've seen that back from 2 million years ago, through to the 1470s. And we see it through the Industrial Revolution. We see it through
the arrival of the car, the telephone, actually to the the start of the 20th century. And now we'll see the same thing in the Exponential Age. ED HARRISON: Yeah. That is fascinating. And it seems that the premise is that we're rising. These things are coming together in an exponential way,
meaning that they are coming together in ways that enhance one another. And let me read to you how your book is touted on the internet, that's coming up. This is what they say, Azeem. They say, in Exponential, Azhar shows how this exponential gap-- this is a gap that you can explain-- between the power of new technology and humans' ability to keep up. He explains how this exponential gap becomes our society's most pressing problem.
The gulf between established businesses and fast-growing digital platforms, the inability of nation states to deal with new forms of cyber warfare, and the sclerotic response of liberal democracies to fast-moving social problems. So this is where I want to go. I want to go to, A, what are these technologies, and then B, that whole paradigm, from a social perspective. Because people, they're interested in the whizbang of the technology.
AZEEM AZHAR: Right. ED HARRISON: But we're also interested in, how do we cope with this? What does it mean for our society? What does it mean for the politics? AZEEM AZHAR: The challenge is that we tend to-- we can focus on the breakthroughs that the technologies can make. Because we can see them more easily. But we don't necessarily appreciate how it's going to have its second or third order effects in society. One interesting example of this is that when Americans started to buy iPhones about a decade ago, they bought more and more of them. And the sales of chewing gum started to collapse. And the argument was that when people were at the front of the checkout, at the gas station, or the mini mart, they were no longer paying attention and grabbing the chewing gum from the bin. They were staring at their phones. And so you have these strange impacts that you can't predict up ahead. Now,
what's happening today is that the rate of change of technology is so fast and the rate of diffusion of that technology in the industry, and then into our economies and into our daily lives, is also very, very fast. And I explain why those things are related, and why we actually see the rate with which a new idea comes and spreads increase. And it's partly because of the internet. It's partly because of global supply chains. But essentially, if there's a trend in Shanghai on Wednesday, it'll be in San Francisco on Thursday and Stockholm on Friday.
And we have this universal access device. That means that we all get access to it. So what that means is that the disruptive potential of the technology mirrors that exponential curve of the underlying capabilities of the technology. But that's only half of the picture. The other half of the picture is the rest of the world in which we live. And those are complicated
institutions. Some of those institutions are very formal. They're like, the beautiful Capitol building in Washington, which is not only a physical institution. It's got people and rules. And some of them are much more informal. They're just habits and norms and conventions.
But social scientists tell us that institutions adjust slowly at the best of times. And so, partly because that's why they're designed-- institute sort of has that sense of solidity, of standing there. And so the exponential gap is what happens when linear institutions that are used to not changing quickly, or not changing at all, meet the reality of technologies and the products and services that are built on them, that can move at an exponentially-increasing rate. ED HARRISON: So let's attack this from one lens. I want to talk-- you mentioned before,
capitalism and companies. Maybe we can attack this via the company. The companies within the Exponential Age-- how do you look at these companies? How do they compete? What are the most salient things that we should take away from companies, technology companies and incumbent non-technology companies? AZEEM AZHAR: Yeah. I think that's a really important and fascinating area. So the argument about the exponential gap is that the technologies create new possibilities and new ways of operating. And really fundamental new
mechanisms and methods. And so they break our old institutional thinking. In the case of economies, and dynamic economies, one of the traditional forms of thinking is that you have diminishing marginal returns. That the nature of running an industrial organization is that it gets harder to get bigger beyond a certain point. So you get some scale benefits.
And then organizational complexity gets in the way. And then it's just really expensive to mine the 1,000,001th ton of coal compared to the 1,000,000, or compared to the 100,000th. And so you have this world, which is governed by the economics of Marshall, on one hand, and diminishing returns, and Ronald Coase organizational complexity on the other. That says that our assumption is that companies can't really get too big in terms of their market share. And if they do, they must have done something dodgy. ED HARRISON: Right. AZEEM AZHAR:
Standard Oil, right? ED HARRISON: Right. AZEEM AZHAR: They become monopolies. And so we then intervene at that point, because they're misbehaving. My argument is that within the Exponential Age, companies rely on becoming platforms. And platforms have network effects. Network effects, meaning that every additional customer adds value to all the previous customers that you've had. They also benefit from the extensive use of intangible assets. And intangible
assets are also assets which tend to accrete disproportionately to anyone who has an early advantage. And then when you're applying AI, which is a core exponential technology, to all of this, AI is an incredible perpetual motion machine. Because AI depends on data. But in using that data, it generates more data to get stronger. So you have platforms, network effects, intangibles, and the AI data network effect. And all of these break the force of gravity of Marshall economics and Coase, to mean that the standard state of nature for an Exponential Age company is to grow very large, become a superstar, and dominate its market sector. And the exponential gap that exists is that that's not how we thought competitive economies would ever work. And so I think that even if we look at these superstar companies like
a Facebook or a Google, and we find that there are bad business behaviors that we might not like, and they may lose court cases here or there, I think that their general state of nature, even if they had been whiter than white and cleaner than clean, and so on, would have been to trend to dominant market shares. And that means that we need new rules if we want to have vigorous, competitive dynamic markets. ED HARRISON: Right. Yeah. So I mean, that's right where the institutions and the regulatory framework meet this new Exponential Age.
Let's dig a little deeper on that. Because I'm thinking about it from a valuation perspective. You mentioned Facebook and Google. But there are other companies elsewhere. In China, Tencent, Alibaba, Ant. You know, Salesforce, Apple. The question for me, from a valuation
perspective is, Apple's a $2 trillion company. Based on what you just said, it sounds to me like they could become a $3 trillion or $5 trillion company. And this whole notion that the bigger they get-- they can't get that much bigger, isn't necessarily true.
So my question is is, $2 trillion, $3 trillion, $5 trillion, what's the impediment to that happening? I mean, is that the right way to look at it? AZEEM AZHAR: I think it is the right way to look at it. I think that if you look at an Apple or a Google, and you say there's no regulatory intervention at all, the question is, what would actually get in the way of them continuing to expand? And the things that got in the way of companies earlier, for example, organizational complexity-- we can't run teams as big as this-- they seem to have really got a handle on. Because they're still growing at 20%, 25% per annum, even though they're huge companies. The other thing that gets in the way is, what are my choices of expansion, vertically or horizontally? Well, these companies are proving they're incredibly adept at vertical expansion. Apple is making its own chips. And Apple makes its own TV shows. They're also demonstrating that they can expand horizontally into adjacent and even far-adjacent markets. So Apple's a computer company. And yet its headphones line of business, AirPods,
is bigger than most other companies in the world. And it doesn't seem to have problems in expanding horizontally. And it's as true as we might-- as when we look at a Microsoft or an Amazon, or a Google. So I think that these companies have demonstrated that there is something about the essence of having that Exponential Age DNA that allows them to rewrite the traditional rules of business that are sort of stuck on that linear institutional path.
ED HARRISON: Right. Yes. And you mentioned four effects that allow them to get there. There are two tracks that I want to talk about. One is the new areas that you mentioned earlier. You talked about synthetic biology, renewables, manufacturing technologies.
Those are three different places that we can go. But I want to think about, first, I want to talk about the four things that are coming together, not just for those areas, but the existing technological improvements in the Exponential Age. So first and foremost in my mind is network effects. When you talk about that, I'm almost thinking, winner take all. What is going on with network effects that means that these companies can grow, and grow, and grow? Whereas before, they couldn't? AZEEM AZHAR: Yeah. It's a really important question. So the network effect simply is that every additional customer adds value to every
other customer. And so what does that actually mean? When you turn it around the other way is that if I'm a customer thinking about making a choice, and one supplier has network effects and the other hasn't, by joining the supplier with network effects, I will benefit from the existing network effects and all future success that they have. And so when I make a customer choice, I will generally choose the company that has those network effects. And Microsoft in the '80s was the sort of great example of that. Because as they got a bigger user base, more developers moved on to Microsoft, because there was a bigger market to sell the software, which meant they had better software, which choices-- which brought more users. So that's the dynamic of the network effect. Now, the question is,
in order to do that, you have to have a platform of some sort. A platform essentially is a bit like a marketplace where people can come and go. If you want to buy shares in US companies, you're better off going to the NASDAQ, which is a marketplace where everybody lives, then coming to my backyard, where I've got a handful of ETFs that I could sell you a couple of shares of, right? You go where the liquidity is. And the funny sort of way-- the platform business model enables network effects in industries that we hadn't thought there would be. So if you think about Uber or Airbnb, there are network effects existing in the taxi business, or in the hotel business. But my favorite example
of how you can create these network effects on a platform and redefine a traditional business is in the area of agricultural machinery. Who would have thought it? But John Deere-- and I'm sure there are Real Vision viewers who bought John Deere stock five years ago and are very happy for it-- John Deere, about five or six years ago, said, we're not just going to make combine harvesters and tractors. We are going to become a platform. And by becoming a platform where third parties can plug into my John Deere, and therefore farmers can access this array of services, we will create a new choke-hold, which is access. And we will enjoy network effects. And they demonstrated that you could deliver those network effects in a much more traditional business, provided you became a platform. And the consequence of that is maybe it's not as strong as winner take all. Maybe it's more like winner take most. But if you look at lots and lots of markets,
you see winner take most, even in online travel, where the booking.com group is the far and away dominant space there. So across all of these different industries, you get this winner take all phenomenon. And a large part of it is driven by a network effect enabled by a platform.
ED HARRISON: You mentioned something about AI that I found very fascinating. I'm thinking about this from an Amazon perspective. I'll give you the example, when we're talking about data. So I go to Amazon. They're the seller that has the network effects. So I'm thinking this company
must have this item that I'm looking for, because that's why I'm going there, because of the network effects. Then I type in what I'm looking for and Amazon autocompletes what I type in. When I click on the autocompletion, it pops up with a bunch of different stuff for me to look through. None of those things are actually what I'm looking for. It's because Amazon knows, because of the AI,
that people like me are looking for this item. And that they're more likely to make a sale of items that they actually have that are similar to this item. And as a result, they display those items as opposed to the item that I'm looking for. They don't display, we don't have this item. They display a bunch of other items that I would potentially buy. And lo and behold,
maybe five times out of 10, I will buy that item. And so Amazon makes the sale. That's how I'm thinking about AI. Maybe you can give us the sort of like macro picture of how that data works. AZEEM AZHAR: Yeah. I mean, you've put your finger on a really great example there. I mean, essentially, what's happening within Amazon is that as I come and search for things, and then choose to buy them, or choose not to buy them, it creates a data stream. As you do the same thing, it creates a similar data stream. It may notice that when I search for a 40 pound kettlebell,
I'm quite likely to go off and then buy a bunch of protein powder. And so when you come and search for a 40 pound kettlebell-- in your case a 48 pound kettlebell, I know, Ed-- it will recommend protein powder. And that's it in its basic. But you have to think that there are hundreds of millions of people doing this for over decades now, in Amazon's case. So their ability to therefore figure out what you might buy next is unparalleled. But you've actually touched on one of the slightly trickier issues of their business, which speaks to the problem of competitive markets. And as investors, we need to see competitive markets because we have to have choices about where we allocate our capital. And part of the challenge that is starting to
emerge is that Amazon is, in a sense, on both sides of the trade. Because it sees the flow. These are things that you searched for, and what choices you make. When you search for batteries for your TV remote control, it figures out what words appeal to you in the batteries you buy, and what price point matters, and what packaging size matters. And one of the accusations is that Amazon uses that information to then build its own brand products. And this has been documented by a number of the very well-known financial papers in the United States. And the question is, does that have a stultifying effect on dynamic markets. Now, of course, retailers have always done this. They've
always used that their knowledge of footfall to figure out what own-brand products to produce, and own-brand products have got higher margins than when you go off and you get a name brand. That tension has always been there. But those markets have never been as concentrated. They've never suffered from winner take most, which these current markets seem to. So I think the Amazon story is a great story about AI. But it's also a great story about how we might have to have new ways of thinking about what it means to operate well and competitively in an economy. ED HARRISON: Yes, definitely. And I think that the nuance there is, how do you
take a regulatory approach there while still understanding that it's winner take most. And to me, the question is, is what does winner take most mean from a data perspective, an AI perspective? I'll use Google, since we're talking about surge. How many competitors-- what percentage of the most that Google should be able to take? And how should regulators deal with the fact that Google's seeing all the flow from search. So their technology, their AI, should make their search demonstrably better than others. What percentage of the market
accretes to Google as a result of that? AZEEM AZHAR: I think that's a really important question. And that takes us back to an idea that's a little bit been forgotten about, which is the idea of a natural monopoly. And from a natural monopoly, essentially saying that, how many sewage systems does any one town really need? Do you really want competition amongst the sewage networks, or would you rather say, this is infrastructure, it should be run once, and therefore it should be, as it's infrastructure, be considered like an essential provision to that municipality. And we as regulators should say, listen. You can't price gouge with an essential facility. It's just got to be there. And once you've built it, you've got this unassailable lead. And so I think that there are things that we can learn from utility regulation.
And I'm not saying take utility regulation and bring it into these markets. I'm saying, look at utility regulation, be inspired, and figure out where it makes sense. So the way that you might then tackle something like Google search is start to say, some parts of what you've built are really unassailable. And you've made a fantastic return for your shareholders from that over the last 25 years. But increasingly, because it's creating some stultification in the market, we need to change your obligations-- your societal obligations when you provide that search. And I'll give your viewers an example of something that's happened in the UK around this. So of course, broadband infrastructure is really important to roll out.
But rolling out broadband and digging up all the streets and the sidewalks and so on is expensive. So a few years ago, the telecoms regulator said to BT, which is our national provider, said, listen. In order to have a competitive market, we are going to have to split your business. And the lower layer, the bit that runs the ducts and the pipes in the streets, and does the backhaul, that is going to have to be treated as a utility. And you're going to have to offer access to that fairly to all your competitors, and yourself. But the rest of you can be as aggressive, and competitive, and capitalistic as you want. And of course, there was a bit of argument about that.
But it's actually ended up working quite well as a kind of a mechanism of saying, there's some part of your business now that has become something that is too essential for us, for our economic vibrancy, and for dynamism. And we need that to be pegged and treated slightly differently as the spicier part of your business. ED HARRISON: I want to look at how this works in some of the newer technology areas, where exponential activity is happening. The first area that you mentioned, when we were talking about this earlier, is the place that people are thinking about the most right now. Because you and I, we're in our homes,
miles away from each other. Who knows-- a year and a half ago, I would have gotten on a plane and interviewed you in the UK. But now we're doing it technologically. And by the way, I like your background. Because it glows green every once in a while, which I find really nice. The thing is, is synthetic biology-- you talked about that
as an area to watch. Synthetic biology is something that we care about with the vaccines, and everything that's going on there. Maybe you can talk about why this is one of the places to look for exponential trends in technology. AZEEM AZHAR: Yeah. It's such an incredible, incredible area, what's going on here. And obviously, we've enjoyed it from the vaccine benefits. There are two different things that we should think about. One is, what is the demand-side opportunity. And the second is, can we deliver on it. And
the demand-side opportunity is absolutely huge. Because as we decarbonize our economies, we're going to need to be able to produce a lot of materials that are currently petro-derivatives. So a lot of plastics, a lot of pharmaceuticals. They're all refinery byproducts. And we produce them by controlled explosions on refineries around the world. And what we can start to do in synthetic biology is look at what nature does, and figure out how to produce those. So for example, you genetically modify a type of microorganism, a yeast. And you can get it to brew
sheets of film that replace plastic on smartphone covers. And they have higher performance characteristics, zero carbon footprint. And that's one area for this new bioeconomy. So the demand side of that is absolutely enormous. And of course, we've seen the demand for mRNA vaccines is going to be 7.5 billion people around the planet, which is as big as it gets. On the supply side, why is this happening? It's happening because there are some biological techniques, the mRNA platform being one, but also genomic sequencing being another, that lend themselves to exponential price improvements. And if you look at genome sequencing, when we first sequenced the human
genome, back when you were probably working for Yahoo about 20 years ago, it cost about between $300 million and $1 billion to sequence our first genome. As of the start of this year, BGI, which is an outfit in China, reckoned it was going to cost them $100 to sequence the human genome. And so, genome sequencing pricing has declined much, much faster than computer chip pricing. And it's one reason, by the way, why we were able to sequence the coronavirus very on so quickly, because we just had sequences everywhere, because the prices had come down. So you have the same
dynamic of rapidly, rapidly declining prices. And essentially, that has meant also that biology has been turned into information. And so it takes advantage of all the benefits of Moore's law and computation, and artificial intelligence. So you combine those dramatic price declines with the huge demand that we have for novel drugs and replacing petro-derivatives, and materials, and so on. And I think you get one of the most fascinating industrial sectors over the next 20 years. ED HARRISON: Yes, definitely. And it was interesting that in that response, almost right at the start, you sort of moved into yet another area that I wanted to talk about. I mean, you talked about manufacturing a case
that's essentially like plastic, but better. Because 0 carbon emissions. Suddenly now, we're going from synthetic biology to renewables. And that's the second area that you think that there are exponential trends in place to go forward. I know that Raoul Pal, who's our CEO, he talks about renewables. He talks about how Europe, in particular, from a corporate
perspective, is perhaps going to really drive forward on that. What about the technology behind it, the exponential nature of that drive forward? Why is it that renewables are exponential at all? AZEEM AZHAR: Yeah. It's so weird. Why would they be? So the first is that the reason prices decline is because of learning. So the more an industry figures out how to do something, every incremental
unit costs less. Because they embed their knowledge in more efficient manufacturing. So this is not because they're buying more inputs. It's just that they're better at making them. And that's a kind of critical pulse of exponential price declines. When you look at solar, for example, solar panel manufacturing is really all about a 2D substrate on which you have to array a set of circuits. And that's a very similar set of disciplines to making a computer chip, which is basically a silicon wafer with a bunch of circuits on there. So we can kind of understand by analogy, and sometimes by direct transfer, why the things that worked in Moore's law should also work on solar panels. And so we've seen solar panel price declines that are just really out of this world.
And now, solar contractor utility, solar prices, are approaching kind of 1 and 1/4, 1 and 1/2 cents per kilowatt hour, which is remarkable. It's slightly stranger when you think about wind turbines. Because wind turbines are these really big, physical things. I mean they're like, sort of cruise-ship sized. 200 meters high, and so on and so forth. Why would they have exponential improvements? So partly, it's because of these learning effects that I talked about. But the second is this really nice idea, which is that the amount of power a wind turbine generates is related to the area that the blades sweep through. And the area is a function of the square
of the radius. So every extra 10% that you get on the length of a blade, you'll get about 21% extra generating power. And that is a non-linear relationship. And it's part of the reason why as we make these sort of linear improvements in the kind of underlying capabilities of wind turbines, we get this nonlinear improvement in their output. And so even though they're sort of big and heavy and need lubricants and so on, we're also seeing wind prices decline exponentially. And that's pretty remarkable. It's not the rate of Moore's law. I use the
definition for exponential technologies as 10% price performance improvement every year over a sustained period of time, or better. In wind and solar, it's between about 20% and 30% per year. In computer chips it's between about 45% and 60% a year. But they're all still exponential technologies. ED HARRISON: Yeah that is interesting. Now, the question is, when you said the one, I was thinking 1.1 times 1.1 is 1.21.
That's where the 10 becomes 21%. AZEEM AZHAR: Yeah, that's right. ED HARRISON: Does that mean that you need to have a bigger and bigger wind turbine, or how does that work in practice? AZEEM AZHAR: Well, we're already building bigger wind turbines. And over the last 20 years, wind turbine sizes have really, really expanded. I think the largest blades now are in excess of 80 meters, 88 yards in radius. So they are getting bigger and bigger. But actually, I don't think wind is just a game of huge size of individual turbines, although that helps. It's also actually just about being able to leverage the capital and build bigger and bigger offshore wind farms. And just before we
came on to do this recording, I read that there's a contract in South Korea now that is approaching $40 billion build the world's largest offshore wind farm. And the thing to think about this is that once you've installed-- you've put out that upfront cost, of course, your electricity roughly speaking runs for free. I mean, there's mechanical maintenance that has to happen. But it's very, very different to the kind of maintenance you need and refueling that you need in a fossil fuel system. ED HARRISON: Yeah. And right at the beginning of that, when we were talking about the solar part, you were talking about-- this is similar to a silicon type of production, in terms of the exponential part of it. It makes me think about
manufacturing technologies. That's the third area that you were saying where we are having exponential growth technologically. This isn't just about silicon wafers and solar, but other manufacturing technologies. What other manufacturing technologies, and how is it similar to what we've seen in Moore's law, which is Gordon Moore of Intel. AZEEM AZHAR: Yeah. So this final area is often known as 3D printing, or additive manufacturing.
And we've got a number of ways of making things. We can do what Michelangelo did, which is we can take a big block of marble, and we can chip away at it. And we're left with whatever's left. Or we can use a cast or a mold to pour molten plastic or molten steel into something, and mold the cup or the knife that we want. Or we can build something and additively, layer by layer, using clay, for sake of argument. And additive manufacturing now is being computer-controlled. And the power of that additive manufacturing is that you can actually create structures that you can't create through casting and molding, and you can't create through subtractive manufacturing, that is, chiseling away, or using a plane or something.
And the value of that is that you can build lighter, more flexible, but also stronger type of components. And so additive manufacturing has taken off in aerospace and in automotive, and in medical implants, where you can afford to pay extra. If you drive the BMW I8, which was their high-end sports car, it had a 3D printed part in the retractable roof. Of course, it's super expensive now. But so were computers in 1974. But there's a suite of additive manufacturing technologies. And they are all, on average, improving on a price performance basis at about 30% per annum. That is, roughly, performance is-- what we care about is speed
and accuracy, sort of pixel resolution, or printing resolution, and the variety of materials they can use. And there's just some really impressive breakthroughs. So what we would expect is the typical exponential pattern, which is slow and boring for a while. And then at some point, it'll get cheap enough and diverse enough as a technology that it will take off. And we've just seen in the last week or two, Desktop Metal, which I think is a listed company in the 3D printing additive manufacturing space, demonstrated 3D printing with wood. And it produces these incredibly beautiful, organic-looking designs that would have been impossible to build just 10 years ago, in any way, shape or form. ED HARRISON: The fascinating thing is the numbers behind that. Because as you were speaking, I wasn't thinking about the cost reduction in terms
of division. I was thinking about multiplication. Because we went through the example of 10%. So if you take something and you increase by 10%, you get to 1.1. Then you get to 1.21. Then you get to 1.331. But if you take something 30% and then you exponential that, you go from 1 to 1.3 to 1.69, to 2.197. I mean, the differential over time is massive in terms of that. Maybe you can
talk, as someone who's like a technologist and obviously probably math literate-- I used to be a math guy back in the day. Just getting your head wrapped around this whole-- I mean, exponents, it's massive differentials, just from that little example that I gave. AZEEM AZHAR: It's incredibly hard to get your head around it. And we started to think about how difficult it was back in the late 1960s, and the early 1970s. Albert Bartlett was an energy professor, physics professor out of the US, who regularly tried to teach people the difficulty of the exponential function. And the rough rule is that if you have a particular growth rate, and you divide 70 by that number, that's the number of years it'll take to double. So if
you've got a growth rate of 7%, which is a little bit worse than the average stock market return, that means that 70 divided by 7 is 10. Every 10 years, your thing will double. If your growth rate or improvement rate is 35%, that means it doubles every two years. So after two years, it's doubled. After four years, it's quadrupled. After six years, it's gone up eightfold. And it very rapidly gets out of proportion. Our brains are not wired to contend with this rate of change. If you can contend with this rate of change, then you stand a chance of doing really, really well economically. I'll give you an example. I was an investor in an AI company back in 2005. And the numbers were not appealing. Like the amount of computation they needed was just really
crazy. And I said to the founder, a guy called Barney Pell, I said, Barney, just talk to me about these costs. Because they're really high. Like, your server costs are really high today. Because what's going to happen long term? How does this turn into something? And he just said to me, three words. Pardon me, four words. Well, duh. Moore's law.
Which you referred to earlier. His point being that he could look at this and every year know that that line item was going to decline by 50%. Because that was the gift that kept giving, which was exponential technology improvement. And as a consequence, investors in this company did very well. And so I think that grappling with the exponential function, this rate of change, is something that we all need to get better at doing.And the coronavirus pandemic, I think,
was a great example of this. People looked at this. I remember saying to people in March, late February, stay at home. They were going, but, there are only two cases. And I'm like, yeah. But virus spread is exponential. It's going to be a lot of cases. We find it really hard. Evolutionarily, we're not wired to see it. Because we've never had to experience exponential processes in nature. They don't exist at our scale. And so as a result,
our brains are not wired to contend with them. ED HARRISON: If our brains aren't wired to contend with them, then as the touting of your book points out, the exponential gap can explain our society's most pressing problems. It just goes right to the social problems. We're talking about exponential change in a linear mind, with institutions, as you said, that are linear. What are the most
important things to think about in terms of what that means for our institutions, not just thinking about regulatory institutions. I'm talking about-- I don't want this. That's the first thing that someone would say, is that, wait a minute, slow down. That's too much change. Let's roll this back in some way. I liked it how it was before. Stop doing these things. What do we do about that urge? AZEEM AZHAR: Yeah. It's a really challenging one. So I think that there are some aspects
of this underlying process that are hard to avoid, because they are not top down. It's not anointed by Silicon Valley tech gods that this has to happen. It's actually every single one of us, acting in concert in the market, prefers the shorter route. They prefer the thing to be more seamless. And that sends a signal back into the market to develop in this way. And so we are, I think, on a path that continues this rapid pace of development. And in many cases, we want it, right? We want to avoid the worst parts of climate change. So we really do want renewables to improve
exponentially. And we really do want synthetic biology to come on stream at scale, quickly, in order to protect wealth and livelihoods, and quality of life, while maintaining our environment. So we want this exponentiality. I think the challenge has been that technology has developed too far away from ordinary society, and where democracy and decision-making take place. And part of the reason was that, actually, it was commercially expedient for technologists in the '80s and the '90s and the 2000s to present themselves as special.
Because that would allow them the headroom to build these enormous companies without much oversight. And the thing is that technologies don't just shape culture. They're shaped by the culture around it. And so part of my plea is-- and it's a book that can be read by technologists, and investors, and policy-makers, is to create that bridge, to give the technologist that sort of empathetic understanding of the impact of their technologies in the economy and society, and to give the rest of us a better understanding of the underlying processes of the technology. And to be very clear that while we might struggle to contain every aspect of the pace of change, we can certainly play a role in shaping the direction, and pointing that change in ways that we all think might be a little bit more beneficial. ED HARRISON: Let me give you some specific examples. Because I think that's good from a generic perspective. I'm going to drill
down. I don't know if you've seen this, but in the United States, there is a chart that shows productivity gains going like this over a 50 year period. You know, and the slope of the line is pretty constant. When you look at real incomes for the median wage earner, it also goes like this, up until, say, the mid '70s. And then suddenly, it flattens out almost entirely.
I would posit that a decent amount of that has to do with network effects, winner take most, and technological change. I think about, say, I go to the local grocery store. And half of the checkouts are automatic. It's technology that's checking me out. It's not a grocery store employee. So that person is therefore not needed, and there's a lot of
downward wage pressure. Now we're having capital getting a lot of the benefits of technology. And labor is not keeping up, as we saw with productivity. How do you deal with that from a societal perspective? Because to me that's where the pitchforks come out. AZEEM AZHAR: Yeah. That's the pitchfork risk. And it's a real risk, a societal, political risk,
a risk for every investor. We need stable, democratic societies. They make better returns than autocracies or societies riven by civil strife. And the timing is really fascinating. I do think it is largely an Exponential Age dynamic. Because it is between the
period of about 1970 and about 2015 where we see an incredible switch in the value of corporate valuations where it shifts from tangible assets that humans make into intangible assets that we can't really put our hands on. And it's related to the declining labor share. And I think that one of the historical coincidences that as an accident is that as we started to roll back the regulatory state in the late '60s and early '70s, pegged to the ideas of Milton Friedman and the monetarists, we strangely took the brakes off technological development, just as technological development started to go exponential. And so there was a slight issue. Because that idea of having no limits is really helpful when you're dealing with sclerotic industries like the car industry in the 1970s or '80s. But it's not what you need when you've got a rocket-powered tech startup. You actually need the other thing.
So I think that there is a moment of reckoning, because we will not want this political risk to magnify. This risk that essentially, 10 points of national income has moved from labor to capital over the last 45 years. It's an enormous amount. And the bulk of it, really in the last 30. And so we will, I'm sure see, in a number of countries, moves to make that adjustment. And I really view it as an adjustment. I think of it in terms of reversion to the mean, in some sense. It's about saying, at some stage of the journey, we have to swap our bike for our running shoes. No one expects you to complete a triathlon on the same mode. And so
we've enjoyed this moment where we emphasized innovation and entrepreneurship. And in that, we've laid the foundations for transforming our economy on this sort of sustainable, digital, exponential platform. But that came in a sense, at the cost of broader equity. And now that we've got that platform in place, and we can see the massive pitchfork risk that is emerging, we're going to move that dial. And I think you see this idea emerging from other domains. I've come at it from a technologist domain, but I know that economists are approaching it with their own theories. And
we're seeing a sort of similar picture. ED HARRISON: When you mention the pitchfork risk, and the pendulum swing, I'm thinking about the Great Depression. Because that's the perfect example of the pitchfork risk. When you think about, just from a macro perspective, there was a whole period of what I would call government intervention that was dismantled because of the stagflation of the 1970s. So the pendulum swung the other way. Now we're at a inflection point where, as I said,
you have this curve here. But then you have sort of a flattened curve. And that's creating this pitchfork risk again, similar to what we saw 100 years ago. The question is, things like UBI, universal basic income, are those going to be the answer? How do you deal with this risk shift that we've had over the last 40 years onto labor? Is it UBI, or how do you think about it? AZEEM AZHAR: I would start by trying to characterize what the problems are and what's needed. So in a dynamic moment of change, businesses need flexibility to grow fast, to shut down quickly if things don't work. I'm going to ask our business people to take risks. Risks means not everything will work out. So they need that flexibility. But on the other hand, we need to also ensure that the labor pool has the right level of security. I mean, partly for
the reasons of political risk, partly because you want workers earning earnings so they can spend and keep the economy dynamic and vibrant. And you also need to ensure that the economy has the skills it needs from the people who are able to work. So maybe UBI is something that works in some countries, or some states, or some cities. I think what you need is, you need to identify the issues and figure out the specific prescription. There's one model that I think has worked well in Western Europe in Denmark, which is called flexicurity. So it's the gives and gets. So the employer
gets flexibility to hire and fire. The worker gets security. Very, very high unemployment insurance, provided that they can demonstrate that they're skilling themselves and looking for work. Now it costs something, right. There's no free lunch here. But it's an interesting balance. And it's a prescription that works for their political culture. So I think that
those conversations are going to happen differently in different places. There's a group in the US called the Mayors for a Guaranteed Income. It includes the mayor of Stockton and a bunch of other cities. And they're running experiments about UBI. I expect we'll see more UBI. I'm not sure we will see UBI universally, because there will be other approaches. But I think the fundamental thing is we can have some agreement about what we need for our societies to function, and for our economies to be vibrant. And once we have that agreement, we can figure
out policies that specifically tackle them. ED HARRISON: So I'm going to make a hard shift here, Azeem, and go back to exponential trends. And one trend that people talk about a lot, that's really dominating the airwaves, is crypto. Walk us through how you're thinking about the crypto
sector from an exponential view perspective. AZEEM AZHAR: I think there are two really interesting things about crypto. One is that for me, it's part of the trend of computation. That it is developing new classes of methods of doing computing very, very rapidly. So of course, there are financial specialists who will talk about Bitcoin as a gold alternative, and central bank digital currencies and so on. But the thing that I find particularly interesting is crypto's ability to provide a new type of computing fabric, of a kind that we haven't really seen since the cloud, and before that, since personal computers. One of the interesting aspects of that is actually
crypto's governance capabilities. So when you get involved in certain types of crypto projects as a holder of tokens, you end up getting some kind of governance rights. That is, you can vote on the direction that that project takes. And I think that's quite interesting.
Because one of the exponential trends as a consequence of all this technology change is an increasing importance of decentralization of economic activity and localism. And so there's a nice concordance between crypto as this fabric that provides a new way of computing, but has baked into it this notion of governance. In other words, you and I are not simply customers of this service. In some sense, by buying into it, we also get to shape its direction. And I think it's early to say how that will end up playing out. But what we've seen in the last two or three years is that there's real progress being made outside of the incredible price appreciation of so many of these crypto assets. ED HARRISON: When you talk about incredible price appreciation, the first thing that comes to mind is bubble. Technology, the dot com bubble.
Let's address the elephant in the room. The Federal Reserve has zero rates. You have negative interest rates throughout Europe and in Japan. Is it possible that we're in the midst of a technology bubble? AZEEM AZHAR: Well, a bubble is expectations running ahead of reality in that context. And it's possible. Of course, it's possible that crypto won't deliver. It's possible that there will be a huge price collapse. And
value may disappear. But I think that there are some really, really deep fundamentals, particularly around the four general purpose technologies that I talk about-- computation, biology, renewables, and manufacturing, that are deep, deep, deep fundamental trends that stand to fundamentally change our ways of life. What happens to assets that are pegged to those trends, and the rational or irrational exuberance that goes around them is something that I'm not best to judge. But is there something fundamental happening? I would say yes. ED HARRISON: OK. Great. Because the question I have is, let's go back to the dot com bubble. It didn't change the pace of technology. The fact that we had a bubble and then it crashed.
You know, innovation kept on happening. I mean, what I'm hearing you say is something fundamental is happening. And it's going to continue to happen, irrespective of the valuation attached to the assets behind that happening. AZEEM AZHAR: Yeah. I think that's right. But also
the difference between now and the dot com bubble is where the hell else is the capital going to go? I mean, there's so much more of it. Because you've got sovereign wealth, you've got Chinese wealth. There are no bonds you want to buy. Do you really want to back a company that doesn't have an electric vehicle platform? No, you don't. Do you want to buy a coal field? No, you don't. So you don't have-- as an investor, you don't have any other way to get a return on your assets, really. So that creates, I think, a natural support
for prices. Because what am I going to do? Well, I'll take my money out of Tesla, and I'll put it in a Swiss bank account where I'll earn minus 0.1%? That's no trade. I'll take the risk on Tesla, thank you very much, you know. So I don't know where the cash goes, other than banking on this future. ED HARRISON: Yes. Well, I'm going to leave it there. I think those are good answers. And give you a chance to tell us. Because
first of all, when your book is coming out, and whether there are pre-orders, and where we can find out more about your book. AZEEM AZHAR: Yeah. No, thank you so much. So the book is out in September. And you can pre-order it from Amazon if you just search for The Exponential Age by Azeem Azhar if you're in the US, or just Exponential if you're outside of the US.
And it's available now. It's actually doing really well. And if you order it from Amazon, you will get it the day it comes out, so the day or two after it comes out. And you can always find me. I write a weekly newsletter, Ed which is on exponentialview.co, or just put exponential view
into Google or Bing, or whatever your search engine is. And you'll find me that way. ED HARRISON: Excellent. You have to promise then, Azeem, that when your book comes out in September, that we can have a follow-on discussion. I'm sure that I&
2021-07-30 19:22
