Techno-Confusionism: Is a New Cold War Happening? - Olaf Groth | Endgame #210 (Luminaries)
The income gap, we always talked about, you know, high, medium and low skills, but frankly, we are not drifting into an era where it's high -skilled or low-skilled. That rift between you and that avant-garde is widening even further, right? Who is to blame people when at some point they feel so left behind, they're working two or three jobs, can't provide for their families, have no educational perspectives for their children? Who is to blame them at some point when they start throwing Molotov cocktails, right? We need to dedicate more time to the computer science folks understanding humans and those who are not mathematically, scientifically inclined, right? Understanding data and AI. We have left droves of people behind in that globalization 1.0 paradigm, and we need to recognize and acknowledge that rather than saying, how can we just put a band-aid over it and get people to just accept? So we do have to break through that, and that's where we need, again, wise policymaking, and I don't think we have that yet.
Hi friends, today we're visited by Professor Olaf Groth, who is a professor at practice at Haas, the School of Business at UC Berkeley, but he also gets involved with UT Malaysia. Olaf, thank you so much for gracing our show. It's a pleasure to be with you, Gita, and your audience, of course. Thank you. Thank you.
Tell us about how you grew up. Uh, you were born in Germany and you got stuck in the U S for the last what, 30 years, and you've been sharing your wisdom to a lot of people in the world. Well, thank you.
I appreciate that. Yeah, I grew up in northwestern Germany between the trading and banking town of Düsseldorf and the Dutch border, small rural community of lots of lakes and fields. So very agricultural setting. And I was born to parents who were both children of World War Two and were both blue collar, had not gotten any opportunity to go to college, but whose dream it was to instill, you know, a more intellectual grounding in their son and so enabled for me this career very much that I benefit from today. And that enables me to have an impact on other people by sending me to college, the first one in my core family and and then later to to study in the United States.
And I was only supposed to be here for about a year and then finagled my way into staying what is now 33 years, shock and awe. Who would have been more influential in shaping your educational journey, your mother or your father or your siblings or friends and. They were both instrumental, but in very different ways. My mother was always the very studious one, very diligent about work, and instilling in me, of course, a certain diligence about getting your work done ideally ahead of time. I think my father was more of the intellectual in the family, and I think tended to opine on societal issues, on economic issues, on business issues, more and more frequently, but by his own self-admission, didn't have the tools really to do so in a sophisticated way or for an audience that he needed to explain it to.
And so I think that combination of the diligence and the intellectual curiosity and the drive really to do something in the public interest, or at least something that's critically reflective, right? Both of them gave me that. I think the latter part, that was my father. And I think both of them, having been children of World War II, and having been through what was called the economic miracle, the Wirtschaftswunder at the time of rebuilding, had that building element in them as well, and that stuck with me too. The idea of building your own venture, or helping build some program, or some academic strategy, or not to sound too grand, but helping build a country, right? Which is what they really had done. I think that ethos stuck with me. Got it.
in addition to your being an academic, you're also involved in Cambrian futures. Talk about that. Yeah, so I'm the the founder, a co-founder and CEO of Cambrian Futures, which is a an advisory think tank that provides clarity for the emerging technology economy around the world. So what that means is we do research, we create insights, we generate foresight on the role of emerging technologies like AI, data, brain computer interfaces, genomics on policies on, of course, corporate strategies as well, investment strategies.
And we hope decision makers see the future evolve more clearly, because there is so much noise going on and so much hype around tech these days, that I think it sometimes takes a little bit of a clear-eyed view to cut through that. So I work that with six other people who are mostly in the United States, but a few in Europe and elsewhere, we have a network around the world. But we're all more or less senior trusted advisors.
This is not a traditional consulting firm. What motivated you to write remobilization? Yeah. You know, this was really between me and two colleagues of mine, Mark and Terence, who during the pandemic saw how there was this swelling of anxiety around the world, of course, understandably, because most of the people alive at the time and in junior or senior decision making positions had never been through anything like this. There was this new phenomenon of this tiny little virus, essentially, crippling the global economy overall.
And we felt that it was incumbent upon us really to help people starting with our students see through that pandemic and understand what lies beyond. The pandemic was really a prism through which we viewed the current dysfunctionality and really the systemic fragility of our global economy or global 1.0, as it were. And so we wanted to use our skills and our networks to unpack that for people and show them, hey, why is that going on? What exactly is that fragility? And how do we see through this and see there's an opportunity to flip that chaos into a new horizon that we're all better off with? And so we decided to first do a webinar that anybody could join that then became another webinar. And before you knew it, we sort of thought we should create this book to have more lasting impact. So that was really the impetus of it.
You talked about the six C's, you want to talk about some of these or all of these. Yes, so we started with COVID, but frankly, this is not about the virus itself as much as it is about the science, CRISPR, and of course, cognitive technologies that enabled vaccines, as well as, I think, a clear-eyed view of why we couldn't get a grip on this virus as quickly as we should have wanted to around the world. So the first C is COVID. And then, of course, there are cognitive technologies that I've already mentioned, so that's AI, data science.
We briefly talk about the coming wave of brain computer interfaces that are already here, and now awaiting commercialization and genomics, and the role of AI and genomics. And then we go into cybersecurity, because you can't really talk about all these digital technologies or even life science technologies anymore without talking about cybersecurity. Crypto, because we believed that crypto and still believe that crypto is much more than coin. Coin is important, and we can talk about the ups and downs of Bitcoin versus Ethereum and other.
But really, the underlying message of crypto is one of governance. It's Web3. It's blockchain as an idealized version, and it is an idealist version of more diffused, democratized governance with all of its promise and flaws, of course. And then existential threats like climate change.
Clearly, you can't really write a book like this about the future of global 2.0, without talking about the one threat that could annihilate all of us, and that is climate change. And finally, China, that's a big C, because you can't really do anything about any of the other Cs without co-thinking China and engaging with China in vigilant partnerships.
And so those are the six Cs. Really, the C that is the cognitive technologies really spans all of the other Cs, and I'm happy to talk about why that is. But that is the command and control mechanism.
Some people call it cybernetics, which is an old term for technology-enabled command and control. And without that, really nothing else works anymore these days. We can talk about 4 billion people around the world who still aren't digitized, of course, but the power center clusters around these digital technologies, because they really enable the further proliferation of that power. And at once, they hold a lot of promise for us to do a better job managing the other Cs and many more things around the world, because things have gotten so complex and so unwieldy that the human brain and human institutions alone can no longer manage the interlocking systems of systems. So we're damned if we do, and we're damned if we don't use cognitive technologies like AI. you know, with the world order becoming a much more multipolar kind of world, intuitively, there's just going to be more proliferation of risk, right? What gives you the confidence that there is an ability to remobilize, you know, by the way you're focusing on each one of those sixies.
Yeah. So we see indicators that people realize climate change is a global threat, and it's existential. The fact that we haven't solved it yet is, of course, a catastrophic reality. But I think the awareness is there that we should be solving it with policy and technology in lockstep. And I think that gives me hope that we will be too late to stay under 1.5 degrees,
but it still gives me hope, call me an eternal optimist, that once it becomes clear as we go beyond 1.5, how much we will be paying for not meeting that goal, that the financial interests around the world will wake up and exert a lot of pressure on the governing institutions and on people in general and other businesses, that something needs to happen very, very quickly, because the fallout for investors, for insurances, etc., will indeed be catastrophic. We're seeing this now here in California, where insurances are starting to pull back. The governor is saying, no, no, no, if you want to do any kind of insurance business in California, you've got to offer people fire insurance.
But insurance companies are saying, but the proposition is not viable for us. And so change can be induced through financial pressure and changing incentives for financial institutions. So that's one area of hope.
The other area of hope is that I teach leaders, of course, of all ages, but I also teach the very young ones. And for all of the skepticism that everybody around the world has about the young generations, ranging from tremendous amounts of depression, of course, induced by social media, etc., and also, I think, some questions around whether they still see democracy as the guiding star, which is a bit of a Western notion, or Western Northern notion, I realize that. Yes, all of those things need to be looked at. But I also see a tremendous resolve to be more socially minded, and to be more equitable, and to be more fair in society, and to not focus on just a singular career that will make you a lot of money in exchange for working 70 hours or 80 hours a week, but rather working portfolios of different roles and careers in parallel, some of which tend to have these social and societal elements in them.
So that's another area of hope. And then, of course, fault me for, you know, bringing a bit of a Silicon Valley view to this, and I guess optimism on that front as well. I do believe that if we govern technology thoughtfully, and we design it thoughtfully, that technology can be a net tool for good. We all understand it's a Janus head, and for everything good you can do with technology, you can also do tremendous bad. And currently, the call is still out whether we aim, which way we're going to aim. But the hope is still there that we're going to turn the corner with technology when we interviewed environmental scientists and economists and the people that go to COP like you do.
They openly acknowledge that traditional government-led processes are failing the world. I think the UN is doing the best it can, but I don't think we can call that a success to date. And they're now turning toward technology saying, hey, you all in Silicon Valley and all these multi-billionaires and VC firms, you really have to step up. And that's a flip. We had previously always heard, no, we actually need regulators and policymakers to step up first to set the right pricing mechanisms in the market so the entrepreneur knows what to go after and which way. And I think for the first time, governments are saying, no, actually, you have to lead the way.
And that's a tricky message. How do you do that without the right pricing mechanisms, pricing signals? How do you, well, I mean, you know, we've talked about this earlier. There is a wide gap between what the developed economies and what the developing economies, much less the underdeveloped economies could afford, right? I'm not concerned about whether or not there is technological capital to address climate change, but the economic world with all is not there. For most people on the planet, I would argue 85% of the people on the planet they can't afford the technological capital that's being availed by the developed economies. How do you think would be a realistic way of bridging that gap? Because it just doesn't look like it's being bridged efficiently and effectively. That's the big question mark because the deeper we get into artificial intelligence and data science, of course, which promises to give us a lot more insight about how the world works and how we can generate growth, the deeper we get down that path, the clearer it becomes that you need massive amounts of capital to do so.
And that is why the big hyperscaler platforms and governments around the world are dominating that playing field. Very clearly, that's the United States followed by China. And the rest of the world just simply does not have the financial resources. Europe in theory does, but cannot get its act together on a common market and on the right incentives for entrepreneurs. But even if Europe were to solve its own problems, there is still all the rest of the world.
And I hate to hesitate to call it the global south, but it includes, I would say, about 70% of all countries around the world. And that is a big concern. It's money, and it's the development of talent for which you need money. And then it's data and the data infrastructure. Because even if you have a lot of data, you have to make it available along a certain infrastructure.
And to do so takes that infrastructure investment, it takes that talent investment, and that is not currently enabled sufficiently. Part of that is that the development institutions around the world aren't digitally native either and lack the competency there. And part of this is that this is just progressing so fast that government institutions aren't equipped to jump onto that running train, as it were, and it's not going to slow down anytime soon. Now, on the argument that you've alluded to, that there's not enough economic war with all, there's not enough educational trajectory, there's not enough ability to compile and structure the data points among most, if not all, of the global south or developing economies. Are you confident that we're going to be able to, in the context of climate change, attain carbon neutrality by 2050 or 60? No, I'm not confident.
I remain hopeful. But that is in large part my personality that I tend to see a path where others don't. And, and that's an aspirational path oftentimes. So I do believe that the deck is stacked against against us in that regard. Now, the entrepreneur in me says the deck is always stacked against me. And my probability of failure is, is greater than the probability of success.
So it's a normal situation to be in. But, but of course, that is not to diminish the potentially catastrophic impact on billions of people around the world. And we describe this in the book, what's going to happen when the global south is not enabled by industrialized and post industrial economies to adapt on time. We see, I think we say in the book upwards of 500 million people migrating north. That is going to, of course, create massive displacement, not just in the south. But in the north as well, massive political problems.
So it's a shared problem that I think most people in the north are not seeing as such. But you see what's happening here in the United States with immigration as a political topic. I know it's a topic in China and elsewhere. It certainly is in Europe as well.
So it's a shared catastrophe waiting, waiting to happen. Technology can help, but we need more capital deployed. Now, some, I think some of the hyperscaler, the tech platforms here on the west coast have realized this.
I think there are some that are quite foresightful when you look at what Satya Nadella does with Microsoft, for instance, billions, billions invested in, in training as well as sustainable energy sources or at least renewable energy sources, including nuclear, of course. Then I think, I think that gives me hope that that's where the capital is. That's where the know-how is. If we now have wise policymakers in the global south looking at those assets and saying, okay, how am I going to, I guess, welcome that capital and that expertise, but also leverage it to create indigenous local capability. And there are, of course, very smart examples in Southeast Asia, East Asia, and elsewhere, Africa, for instance, as well, where that has been done.
But that is a precondition, right? Because what you want to do is leverage that northern industrial, post-industrial economy knowledge and talent and capital, while also building local capabilities. Because you need those local capabilities to get to the next level of growth that can't just be fueled by foreign investment. There needs to be local capability building on must. And currently, the few billion dollars I see invested are not enough. They're a good indicator, but not enough, clearly.
If you take a look at how much China built on coal, it built about 43,000 megawatts of power generation capabilities using coal last year, which is about 13 times the amount that Indonesia built. It just shows that as much as everybody wants to be environmentally friendly, the realism with which countries like China want to pursue AI, I think climate change is a tough one. It's a wickedly tough one, and it's the issue of short and medium term necessity and long term aspirations, or medium to long term aspirations. And we're not just seeing that, by the way, in Southeast Asia, we're seeing that elsewhere, and not just in China, we're seeing that here in the United States.
Kamala Harris has said, look, she is in favor of fracking. I don't think she says that because she likes fracking or its environmental impact. I know, but we have got to rely on it.
Yeah, exactly. It's a medium term necessity if you want to stay energy independent. And for some very good reasons, the more you depend on foreign oil, the more you get embroiled in geopolitics around that oil.
And so yes, I think she, for instance, is a realist on that front, and understands that if he were to just go with renewables, he sacrifices potentially slower growth than even he has now. And that means social upheaval, political upheaval. And as my friends in Beijing have said, when 1.4 billion Chinese are at war with each other, the rest of the world will suffer as well. So I think we need to respect that.
But on the other hand, China has also invested whatever it is, $300 million in solar. And that's, I think- Oh, they're well ahead of many. Well ahead of many on renewables. but they still rely on coal. They need the energy. I want to move on to the cognitive economy part, but before we get there, you've used five letters in your book, F-L-P-I-T.
Talk about that. Yeah, we call it FlipIt, F-L-P-I-P-I-T, FlipIt. And it's about it's a framework that helps you flip your mindset and really your strategy from this moment of chaos, or as Genevieve Bell in the book called it, this liminal moment. We're in a liminal moment where we've thrown everything up in the air. It's very anxiety inducing.
We're not trusting our institutions anymore. And they're not providing for us the public goods for us anymore. And so it feels chaotic, potentially catastrophic. We've been here before, after World War II, of course. And there have been other periods in human history where we've been at moments like this.
And now it's all about looking at the pieces that are up in the air and seeing when they come back down to Earth, as it were, how do we arrange them and maybe add some new ingredients in to create a better system. There were a number of thinkers after World War II that got together here in the United States that thought about what kind of system do we want. And the thinking around the United Nations, et cetera, was in part spurred by those deliberations.
We need that kind of thinking now. And so we have set this flippant framework and help you understand, what are all the forces that are pushing us in various directions, putting all these pieces up in the air? And when they collide, what kind of logic do we see unfolding? And how do we shape that logic? And then the question is, what kinds of new patterns do we see or do we want? And what kind of action do we take based on that? So it really takes you from looking at forces and building blocks up in the air, bringing them back down, looking for better logic and new patterns, better patterns. Exactly, first print. Well, in fact, though, Gita, it's really about zero principles. And that's really, we picked this up from interviewing Brian Johnson, who has been in the news a lot with his Don't Die initiative, the former CEO of Venmo and Braintree.
And he pointed us to this concept of zeroth principles. And that's really to say, we got to go beyond first principles. First principles are sets of rules that we have accepted to be axiomatic. But they may, in fact, just be human constructs because we can't think beyond. And so, for instance, first principles are always, more growth is better. I think you and I know that that's not always true.
But our markets are still very much focused on that. And I'm not talking about degrowth necessarily, but smart growth. Well, what does that mean? And so getting to understand new building blocks and seeing those building blocks and those new logics, those new rules, is really what leaders need to be doing today.
Not everybody does. I would say the vast minority does. The majority still plays to old rules that are set in global markets, whether that's Wall Street and the city, meaning London, whether that's gap rules and principles. But we're playing to old rules that are no longer fit for the world we're in now, much less the world we need in order to step away from the climate change brink or World War III, for that matter. Interesting. Let's get into the cognitive economy part.
You've been spending a lot of time talking about AI. I get the sense, and I've shared this with some people, that it's not the AI narrative is not being pushed forth in an adequately multidisciplinary manner. And intuitively, just since it could be problematic, that it's not roping in the environmentalist, the culturalist, the sociologist, the economist, whatever dimensions beyond technology, right? Do you share that sentiment or pulse? Yeah, I think we are bifurcated today between those that are the accelerationists, as they're called, that are saying, look, the train has left the station.
We need to push as hard as we can to get the greatest degree of AI advance over the next couple of years. Everything else we can figure out, things will fall into place. And then you have what's commonly referred to as like the Church of Doom or the doomsayers that are typically people concerned with AI safety, existential threats, and generally the AI ethics and governance crowd, which is very unfortunate because governance, as you would know, is a tremendously important toolset in order to channel flows, whether there are flows of money or flows of technology, IP, or talent, or data, or genetic code. It's directing it in ways that yield the highest advances. And of course, you need to define what that actually means for a given society or a given organization.
But governance is a very pragmatic approach. So we shouldn't have this black or white understanding really tribalizing each other. We should come together in the middle and say, yes, look, we want growth, but we want healthy growth, sustainable growth at a rate that is palatable for society, and that doesn't just create more wealth for those who already have great wealth. And that's typically those who own the technology and the IP or the capital that allows it to build it into ventures. And so governance really should be sitting in the middle. But yes, I think currently the driving forces behind AI advancement are clearly still in the accelerationist camp, if I were to use that language.
I think slowly we are starting to look at, but how do we bring governance in there? And there are some great precedents. There are some tech companies here that are outstanding, blazing the trail on this. But it's not yet the majority of platforms and tech companies, and it really needs to be.
Some of that will be driven by, again, coming back around to the importance of investors in the financial markets as a change mechanism. Some of that will be driven by investors who say, number one, you're not getting us the right ROI or the right quantifiable benefit, bottom line, top line. And some of them being insurers who say, if you all screw up with AI, I have to pay. And we're seeing the role of insurance is actually quite important in climate change.
So it will at some point become so in AI as well. If you take a look at the amount of money that China and the US are plowing into AI, it's disproportionately much larger than what I'm seeing Europe is, much less Southeast Asia. I want to ask you a few questions based on these observations. Why is Europe seemingly missing out on this digital revolution? Yeah. So it pains me to say this since I hail originally from Europe.
Europe has lived obviously through world wars and a very painful history of bloody conflict for the past many centuries. There is a consensus mindset in Europe that lends itself to thinking about stability in a look, for instance, at where I came from, Germany. There is a predilection, I think, in Germany to be uncertainty avoiding. When you try to conquer a new frontier, whether that's AI or anything else, genetics is another one of these, you need to be able to experiment to take certain risks.
But when the majority of your population actually says, no, no, no, no risk whatsoever. Don't mess with my data. Don't mess with my genetics. Don't mess with my work conditions. And when that becomes the governing principle for everything you do in AI and data, it's clear you won't make any progress. So the reason for why Europe is not leading in artificial intelligence, and I think we do have to clarify what that means.
Obviously, there are leading lights in AI science in Europe and AI scientists in Europe play at the very top of the science pyramid. That's very clear. We also observe that there is massive amounts of entrepreneurial AI talent in Europe, but that combination does not seem to get Europe off the couch when it comes to actually generating creating ventures that can come to global standing and project Europe's view, which is one of individual dignity and the protection of the individual throughout the global economy. Some of that has to do with policy and regulatory frameworks, with capital formation frameworks and incentives.
Venture capital is nowhere near where it should be, given where Europe is punching as a global economic region. But really, most of it is rooted in the fact that Europe has a cultural legacy and a cultural DNA of many centuries of bloody conflict and has finally gotten to a stage up until recently. Well, I guess you could also look at the conflict in Yugoslavia as the first element of this. Really, this peace dividend that they thought they had after the Cold War of a successful attaining of regional peace, that's really falling apart now.
The instinct of clinging very hard to stability in light of history and in light of recent events and making that the guiding North Star for everything you do is inhibiting advances in artificial intelligence and data science. The instinct is to regulate first, make sure that you shut off anything data related because that could be used to abuse people. And I might add, with some good reason, when you look at some of what the US platforms have been doing and what the Chinese platforms are doing, it's not that Europe is wrong, it's that Europe is ill-balanced. What happens then, of course, is when you don't create enough spaces for experimentation that you will never get to that frontier of actually transporting, agenerating enough growth for yourself, but also transporting European values into the rest of the global economy. They're trying to do that with regulation and it's working to some degree, but it's not sustainable because people want livelihoods. They don't just want to have protection.
Eventually, they need good jobs and we're seeing what's happening with the European economy today. I say a bit provocatively, Europe needs to watch out that it doesn't die an analog death. The GDPR, is that a manifestation of how or the extent to which Europe has not been a beneficiary of the digital revolution? Yeah, I think the GDPR was the first the sheer.
rigidity of it. That's exactly right. And I think we need to recognize it was a stake in the ground, the first of its kind, that said, well, wait a minute, there are other elements here, as you were saying earlier, that we have to pay attention to. And those are fundamentally of ethical nature. However, it was overly rigid.
I think the European Union took a receipt of that. I think there is widespread consensus that it was a dual-edged sword. I think the EU-AI Act that we got last year is, I think, a more well-balanced set of instruments. And it does acknowledge the role of experimentation with sandboxes. I think it does not yet address the needs of the European entrepreneurial landscape.
And some of that has to do with the common data market that is really not in place. It's in place on paper. But what's on paper doesn't really matter to an entrepreneur as much as what's happening in the street, actually, and in agencies, actually, every day.
So the fact that the administrative procedures are not homogenized means that an American, or for that matter, Chinese or Indonesian entrepreneur, can't scale out efficiently through a 500 million people market, which theoretically should be bigger than the US market. But it won't be for the foreseeable future because of the disparity and the fragmentation. And so that needs to be addressed. I think a more pro-entrepreneurial view, more entrepreneurial voices in regulation need to be heard. I've been hypothesizing that the Internet has been very good at democratizing information, but it has not been good at democratizing ideas.
It has not been good at democratizing economic capital, as a result of which we're seeing certain economic phenomena that are disturbing, rising inequalities of wealth, income, opportunities. And now we're seeing rising centripetality of economic development, i.e. GDP per capita growth in primary cities, much more accelerated than that in secondary cities, globally speaking, developed, developing, and underdeveloped. And it's a bit paradoxical because in some developing economies, the talent and natural resource are actually in the secondary cities, but it's been extracted and pulled to the primary cities, right? And I just think intuitively AI is going to further exacerbate this pre-existing condition or situation. What's your take on this? Yeah, I think that's right. I think economists have actually shown that AI and the importance of data and compute power and very expensive talent to develop it and to scale it out will actually lead to a further concentration of wealth in the hands of innovators in these innovation capitals around the world.
And they're usually first-tier cities. Not to say that San Francisco is one, but I think San Francisco and the Bay Area and Silicon Valley, certainly in innovation terms, is that. And so that's certainly true. If you think about the evolution of the Internet, the idea was a democratization and a diffusion, but pretty quickly in Web One, for instance, we of course had the telcos and we had the likes of AOL and MSN that had the infrastructure to build and then to add application platforms on top. And then we got the over-the-top Internet platforms, as it were, that are now forming what is maybe technically, legally not oligopoly, but structurally in reality in terms of people's lives.
It probably is an oligopoly. And so now you have all this concentration of data and money. I mean, these trillion dollar valuations speak for themselves. Where is the next wave of more diffused innovation going to come from? And that's, of course, where the idea of Web Three came from and blockchain and the idea of crypto.
And so we're going to have to see how far that will carry. But it's supposed to go back to the original thought of the Internet being a more democratized tool for self-expression, but also for growth and for ideation and innovation. What's, what's the risk that only China and the U S are just going to be completely dominant in the next 10 to 20 years or 20 to 30 years in the field of AI at the expense of some of us in Africa. Yeah. Southeast Asia, and many other developing economies, to some extent Europe.
Right. Yeah, to a large extent, Europe, because when you look at Europe, it has all but just a handful of platforms. I mean, we have SAP, we've got Spotify, but really, it gets very scarce beyond that.
So the risk is substantial, and there are mitigating factors, right? And again, being the optimist, I'm hoping that those factors will come to bear. Clearly, we already talked about we're at this point today because we have all this data. Where is that data? Now, it is stovepiped behind walls, either in the big internet platforms or the data brokers and advertising networks that do business with them, or of course, large governments. But that's a handful of very centralized institutions. The same with capital, and the same with the talent that can be bought with the capital. I still remember one episode where I had a meeting at Stanford, and a fellow professor came in and said, I'm sorry to be running late from a meeting where a young AI PhD was defending their dissertation.
And he said, we tried to offer him a tenure track role at Stanford right off the bat, but he had to decline because Facebook offered him a million dollar total compensation. And Facebook can do that, and many companies on the West Coast can, but 99.9% of all companies around the world cannot.
So how will that be mitigated? Well, institutions like Berkeley and Stanford, and of course, many around the world, UT Malaysia comes to mind, right? And I'm sure there are a whole slew of institutions need to think about how they put more talent through their pipelines. That's an institutional capacity problem. And at the same time, we're also now starting to see and hear from cutting edge AI talent that, in fact, the era of big data may be coming to an end, at least the necessity for big data for every single application area may be coming to an end. We may be able to do with small data, wide data, which is pockets of small data for certain types of use cases quite well.
And this idea of billions and billions of parameters in these models that may be coming to an end at some point as well, especially when you look at the success of open source or semi open source models like LAMA and others. So we shouldn't straight line extrapolate from here, but certainly currently that's where we are. Yes, that it's not looking good for countries that don't have capital, don't have the installed base of data or talent. I just think it's going to get much more elitized, you know, at the rate that only a few companies are going to be out spending the other billions. Yes.
And that is why we need policy makers to say, you come to my country, you have to train X amount of people. ...shackled. There are some that are and some that are quite forward leaning. Right. So so because of my involvement in Malaysia, for instance, I can say that the least the aspiration and the political will is. I was thinking of here.
Here. Yes. Yes.
So so well, we have we have various government commissions saying we need more talent, whether that's on the national security front or economic development, more evenly spread throughout the country. But of course, yes, the the the the giant sucking sound into Silicon Valley and the East Coast is is is definitely a formidable force. Yeah. I think there's more hope in terms of re -governing or up-governing in other countries, including some of us in Southeast Asia. I want to, you know, there's this inherent apparent structural limitation. I call that energy.
Anytime you hear the gods of AI talking about 10Xing, 30Xing, the pre-existing capacities, they just simply stop short of mentioning the energy requirement. And I think, you know, everybody's not going to be able to fulfill the incremental energy requirement, which I think is going to just serve as a structural limitation in terms of how much AI is going to grow. And that I think will speak very nicely for utilities companies around the world. Yeah. Well, they're going to feel the pressure very soon to provide greater load. But where is that load going to come from? I think that's a fair question.
I think certain companies like Microsoft, for instance, have recognized that and are now, as you saw in the news, Microsoft just bought Four Mile Island and is bringing that on stream, wanting to convert that into safer nuclear. I think part of that is Bill Gates' investments in smaller and safer nuclear. China is building, I think, a nuclear power plant per month. They're doing SMRs, small modulars. Exactly.
Alongside coal, of course, as you said. I'm not sure that nuclear or even a hard ramp up on renewables is necessarily going to do the trick if we continue on this path. But again, I think the mitigating factors are that even many people in LLMs and in transformers right now are saying, look, we're going to hit a ceiling at some point.
These LLMs are not the path to what many aim at, which is artificial general intelligence. It will require different clusters of different technologies coming together, and LLMs are really only one thrust of that. There may be a ratcheting back on the energy requirements at some point. If that doesn't happen, then yes, energy will become a limiting factor, and that will put a lot of pressure on systems, on inflation for that matter, because it'll become more expensive in people's lives. And again, we're talking systemic fragility.
In your book, you talked about, well, in your lectures also, you talked about the three shifts, right? The shifting from things to insights, west to east, you call that techno-confusionism, and then third one would have been the shifting from, I forget, Um, shifting from well west to east and then things to insights and Uh, it's my book. I should know the answer. Um, let's. about the two shifts first, until I remember a third one. But it's interesting how you talked about techno -confusionism and how things are shifting from the West to the East.
Yeah, I think I think there is a pronounced shift that is hard to swallow for many in the west, or the west dash north, because they're not used to a paradigm where Asia co-innovates and co-leads, but clearly China has shown that it is playing at the very forefront of some of these emerging technologies. And, and in short order might be co-leading and so, so that is a paradigm shift and a mindset shift on the part of the many that are then of course conflating that innovation power with the power of values, specifically from and in China. I call it techno-confucianism, because that particular mix up until this moment, and there is a high degree of volatility right now in China, China and around Asia, of course, up until this moment, we have seen a very technology forward economic development paradigm in China that has lifted what 6-700 million people out of poverty, arguably at great cost, of course, when we look at the cultural revolution, 30 million people dead, etc. So yes, if you wanted to go there, then it's a question of trade-offs. But by and large, I think China's population today would say, well, that was a success and it's very technology forward. And they have not yet seen some kind of cataclysm that flows from that technology or the government's control of that technology.
And so the experience has been, that's a positive paradigm. And it's a very pragmatic paradigm in terms of the role of Confucianism, there being a set order and a structure to things and who holds control and who has to be revered and respected for what reasons. But I think what we meant in the book was that that stands for a pragmatism that is very East or Southeast Asian. And that is, look, whether you call it democracy or whether you call it any other name, whatever the paradigm is, it needs to provide for people. And I'm not saying every government by far in Southeast Asia or East Asia is successful at that.
But there are notable examples of where that has been very successful, where really it's the outcome of governance in terms of public goods and the increase of living standards that has been the determinant of success and of satisfaction amongst the population. Singapore obviously comes to mind as the shining example. Korea and Taiwan are not far behind.
We'll have to see whether Vietnam can do it or whether Thailand can do it. You seem determined Vietnamese. That's right. Yes.
And so we'll have to see whether the Vietnamese can be as pragmatic as the Chinese have. All indicators are that they might be. But there's also still a lot of corruption that I think China has addressed head on and keeps addressing head on. And I think Vietnam is showing indicators that it will as well or has been as well. But it doesn't quite have the longevity yet. The proof case has not been as long and sustained.
anticipate or foresee a further decoupling in the digital space between China and the U.S. Oh absolutely. We're seeing an attempted decoupling.
I think there are lots of points of friction around that. You saw Elon Musk going to Beijing to ask to please take some of that data that Tesla's cars are generating in China back to US servers at the same time where we are saying to China you cannot have US data, you know, US TikTok user data. So that decoupling is painful when you are economically still very much coupled.
We have not seen a decoupling. We have seen a decoupling economically in certain very sensitive technology areas like AI, like chip technology, compute technology, but the economies are still coupled. They're coupled through longer and more complex supply chains and so that inhibits efficiencies. It may increase resilience because you are now no longer beholden to just one place and a direct link, but certainly efficiency has not increased, has not benefited.
So I do see an attempt, but there is that tension of but we can't really do without each other in the global economy, nor should we, because there are problems like climate change, migration, education, public health, even collaboration around natural resources, whether you're thinking about the Arctic or you're thinking about space where we need, we can't. Astropolitics. Exactly. Otherwise we're back into that Cold War paradigm and that's not going to end well for anybody. If you've talked about Southeast Asia, China, Europe, US, what's your take on India in the context of the digital revolution? Yeah.
So India is stuck at that middle income trap. And we'll see whether Modi or whoever comes after Modi is going to be able to manage India out of that. India is now the destination of choice for many Western companies, of course, sanctioned by US policy. So we've seen that in installed base of capacity or of talent and know-how in semiconductors.
And so the question is, how fast can that really be the alternative to China? It is mainly a services-driven development model. But the proof's in the pudding right now. India is very much also politically a fence -sitter, as it were.
And we call it that in the book, because it establishes itself as a partner to the United States on certain fronts, but then also sources weapons from Russia on other fronts, and I believe energy as well. And so there is a host of countries around the world that will keep doing that. Singapore has done that for ages, right? I have friends who served in the military in Singapore. And they tell me, look, we do exercises with the naval exercises with the Americans and the Chinese.
And we also do land-based exercises in Taiwan. And so that's a country like many others. Vietnam, obviously, is one of those. Cambodia is another one of those. That's straddled a fence, but it is particularly notable. Multi-directional foreign policy.
Multi-directional, sometimes by necessity. I think whether India can hold that is a big question. It is a very powerful, large economy. And I think at some point it will not be able to do without US help. I don't think Russia will help much in terms of getting over that middle income gap into that next growth frontier.
And China isn't for obvious rivalry reasons. And Europe is nice to half, but it isn't going to be the driving force of economic development for India. And so there's really only one left that has a resilient economic engine that, frankly, Europe and others have been depending on over the past couple of years.
And so it needs to start looking toward the US more. And that's uncomfortable. With philosophically, which one do you embrace, open source or closed source? That's a great question, and also an uncomfortable one. Generally speaking, I'm in favor of open source as a matter of sparking innovation, as a matter of diffusion of capital, and you spoke of ideas earlier, diffusion of the creation of ideas. And I think it is always a good thing to have counterpoles to the large platforms. I will say, and maybe an additional point is resilience based on open source, because reliance on just a few global platforms, especially when it comes to AI and its cascading effects, is not enhancing resilience.
And so you want many, many actors. That's the initial idea behind the internet as well, as you know. It was conceived of at DARPA as a means to make its national security communication more resilient. And diffusion and decentralization was an essential part of that. And so from that angle, I am in favor of open source.
I will say that that combination of open source and diffusion while you have currently the dominance of the global platforms that provide models to the open source community, meta being the prominent one, of course, that mix is still very dicey at the moment, because we don't have truly open source that comes from the open source community in terms of where the inception is. The inception is really still very much in the platform, and that's the LAMA model. But when you have regulation that will tell meta to shut down LAMA when there is threat of massive catastrophic effects and everybody else downstream gets shut off as well, then, of course, you don't have resilience, right? So we need to figure out this balance, this tenuous balance between central and decentral. How does that work in the world of split-second AI model decisions? So that's one element that is critical. And the other one is, do we have regulators who understand how to regulate open source AI? It is, I think, Sam Altman was right when he said it's a lot easier to regulate four, five, six, seven large platforms than it is to regulate seven million open source models. And so we need to create new capacity in our regulators to deal with that.
I can buy that argument, but it's tougher for me to swallow the not-for -profit versus for-profit debate. I think shifting from nonprofit to for-profit on whatever platform we were talking about where Sam is on just seems troubling. I want to take you to a different level where there's this observation that Silicon Valley was this entrepreneurial experiment on the back of massive support from the government of the US, on the back of this thing called the Cold War, which the academic institutions in the Bay Area would have been very supportive of, and they helped actualize the aspiration. Now, in the absence of a Cold War, how do you think the academic institutions in the Bay Area are going to be able to maintain or help maintain the technological innovation that we would have seen in the last few decades, at the rate that China just seems to be taking everybody else's lunches, technologically speaking? Yeah, very hard. We've seen, of course, over the last couple of decades more and more focus on R&D by corporations rather than the government relatively in proportion.
And you're right to point out that Silicon Valley was very much founded on and still benefits from US government involvement, whether it's DARPA and the Defense Department or other actors. And I think when I sometimes hear my friends in Silicon Valley talk about the fact that Silicon Valley is so good at what it does because it's far away from Washington, I can't but chuckle because it defies the reality of today and history. You know, when I look at where CDMA came from, or early contracts.
It's the government, child. Yeah, it's the government, exactly. And still today, we wouldn't have had CRISPR without years of research at DARPA. And so I think leaving that aside, because it's just not credible, I think the US government has realized, arguably under the Biden administration and really even reaching back to the Obama administration with the BRAIN program, that it needs to get back in the game. Now, I think under the Biden administration, you have seen, of course, and previously under the Trump administration, you've seen, I think, a stronger perceived threat from China. And so looking at the Chinese model became more pronounced in the Trump and in the Biden administrations than it was under Obama, at least in my view.
And so I think that's why you've seen the US government do this 180 on industrial policy as well, and on fueling science and technology development so the US can stay ahead. And I think that is a shift that is here to stay. We can have the conversation amongst economists, whether that's a healthy thing. There are many critics of the industrial policy right now, and I see some bad elements of it. And we can talk about whether 100% tariff on Chinese vehicles is a smart thing.
But I think if you're asking me, as somebody who looks at futures and foresight, whether that's going to abate anytime soon, no, it won't. And I do think it will aid Silicon Valley. And I think that's where your question was headed. And it sort of harks back to the early days of Silicon Valley.
But there is definitely a head to head competition now with China that is not going to go away. Or maybe the decoupling between China and the US is going to be the new motivating force or element, right? It's the motivating. The next phase, yeah. But is that feasible? When you're looking at what we're doing with these semiconductors, it's one thing to focus on the most advanced chips so that we determine the technological playing field on a 15, 20-year horizon. It's another thing to expect that, to prevent US chips from being found in Chinese fighter jets.
A lot of the technology that you need for those kinds of weapons is not two or three generations out, it's actual commercially available technology today. So yes, I think the paranoia and the system competition that we see that's in our heads is driving that industrial policy and is driving those monies to flow, whether that's actually going to lead to a real decoupling today or tomorrow, I doubt. You've alluded to a number of times to the fact that as much as data is there, only 20% is structured and only 0.5% is analyzed. That was shocking to me.
I would have thought the percentages would have been much higher for each. And they may be higher today, we wrote the book two years ago, it may be higher today. But data is the key fuel of what we have called that cognitive economy, right? And remember, cognitive economy means command and control based on data across different pillars in society. So corporate operations is one pillar, infrastructure is another, ecological sustainability is another, our biology, our psychology, our two more pillars. And technology now, cognitive technology enables us to exercise command and control across these based on insights that we have across these. But none of that works without data.
And so I sometimes say we need to not have such an exclusive focus on the models and rather look at data as a topic. And so whether that is the fact that there is still too much noise in the data we create, or whether we look at bias in data, or the fact that data rots and degrades over time, and certainly how we treat data ownership, and that we're not really cutting people into business models enough. And when I say people, I mean the creators of data that are rightly saying, hey, you're doing many more things with my data than I paid for, or I thought I was paying for, or that you are paying me for, I should say, with your services, right? You're selling that data to second and third order parties, and I don't see any of that revenue. All of those questions are data centric questions, and they're heavy questions. And there is valid views, pro and con, on all of these questions.
But we need to have dialogues on data and not get so hung up on the shiny new thing, which is these humanoid AIs that we often depict. this in the context of the comparison between China and the rest of the world in their respective treatment to data. I mean, if you've alluded to the fact that China doesn't allow companies to transact or trade data, right? And you've also made reference to China being wobbly on the economy, the US being wobbly on the politics. And take us to how data is going to play a role. Yeah.
So the reason for why the US... One of the reasons, many reasons, why the US economy is so resilient, leaving aside the fact that the Fed has done a good job, is because of the free flow of information and the reasonably free flow of data, especially on the enterprise side, with some safeguards, but not too many. So the US leads on flow of data and corporate data, enterprise data. China leads on the flow of consumer data, consumer application data, based on its own legacy of social media. The difference between those two is not just the buckets of data, but it's also how both countries treat the trading of data. So in the United States, you can sell data.
So Facebook's collaboration with Cambridge Analytica, I'm told by a lawyer, friends of mine, was perfectly legal. It's not legal in China. You can share data within the same stovepiped organization, vertically integrated organization, or even horizontally integrated organization, as long as it's one corporation. So the fact that WeChat uses user data across different business areas within its own corporation, that's very legal, but it couldn't trade data with other hyperscaler platforms across the Chinese economy.
So that's illegal. It's not illegal in the United States. So domains are different, and the data sharing paradigms are different as well. But both generate enormous amounts of data.
Now, Asia overall, and certainly with China at the center, is leading on data growth. It generates more growth than any other region. But the US, because it started earlier, is still very much at the intersection in a hub for global data flows across borders.
Its platforms are far ahead of the Chinese platforms when it comes to unique users, because it started earlier and it rolled out internationally earlier. The Chinese model does not work for as many countries around the world, and it is not as established as the US model. So there is a time disadvantage. So the US still leads on global data. The question is, who is going to be first in defining what that future global data economy will look like? Will it be that US model of discretionary data trading across organizational boundaries and across international borders, which is very laissez faire? Or will it be a more controlled data trading paradigm that is the norm in China, where governments will determine whether that's allowable or not? Currently, we're seeing evidence that Asia is a little ahead when it comes to intergovernmental agreements on data trading and the digital economy at large.
For instance, through the DEPA, the Digital Economy Partnership Agreement, that Japan, I think, is part of the Philippines. Small places like Guam are part of this. I believe China wants to join. I think Singapore is in it. And there, I think, it's the first international sort of agglomeration of national government saying, yes, we want to trade data and digital goods, but we got to figure out how, what's acceptable.
That has not yet happened widely, as far as I know, amongst the existing Western regime, whether you call it OECD or whether you call it G7, Western-dominated regime anyway. And so that's where I see the future headed. What s
2024-12-14 06:59
