Ep 103: Jayant Sinha on AI Governance In A Fragmented World

Get ready to enter a brave new world with your host, Vesant Dar. Hello and welcome to Brave New World. My guest today is Jayant Sinah, who is a globally reputed policymaker, thought leader and investor who served as a minister of state for finance and civil aviation in the Modi government. He is currently a visiting professor at the London School of Economics and the president of Everstone Group, one of India's largest investment firms that focuses on new approaches for sustainable development. He's also the author of a recent book called India's Green Start-ups. Jayant, welcome to Brave New World, delighted to have you on the show. Wonderful to be on the show, Vesant. I have been listening to the show regularly for years now and it's such a privilege and honor to be with you today. Thank you, really look forward to this. So, I usually start by asking my guests to tell us something about themselves. You've had sort of an interesting background in journey, your former member of parliament, a minister of state for finance and minister of state for subliteration in the government of India. And now more recently, you've sort of pivoted into the business world as an investor and a policy analyst. So, take a few minutes and just tell us about your journey and what's brought you here. Well, Vesant, I have bounced around quite a bit. I was born and brought up in Charkant, which of course, at that time was Bihar. My father was in the IAS, so he was transferred around quite a bit. So, I was born in Giridi, lived in Hazaribar, Khudumkapatna, and then ended up in Delhi, after which my father was posted actually to Germany. And I was eight when we went to Germany, we were there for almost four years. I went to a German school, I learned German, I can still speak German till today. And then I came back to Bihar in Patna, and finally ended up in high school in Delhi, which is when I took my exams for the IIT, and I was in the last five year batch, and ended up in IIT right after my 11th. And I did my BTEC there, and then got a scholarship to come to the US, to the University of Pennsylvania, to do a degree in energy policy and management, and after that I have been in consulting, and then, of course, investing in public service. Yeah, I realized we have a common Alma Marta in IIT Delhi, you know, I mean, I, you know, I talk in my book about I have no idea how I got in, because there was this one entrance exam, and that's all that mattered, you know, and I still remember going to the IIT to take the exam. And the first one was in math, and that was a three-hour exam, and I came back after lunch and half the hall was empty, you know, you know, people didn't come back for physics, you know, or the other subjects. So anyway, I have no idea how I got in, but, no, you're exactly right, percent it was an intensely competitive exam, but it was obviously a life-defining experience for me to be there on campus for five years, which such super smart people, and of course, I was very fortunate to meet my wife at IIT also, so we are one of the very few IIT couples mainly, because they're so few IIT girls, there were just four people in my batch, and I married one of them. How about that? You know, by the way, I don't know too many politicians. Well, so I'm just curious, what's, what was that journey like? It's a fascinating experience, and I think all of us should spend time in public service for some, because those of us that lived in sort of this professional world have a lot to learn about, you know, how things actually work in public service. I was, you know, privileged in the sense that, you know, my father, of course, even though he had been in the IIT, so resigned, enjoyed politics, and then had a very long and successful career. As a politician, he was the finance minister of India, he was the foreign minister of India. And in 2014, when he decided he didn't want to fight the looks of our elections, because I had been working with him and helping him in his elections and campaigning and so on. The party said, well, you know, if you're not campaigning and fighting this time, Mr. Senna, why don't you ask Jayanth to run? And so that's how I got the opportunity. I fought the looks of our elections in 2014, one by one like 60,000, 160,000 votes, and then one again in 2019. So I spent 10 years in parliament, 10 years serving a constituency of two districts, roughly about 25 to 30, like two and a half to three million people. And it was an eye-opening experience to be in public service and dealing with the problems that people have to face on a day-to-day basis. I bet. I bet. And like you said, you know, most of us have so little knowledge about, you know, how the wheels and politics actually turned. You got quite a unique background here. So let's talk about innovation, which we've talked about casually, and, you know, get your take now that you're sort of in the startups and VC and investing world. What do you see as the sort of big spaces of innovation in the world? You know, AI is obviously like one of them and that's you know, super hyped up. But what are the sort of big areas, the big spaces in innovation globally? I mean, there are at least four or five areas, which are absolutely in takeoff mode right now, that one of them, of course, you've already highlighted AI. And AI like electricity and the internet is going to be all pervasive. It's going to spread across many different aspects of the economy, pretty much touch everything. So artificial intelligence, of course. At the same time, what's happening is that there are very fascinating developments in the area of energy. So apart from solar power, which is of course now more cost effective than any other form of electricity right now. We have battery storage, which is changing the way in which power is actually delivered to you. So battery and battery storage is going to be an incredible area, which is going to transform at a minimum transportation, but many other industries as well. Then of course, we have very exciting developments in nuclear fission and of course, nuclear fusion. So that obviously is going to be profoundly important in terms of delivering base load energy. A whole area of energy is switching away from fossil fuels, switching to these other sources of energies and major major area of innovation as well as investment. Then we have biosciences. And here again, you know, incredible things are happening. Obviously, you know, in some ways stimulated and driven by AI like alpha-fold as you know, as soon as which is changing the way people think about proteins and developing proteins. So all across biosciences, we're going to see incredible developments happening. Space of course is always the final frontier and here what's happening with SpaceX, the trip to the moon with the Artemis program. What China is trying to do, what India has been doing in space satellites up again, incredible quantum computing, another major frontier area where lots of exciting things are happening. So you add all these biosciences quantum space, energy, artificial intelligence and the kind of explosion that we are seeing in technology and innovation is absolutely mind-boggling ascent. And I'll see one more thing. You know, we talked about kind of my childhood and, you know, what were the things that that were of interest to me during my life's journey. And I will say one thread that connects pretty much everything I've done in my life. And that is the transformational power of technology. As a boy, you know, I was a real science junkie. I loved science. I used to consume science fiction and science books all the time. I was incredibly lucky to go to IIT and study engineering and be with a lot of super smart technologists and engineers and get to really understand what engineering is about. And then the work I've done subsequently, both as an investor and as the policy maker connected with technology really has been incredibly fulfilling for me. And now, after everything we've already seen in technology over the last 50 years, to be at the brink of what I think is going to be just an extraordinary explosion of new advanced and, you know, really, really important technologies. It's super, super exciting. Let's start with the first of these AI. How do you see AI working out, you know, in different parts of the world? Do you see sort of different parts of the world having different AI strategies similar to the way the internet was used differently in different parts of the world? Now, that's a very, very interesting observation, Vasant, and you're absolutely right. So, different societies, different cultures adopt and use technology in different ways. The US is completely wide open. It's largely in the hands of large companies. The government has a very light touch and it's a free for all in a way. And that results in incredible innovation. At the same time, you can also argue that it creates a lot of impact on society, which is difficult to regulate and control. Europe, on the other hand, is a more heavily regulated environment for the internet, for any other technology. You've seen that in AI also. And as a result of that, there's less innovation there, but maybe society is able to deal with it better. China, of course, as we know, is the great firewall where they have built kind of their own homegrown approach towards all these different technologies. You're allowed to do certain things, not allowed to do other things. There's huge surveillance, there's no real sense of privacy and so on. So the China model of these types of technologies is very different from either the European or the US version. The exciting thing was that in my lifetime, you know, led by visionaries like Nanda Nellikeny, we have been able to develop a different approach to technology from these three other models. Certainly we've done that in the internet and I hope very much that we'll be able to do it in the AI age as well. And let me characterize the Indian approach with three very important aspects. Number one, it's incredibly affordable and accessible, right? So today the internet is largely delivered through the mobile device in India. And we have more data being consumed on the mobile internet than in any other country. And it's at a rate that is inconceivable even in China in terms of the cost per megabits or whatever, however you want to look at it. So number one, it's very affordable and accessible. I could be in the remotest village in Charkin and I'd have 4G coverage and you know, be able to access any service I want in the world, which is incredible. That's how accessible it is. Point one, point two, we have put together a set of standards and protocols, which is the famous India stack that has enabled us to develop our own approach to digital public infrastructure. So between the identity layer at Adhaar, which is totally unique, of course, UPI, the payments system, ONDC, which is an e-commerce layer that we are creating. We have built our own protocols and these are now being looked at very carefully by many other countries in the global south so that they can adopt something similar. And then finally, that there is a genuine public-private partnership in India so that government is playing its role and the private sector is playing its role so that we have innovation and we built billion dollar unicorns, of course, unlike Europe where there's less innovation. So we've got significant innovation as well, even though we have a well-defined digital public infrastructure. So those three aspects of the Indian model, I think, make it genuinely quite unique and remarkable. I was going to ask you, since many of my listeners, in fact, most of my listeners are not Indian, you talked about digital public infrastructure, which is a phrase that is now quite common in India, but I'm not sure that that's so common in other parts of the world. And I did have Nandan Nilakani as an early guest and promote Varma, who's sort of a major architect of this approach. But let's go back to why did it develop this way in India? Why did India come up with this notion of a digital public infrastructure? What's the genesis of it and the thinking behind it? So most of the credit for that was, and truly does go to Nandan and the team at I-Spirit, which is a nonprofit that has been at the forefront of defining India's digital public infrastructure, led by people like Pramod, whom you've interviewed, Sharat Sharma, who's another key architect at I-Spirit. So it all started with Adhaar, which is, of course, a digital identity system. And the architecture of it is incredibly elegant and truly, truly outstanding. Of course, great credit to Nandan and Pramod for putting it together. But the notion that each of us can have a unique identifier, but it's only an identifier. It is not a certification of anything else, but it is a way of certifying your true identity. And rolling the entire population of India within a decade through biometrics, which is fingerprint as well as facial recognition. And then enabling it through a set of redundant data centers so that you can do a very fast look up and say, if these are the markers, whether they're just your name and your father's name and so on, just those kinds of markers or your face or your fingerprints. And with that, you can do a quick look up and say, is this wassant or not wassant? That ability to do KYC digitally at a very, very low cost was truly transformation. And once that was done, and this was, as I said, the architecture that Nandan and his team developed, once that happened, we all realized that we were onto something very powerful that once digital KYC could be done, a whole host of other services could be built on top of that. And then we built a system which was an orthogonal to that, but also supremely complementary, which was the unified payments interface. Now, so also hold on for that for a second, because I want to dig a little bit deeper into other, before we get into payments. So I know that there was some skepticism at the time that this was proposed that this would be like a big brother kind of a thing, right? That it would be used for surveillance or used for control, that it would lead to government overreach. And I remember Nandan at the time saying, there's only one thing this thing does. And in fact, he said there was a sort of temptation, you know, on many, from many parts of the government to say, well, what do we include all of these other things in it? And he said, no, this thing is supposed to do one thing and one thing only, which is, are you who you say you are? Period, right? And that's it. And in retrospect, that was also brilliant to actually limit the scope of this to just pure authentication and nothing else. How did that whole thing play out? The skepticism, is there still skepticism around the system? And that it could be, you know, subject to government overreach? Like, well, how people think about it now 15 years later? So necessity is the mother of invention, as you know, as a, and because Nandan knew that if he started to step on to some of the most sensitive areas, like citizenship, you know, like surveillance that he would get enormous pushback, he designed it as a very lightweight, elegant architecture, which exactly as you say, just allows you to do authentication. And it doesn't keep any record of who was authenticated when they were authenticated, how they were authenticated, they just send you back a yes or a no. That's it. And then that whole transaction is perished from the system, no trace of it remains. So it's just doing one simple function, which is an authentication when certain markers or characteristics are presented to you. And that's just a thought. Now, over time, and in fact, I have been part of a lot of these sort of non-making efforts around it, we had to actually get it passed through Parliament, we had to get it verified through the Supreme Court of India and so on, because lots of people correctly said that they should not become a precondition for, you know, universal welfare programs. So if you didn't have an Adha number, you could not be eliminated from a welfare program or you could be denied any of your citizenship privileges. So it had to be designed in such a way that, you know, it was not mandatory, it was voluntary. But, you know, as it has become more and more pervasive, we've been able to enroll more and more people. But again, as I said, the law making and the policies around it had to be designed in such a way that it could be near universal without it being mandatory. And again, as I said, necessity is the mother of invention. And so we innovated around the architecture in such a way that it could be, you know, virtually universal without it being mandatory, which is a very difficult tension to pull off. But I think Adha has done that. And what is the attitude towards it now in India? I mean, do you still have skeptics and naysayers and people worried about overreach? Yes. And there should be because obviously, as the Supreme Court declared in a very important judgment that privacy is a fundamental right in India. And so you can't, you know, sort of interfere or violate people's privacy. So we have to continue to ensure that we know we are supremely compliant when it comes to privacy and data security and so on. While at the same time providing people the convenience that Adha can and does deliver. So yes, we are walking that fine line. And by doing so, I think, as I said, we've developed out of necessity and architecture that is very extensible and very, very popular around the world. And so there are many countries now that are taking the essence of this architecture and implementing it themselves. So now let's move on to sort of the next level and I interrupted you. I'm sorry about that because you were getting into payments, which was kind of the next layer. Yeah. So on payments, I think again, two or three things were sent out very well worth noting. Number one, it's free. So unlike using Mastercard or American Express or whatever in the US, this is absolutely free. And so even if you're buying fruits and vegetables in the Bazaar, you know, very well, you're doing it with UPI. So I of course carry no cash in Ilhan Gub because there's no need to everybody, the Pan Ki Dukan, obviously uses UPI as well. So UPI is free and therefore it's become supremely accessible. Number one, number two, it links virtually every bank account and mobile number to another one. So the moment I know your mobile number or you have a QR code, you can show me, I can pay you. So all bank accounts are pretty much interlinked as a result of that. And again, this is back to architectural design. There is a payments clearinghouse with a very big switch where mobile numbers and bank accounts can be interlinked across banks and it works here. seamlessly and transparently in milliseconds across every different bank through this massive switch that we have, which is this national payments clearing out. So that's pretty remarkable that we built a section eight nonprofit company called NPCI National Payments Clearing Institute. And that does all of this back office clearing of transactions. So that's quite remarkable. And number three, of course, because there is a bank of India forced all the banks, all the banks now are part of this. And so anybody can pay anyone if you have a bank account. So these are remarkable things that this payment system is done. And again, I'll repeat, it's free. And now Singapore, Dubai, a few other countries are also adopting the same architecture. And so that's made it incredibly, incredibly universal as well. I was going to ask you, it seems like other countries should be adopting it. If it's safe, it's free, low cost, doesn't it wait privacy? Yeah, exactly. And it's called built in security, right? Because you have the clearinghouse in the middle. So it's not if you're getting access to my bank account, you're just getting access to an identifier in the clearinghouse and there you can be stopped, etc, etc. So the various ways in which security is built into the architecture. And again, for the global south, this kind of a lightweight payment system, which is free and accessible is quite attractive. And should be very attractive. Now, of course, in Africa, we've had MPSA and other sort of universal payment systems as well. And in fact, MPSA was a very important inspiration for the design of UPI as well in India. You know, what's actually boggled my mind is that Europe is considering adopting it, you know, of all places. You know, I remember this lunch I had many years ago with Nandan and with Margaret Westhag. I don't know what her title was at the time. But you know, and you know, the world was just beginning to get familiar with, you know, Adhar and the notion of digital public infrastructure. And I was actually amazed that she expressed a strong interest in it, you know, as something that could work in Europe, you know, where the notion of a national identity seems, you know, something that wouldn't be embraced. No, that's right, was it? But again, let me clarify that Adhar is not a national identity. It's an authentication architecture that is available for people and it's voluntary. It's not mandatory. But it is so convenient and accessible that pretty much everybody wants to have it. That's why I'm saying we've kind of straddled that fine line between being voluntary but still universal, right? So you have to opt into it and all of that, of course. So this notion of opting in and consent has, I guess, extended since the, you know, early days of Adhar and, you know, you've been writing about it in some of your sort of policy kinds of articles, where you say, I don't know what I'm butchering what you actually said, the way you can clarify it, that, you know, that AI would work best with, you know, with some sort of a consent-based architecture behind it. Am I getting it right? Well, partially, wasn't so the way AI is evolving right now brings up many very interesting questions around sovereignty, nommaking, human rights, and of course, you know, the technology itself, right? So the regimes that are starting to develop and evolve are going to have to struggle with these four five very, very important issues. India, of course, is struggling with it. I think Europe is struggling with it. The US seems to have adopted a certain model. Again, a model where large companies are pretty much in charge and, you know, people are moving forward without worrying too much about regulation and just it's free for all in that sense. China, of course, is a very close model as we know. Europe is, you know, a very regulated model. They are trying to control AI in many different ways. But countries like India have a particular dilemma. And the particular dilemma is that compute, as you know, is enormously expensive. So where the models are running and how the models are running is something that you will not be able to control if everything is running in a hyperscaler, in Utah, Virginia, or Oregon. And if 10 or 20% of India's compute needs are domiciled in India and the remaining is outside of India, it could be in Malaysia or it could be in Utah, Virginia, wherever the big model providers place it to be. So that's kind of what's evolving right now. And as it's evolving that way, for India, which has already well developed DPI, as we've discussed with Adhaar, with UPI, with ONDC for e-commerce, with privacy protection and so on, we need to evolve our own architecture for AI as well. And that's really what I have been writing about also because obviously in my policymaking life, I was quite deeply involved in the way DPI evolved in India and I myself designed and implemented a very important aspect of it, which is Dijayatra, which is a national digital traveler program. So I've had a lot of experience with UPI in India as it's evolved and of course Dijayatra and Sina Adhaar and all of that. Now seeing that and now seeing what's happening with AI, it's very clear to me that we need to put some guardrails in place for AI as well. And that's what I have been writing about so far. So say more about that. Like what kinds of guardrails? Because I think you say something to the effect that these guardrails should be sort of built into the architecture itself. So what do you mean by that? I found that really intriguing. So just like we did with DPI where for example, architecture in Adhaar, the only thing that the system allows you to do is authentication. Or if you look at UPI where we have a payments clearinghouse in the middle acting as a switch, which is a utility, a free utility that allows you to do that paid for by the banks. If we architecturally get it right then solving the problems that I was talking about earlier becomes much easier. So what are the three for problems we have to solve and how can we take an architectural approach to it? That was really what I was thinking about when I wrote those opets. Three of the problems are pretty obvious when you look at AI as it is right now. Number one is I was talking about earlier. The compute is all over the map. So the compute can be anywhere. It need not be in your country and then your data and your model is not localized. So what do you do about compute number one? Number two, who owns the data right? Does the individual still own the data but as you're interacting with AI or AI is developing data about you? Do you have any control over your own data? That control is not fine-grained. How does it get shared? So data control becomes a second very important thing. And then related to that obviously is this question of privacy right which is does everybody know exactly what you're doing? How you're doing it? And how do you prevent that from getting out of hand and out of control? So these are the three big problems you know sovereignty of a compute consent and the control of data and number three privacy protection as you're seeing for example in Australia where kids under 16 can't be on social media. Maybe kids under 16 should not be on AI either for that matter right? So how does that get controlled when I talk about privacy? So these are the three big problems right? Now how do we architect and put in place guardrails so that the big providers you know the make you know the obviously the Microsoft's and the Googles and the Metas and Open AI's at the world operate within those guardrails enabling societies to have sufficient control over these three very big problems and that's really the architecture I've talked about. So number one is I've said is you have to have sovereignty over compute which means that whoever and however the model gets set up it has to be partition properly so that the models that are operating for India are different from the models that are operating for the UK or Nigeria or the US that partitioning is very important the cross contamination of data and so on should not happen. Now when I've spoken to the model providers they have told me that will break their AI models that this is going to be a really big problem for them because obviously the weights are all over universal they are global etc you can't have a set of weights for India you can't have a set of weights for Nigeria but again my belief is necessity is the mother of invention so is there a way in which the weights for each country and the models for each country can be controlled by that country I think that's a technological architectural problem that we have to solve. So that's that's the first area. Second of course when it comes to data we need to have consent people need to be able to do an opt-in there and there we have a full architecture called the DEPA architecture data and our empowerment and authorization act where we've designed a method by which you can through a data broker give consent for certain types of data so that consent architecture also I think needs to be part of this and then number three as far as privacy is concerned you have to be able to see your own profile manage your own profile etc again that needs to get architected so that the equivalent of the cookie in the AIH is something that's transparent to you. So let's come back to sort of the big spaces of innovation and what you see going on and you know what you see sort of as the table stakes as being right who's really in a position to make progress in these big spaces. So Vasant it's really the golden role which is he who has the gold rules he who has the gold he who has the money in of it and that's essentially what's happening right now we talked about these big new innovation areas obviously AI quantum computing, new sources of energy space, biosciences extraordinarily important profoundly transformational but At the table stakes, the Senth are trillions of dollars. The table stakes are trillions of dollars. Right now in the US, just as far as AI hyperscale is a concern, there's 500 billion dollars a year of investment happening. China is investing people say somewhere in the range of 100 billion dollars a year. So for India to be able to be competitive to have models that are built and trained in India or to develop AI applications, etc. DB have to also spend hundreds of billions of dollars a year at a minimum tens of billions of dollars a year. Are you talking AI here? Yeah, just AI. Just AI. Then let's take small modular reactors or let's take fusion. I mean, there's five, 10 billion dollars a year in the US going into fusion every year. So if you have to keep up on fusion, that's the kind of investment we have to do in fusion. China is investing at that rate. I mean, their fusion advances are quite spectacular also. And of course, in quantum computing again, billions and billions of dollars a year. So innovation is happening in these areas. They're all incredibly important and they're all interconnected. I mean, there's discussion Elon Musk and others have talked about about putting hyperscalers in space using space technology so that you can take advantage of the absolute zero and the cold of space to run these very heat producing data centers. So all these are interrelated. I mean, if you unlock nuclear fusion, you can build bigger and bigger scale data centers, which you will not be able to do if you don't have access to that kind of power. Similarly, if you have cheap solar power, you can build these data centers. They're all interlinked percent. And unfortunately, the table stakes to push the frontiers of these technologies to push the frontiers of research is literally going to be trillions of dollars for India and Europe to catch up with the kind of investment that's happening by the two superpowers right now. Looks very, very difficult. And so then the question becomes, how do you stay at the cutting edge and maintain your strategic autonomy, maintain your ability to control your destiny going forward and avoid becoming completely dependent on the US and China. And let me just very quickly sort of add to that because right now, if you look at AI, all of India's AI usage and India is the second largest user of AI services right now, is happening out of US data centers. Similarly, every solar panel of battery coming to India is full of Chinese elements. So on the one hand, when you look at energy, we are very dependent on China. When you look at AI, we are very dependent on the US. How will that play out over the next 10 or 20 years as you become increasingly further and further behind with this kind of investment happening is a question that really worries me. You're making it sound like it's almost like game over in terms of being able to invest in these technologies, right? But those kinds of table stakes. What this really reminds me of is I gave a talk a couple of weeks ago at Standard and Poor's and the chief economist Paul Drunwald asked me, he said, you know, you talk about this sort of impending bifurcation of humanity due to AI, do you see this applying at the level of companies and even countries, you know, which I thought was a really intriguing kind of question. And the way you're making it sound is that it's it almost seems like it's like game over in terms of like this fundamental innovation in these big spaces for most countries except the US and China. So the sun we've seen it happen in history so many times, right? So just take the industrial revolution when the industrial revolution began the two largest economies in the world were India and China in that order. They were both 25 30% of global GDP. Then the productivity that you got out of the industrial revolution resulted in a small country off the coast of Europe that is of course the United Kingdom becoming, you know, this global superpower because they had access to this technology coming out of steam power and you know better ships and better cannons and everything else. So yes, technology does result in a fork in terms of comprehensive national strength. And when that happens, then certain countries become very, very dominant. Similarly, the same thing happened with the US post world war two with the control of, you know, so many key technologies, whether it was space, whether it was nuclear, whether it was computing with control of these incredibly important powerful technologies. The US rapidly became, you know, the superpower which post Berlin wall became the sole superpower. Right. So yes, technology does result in geopolitical sort of changes as well. And with AI now, it looks like, you know, the same thing could very well happen. And India having been left behind twice or twice in this technological race, I think really we need to worry about what we can do to stay in the race. And a lot of my thinking and policy suggestions are really designed around that. I mean, how does a country like India or, or, you know, you know, how does Europe, I mean, how do different parts of the world stay in the game, you know, as opposed to becoming completely marginalized because they're just not participating in the table stakes. So the way you do that is by unleashing innovation in an open market economy. For a democratic country like India or for Europe for that matter, there is only one way. And in fact, the Nobel Prize has just been given to a set of economies, Dan and Achemau Glue, James Robinson, Joel Mocha, others that have spent their lifetime studying how innovation happens, how productivity growth happens, and how society is progress. Now of course, the exception to that rule that they've identified is in fact China. So we still don't understand China very well. I think one aspect of the Chinese model, which is very interesting is that capital largely is abundantly and freely available there. And obviously they have developed an incredible, incredible strength in engineering talent and scientific talent because of the longstanding sort of strength in science and engineering and so on. So we have two very clear examples, you know, lots of economists have won Nobel prizes for showing that innovation in a market-driven economy is really the way for you to flourish and prosper. There is the China model, which nobody can quite figure out, but it looks like it's being very successful right now. How long that will endure is still a bit of a question mark, but certainly they've shown over the last 20 or 30 years that they can now go ahead to head with the US, which is kind of incredible in a way for those of us who know the US innovation system well. But yes, these are the two different models. Now for India, it's very clear that we can't follow the China model. We have to follow the model that open market economies have followed and we have to therefore really, really emphasize innovation. And most importantly, we have to put enough capital to work and remove any restrictions or problems that come in the way of innovation. And in a way, Vasant, that brings us back full circle to where we started the conversation, which is really about education and the IITs. So if you look at, you know, what is remarkable about the US, both UNI, of course, have benefited from, you know, studying and working in the US and being associated with great universities. For the IITs, it's extraordinarily important now to be at the forefront, to be the vanguard of innovation and technology development in India. And it's really the older five IITs actually that are going to have to take up this mantle. But unless we can make, let's say IIT Delhi, the equal of Qingua or MIT, it's going to be very difficult for India to compete. And it all starts there, just like Stanford and MIT and Qingua have done it. It's IIT Delhi, IIT Bombay, IIT Madras and others have to do it as well. Anything is possible? Well, we have to again start with the money. So let me explain. MIT's budget is $3 billion, Vasant. $3 billion here. Qingua's university's budget is $1 billion here. IIT Delhi barely gets $100 million here. So we've got to put a tremendous amount of money into our top institutes. Then of course, we've got to strengthen the VC ecosystem, which is okay, but we need to do much more there. Corporate innovation has to really take off. Defense procurement, the government to DARPA and others have to become a lot more innovative and agile. And thankfully, our capital markets are doing quite well. And our capital markets then have to continue to fund this also. So the whole ecosystem in India has to really take off, ignite all at the same time. But unless all these jet engines fire, it's going to be difficult for us to find the initial. The US has a unique university ecosystem developed over centuries with dozens of top-notch institutions that compete and cooperate at the same time. Which is why I asked when you said, "Been you do asset, is it possible to do this in India?" And I guess you're optimistic. Well, it's very, very hard, but we have no choice. A necessity again being the mother of invention. So we have to find a way of doing it. And it's quite subtle in a way also of a sense. So some years ago, we did a benchmarking study for IIT Delhi in fact to see what is it that IIT could do to compete with an MIT. And all the obvious stuff came up, obviously. the money, you know, the independence and autonomy for the faculty, enough research grants from the government. All those factors came up for sure, but at the same time, another important item came up and that was livability. Most of the universities actually are in very nice areas to live in, whether it's where you are in New York, in Boston or in Cambridge, UK. These are very nice livable cities and people want to live there. And that's part of the reason why these universities have flourished. Yeah, you were talking about livability, Jent, and that is one of the huge drawbacks of Indian cities, is that they've just become completely unlivable. You have the air quality index and Delhi approaching a thousand, it's just like off the charts, right? You're absolutely right. And for us to be an innovation-driven economy, we are going to have to make sure that we keep the best and the brightest and ensure that, you know, our talented people can make good high quality lives in Indian cities. And that was in fact, the case in Bangalore. So Bangalore was a very nice livable city. Traffic was not a problem. The weather was very nice and that's part of the reason why Bangalore became the high tech capital of India. Of course, that's not the case any longer in Bangalore because of the traffic and so on. But as long as we can get those livability characteristics right, our cities will flourish. And certainly, you know, we need to do that for Bangalore, for Hyderabad, for Pune, for Delhi, and so on. But unless we can improve on the livability quotient, our best people will end up in Palo Alto, or in Cambridge, Mass, or, you know, in Dubai, or in the UK, because all of these are very pleasant and nice places to live in. So livability is a key element in the overall competitiveness framework. But it seems to have been like a really long-term slide, isn't it? It's a real crisis now and there is research now that shows that as a public health emergency, it's worse than COVID. More people are actually finding it difficult to survive than, you know, was the case during COVID. So we really do need to do something about livability in Indian cities. But again, as I said, once we reach a crisis, you know, action does happen. The government does take the right steps and an appropriate policy intervention is put together. So I'm hopeful that it'll happen. There's no silver bullet. It's complicated. There's so many different sources of pollution. They all need to be dealt with in a disciplined, dedicated way over decades to get it right. But Beijing has done it. London has done it. And I think we will also be able to do it. But yes, we have to tackle it on a war footing like we tackle COVID as well. Yeah, yeah, for sure. Because the way things are going, I mean, I can't imagine anyone wanting to live in any of the Indian urban centers where people would just flee. So Jaya, they want to come back to where we started, which is how you see this sort of global innovation race playing out in these big spaces that we identified at the beginning of our conversation. Nothing could be more important for India, but then this innovation race that we are on. And it reminds me of my time at the Paris Olympics last year, when we were in the stadium watching Nidh Chopra, Vinah Silver Medal in Javelin. And we were super excited. You know, we kind of had a photograph taken with him. We took the Indian flag and you know, it took photographs of that sent it all over social media handles and everything. Because it's so exciting to see an Indian Vinah Silver Medal and repeat a medal, you know, from the last Olympics as well. The next day, we ended up going to the diving stadium. There were eight gold medals on offer, China won all eight, and the entire stadium was full of Chinese flags. That's the innovation race. Whether it's the Olympics, whether it's different technologies, whether it is manufacturing, you know, we are up against formidable sort of opponents competitors. And we really have to ensure that our systems, our institutions, our processes are truly world class so that we can keep winning gold medals. And that's the race that we are in right now as we discussed earlier. Table stakes are trillions of dollars. So we've got to find a way of putting more capital to work, keeping our talent in the country and finding the technological areas where we can genuinely compete against a US or a China, you know, Europe, Japan for that matter. Otherwise, the next phase of technical colonialism will happen. As I was mentioning earlier, it's technology that determines winners and losers. And so unless we innovate, unless we develop the new technologies of the future, it's going to be very difficult for us to maintain control over our destiny and our strategic autonomy. You know, it's interesting that you're talking about sort of the India-China contrast. You know, I was at a mortgage-sendly event last week where I was talking about my book. And after the event, one of the people who has actually spent a lot of time in Asia and has studied China and India extensively came to me and we were chatting and she said, you know, it's interesting. I went from India to China and I could see immediately what all the problems were with India. And someone interrupted us at that point unfortunately and I never got to hear the answer, but when I do hear the answer, I will share that with you because I was most intrigued by what she could have met by that, you know. No, I'd love to learn. I'd love to learn. And really, I sent the question, isn't so much about one area or another area. It's not about just competing in the javelin or in diving. It's about how you consistently win tens of gold medals in the Olympics, you know, year after year after year. How do we build the systems to do that? Another great example is high-speed rail. So we are building a bullet train track from Mumbai to Ahmedabad and you know, that work has been ongoing. It will still take another few years to get that completed. During that time, China has built tens of thousands. Some people claim some 30 to 40,000 kilometers of high-speed rail. What are their systems? How are they doing it? Where are they getting the capital from? Where are they getting the resources from? These are the things that we need to learn and we need to be able to, you know, benchmark ourselves against. That's the competition and that is what it means to be in control of our destiny. Yeah, and that is quite a formidable challenge. I mean, if you look at the number of research articles published by, you know, Chinese universities, you know, they're comparable to, to, you know, the kind of volume of publication you see, you know, coming out of the US, it takes a while to get to that level and you have to do something right. And so India certainly has its challenges cut out for it. I want to also come back to, you know, India's green economy and what's going on there? What makes you optimistic? It's a great segue, Vasant, because, you know, we were talking about what China has accomplished. And what China has accomplished is because of companies like DeepSeek and BYD and CATL, so it's the really innovative companies of China that are really showing, you know, how they can compete well and how they can win globally. It's the same thing in India. If we can find and scale the right companies for whom sustainability is their business model, then, you know, we can demonstrate that we can also innovate at scale. And by the way, we've done that, Vasant, the good news in India is that we have many good examples of companies that have innovated at scale. We talked earlier about a remote villager, you know, in Jharkhand being able to access all of the internet services at a very affordable price. That's been done by telecom companies like GEO and AITEL. So we have scaled up that internet connectivity across India and done a super job of providing those services in an affordable way. Similarly, Indigo, despite, you know, the problems we've recently had have been able to at the most affordable price be able to connect Indians through an aviation system. Our banking and financial services is among the most efficient and cost effective in the world. UPI, as I said, is free. So we have been able to innovate at scale. We have been able to compete, but it's been through our companies. And so through my book, India's Green Startups, my goal was to be able to illustrate and explain how certain companies are innovating right now. And as they scale up sustainability, being their business model, we can definitely, definitely find a way to get ahead of many other global competitors in the innovation race. It sounds like what you're really describing in India is sort of innovation in terms of services and access to technology. Whereas the kind of innovation in China seems to be at a fundamental level, right, in terms of like battery technology or electric vehicles or drones or stuff like that. So they've actually become sort of global leaders in these spaces, right? That they've actually taken the evolution. That is an evolution. I mean, 15 or 20 years ago, China was also obviously known for doing very low cost efficient manufacturing. Now we obviously done that in services. But over the last 20 years or so, China has climbed up the technology stack and really pushed the boundaries. on lots of frontier technologies. And we need to do the same, we've reached the same point in our evolution that China had 10 or 20 years ago, we're at $5 trillion of GDP right now. And we need to now build the innovation ecosystem so that we can genuinely be competitive and genuinely grow our economy, the way China has been able to. And here, Vassan, the framework that I always like to highlight is the economic development framework that I have been promoting for a while, which I call the green frontier framework. What India needs to do is to really adopt a green frontier development model. And I contrast that frontier development model from the traditional conventional view that economists have about development, which is farm to factory. Farm to factory. So what you have to do is you have to take, you know, people who are working in the farms with low productivity agriculture, bring them into more high productivity, economic occupations and factories, you know, building shoes and toys and computers and through industrialization and urbanization, grow the economy. So that's been the traditional farm to factory development model. I think that model will not work any longer, particularly now that China has this incredible manufacturing capacity. And the world is awash with that capacity now. What we need is what I call the farm to frontier development model, the farm to frontier development model, where we get to the competitiveness frontier, the innovation frontier, the technology frontier, but we do it in a way that's sustainable. And that's why I call it the green frontier. And so when we think about economic development in India, we have to think about combining competitiveness on the one hand and sustainability on the other hand. Because as we were talking about earlier, if our cities are not clean and beautiful and livable, we are not going to be competitive. The two go hand in hand, the two sides of the same coin. In Hindi, what I say all the time, Vasantipil, excuse me, is, "Hame viksit bharitvi banana hai" or "harit bharitvi banana", which means we have to have a developed India and a green India together. The two are two sides of the same coin. So competitiveness and sustainability is absolutely crucial for us. That's what we need at this stage. And that's what I mean by being at the green frontier. So when you put on your investor hat, are you most excited about? So when I put on my investor hat, what I do see is incredible innovation in India happening, not yet at the deep tech level, though that's also beginning. We are certainly seeing signs of a lot of deep tech research happening in quantum, in space, certainly AI applications as well. So we do see certain deep tech applications starting to happen in India. The good news in India, of course, is that the VC ecosystem, which I was one of the people who helped build it up, as really matured, we have capital available. There's domestic capital now, lots of domestic capital available as well. And exits are available through our booming capital markets. So those are very positive signs. In a way, the fact that the rest of the world is being more restrictive on immigration will actually help in their innovation, because many of our best and brightest that may have ended up at NYU or at Stanford, may very well stick around in IIT and develop new unicorns. And many people who cannot get a green card or whatever may come back to India and start great companies. Many of our great companies, in fact, were started from people from the diaspora who came back. So I'm hopeful that we'll be able to put more capital to work that our corporates will step up and that much of our talent will stay in India and have built in their technological and innovation driven companies. What companies or type of companies are you most excited by? That address a large problem space that have a clear environmental impact, scalable business models, and are aligned with Indian policy thinking and momentum. What excites you the most? Many of them have been profiled in my book, "India's Green Startups," where we're trying to illustrate that if these companies can scale up to the population scale of India 1.5 billion people, it can have transformational impact. Let me give you four very quick examples. One is battery smart, which is a battery swapping network for three-veilers, enables three-veilers to switch from petrol to lithium-ion batteries. And it is much, much more affordable for them to be running on electricity than to be running on petrol. So battery smart is doing a super job. Second example is new tree fresh, which is hydro-ponic farming, climate-controlled agriculture, reducing the usage of fertilized water, energy, and producing vegetables at a far lower cost. All that is being done through greenhouses and so on. Very interesting company scaling very quickly. Number three, zero cow, which is developing milk proteins in a reactor as opposed to through cattle, obviously. This again is much better for the climate and is much more affordable for people because many Indians are vegetarians, they need milk proteins. And then number four example, I'll give you, is a company called Rebthin, which provides financing for two-veilers, three-veilers, four-veilers that are electric. And they've developed proprietary algorithms to be able to do the credit scoring, to figure out residual value of electric vehicles, longer-tenor loans, et cetera. All the things that are necessary to make electric vehicles a lot more affordable on a total cost-amonorship basis. For examples of companies, if they scale up like an Indigo-esque scale up or like a geo-esque scale up, it can be truly transformational for India and have a profound impact on our ability to deploy and use sustainable technologies. And as I said, net zero is net positive for India. So net zero is net positive for India. They're all picking India on the net zero trajectory. All right, well, let's see whether you make out like a bandit on those investments. By the way, zero cow reminded me of a company out in Emoryville, California called Upside Foods that is making cellular meat without slaughtering any animals. And Uma Villetti, the CEO of that company was a part-gast guest last year. They're now actually at that point where they're trying to scale the product. So they produce a product, but a taco will cost you $20 at a current scale. He did promise me the next time he's in New York, he'll bring some of his product along and will make a Rogen Josh, but I'm still awaiting that. But during our conversation, we used whether Indians would actually eat cellular meat, cellular beef. There's no cow slaughter. Would people eat that interesting thing to muse about? No, no, that is a real moral dilemma. And I think you need to ask a moral philosopher like Deepak Chopra whom you had as your guest recently to give you an answer on that. Because of course, how do people think that true? I'm not sure I know the answer to that. So, Jan, before we wrap up, I offer ask my guests about their advice to young people. I get this question all the time from students as to how should I think about my future. And I have my standard answers. How do you think about it? Do people ask you this kind of question? And Vissant, I get that question a lot as well. Because people say you kind of stumbled around in your life doing different things. What do you think people should do? And I think your terrific new book, Vissant, provides a great answer, which is that you really, really have to lead and think and get to understand the world and build your own world model that can really push the machines. So we were talking earlier about the innovation race between let's say the US and China. Actually, humanity is in a race too against the machines. And as you have said in your book also, we might have a situation where there's a bifurcation of humanity. There'll be people who use AI as a cognitive partner, as some set of tools that really amplify their potential and what they can do. And then are able to do amazing, amazing things. And the rest of us that are not able to put AI to that use, given that AI models now have all the answers. It's very important for young people to be able to ask the right questions, to push the machines, to push the models to the edges. But to be able to ask the right questions, you have to know an immense amount, which means you have to read and read and read. And perhaps also listen to great podcasts like Brave New World, because that's the only way you'll build your world models that allow you to push the machines and use them in a way that is truly value-ed. Otherwise, if the machine is adding all the value, then your contribution to what's happening in the world goes down. That doesn't necessarily mean that you won't have a happy prosperous life. You may very well, if we can solve the abundance question and if AI really makes us super productive, then as we discussed, some part of humanity, 5%, 10%, will be productive enough to be able to supply all the rest of humanity, 95%. the tax system in some ways we're already doing that right. So we'll find a few people who will be super value-ad because they know how to push the machines, program the machines, and really stay ahead of the machines. And there may even be a fusion of sorts where these humans merge into the machine in some ways. And then with that productivity unleashed the rest of humanity can really lead very good prosperous lives. So we may end up that way. But then you have to decide which side of that fork do you want to be on? I know which side I'd like to be on, which is obviously love to be able to use AI as a cognitive partner. But that means I have to invest a tremendous amount, a tremendous amount to really understand what the hell is going on. Read great books like you was listened to fabulous podcasts like Brave New World. Otherwise, as you know, in Brave New World, we ended up with Alphas and Betas. Yeah, really. Yeah, really. No, I mean, in fact, someone asked me this question as to, do you have any idea about what proportion of people will be on either side of the divide? And I was like, I don't. But that's something that's probably in our control. It just depends on whether we get with it and are able to use the machine as an amplifier as opposed to a crutch, which makes us dependent on it and leads to cognitive decline and we stop thinking. So yeah, I tell people that all the time. And I'm seeing it in my students as well, you know, where some of them become incredibly productive, you know, because they able to leverage the machine and ask questions and answer them that they wouldn't even able to do otherwise. So it's a whole new world out there. No, no, you're absolutely right. And I think it comes back to your innate personality also present, which is that some people are shapers and others are happy to be followers. A lot of people might say that I'm quite happy to lead a good prosperous life and make sure that the survival questions that I struggle with on a day-to-day basis are taken care of. So as long as those are taken care of, I'm happy playing sports or doing my creative pursuits or spending time with the family. And that's perfectly fine. That's a great outcome for sure. Others, and I would characterize myself in that category. Our shapers are want to see, you know, the world evolve in certain ways. And for that, if that's what you want to do, then you have no choice but to get ahead of the machines. Now in the Elon Musk case or in the Sam Orlman case, etc. Not only will you become insanely rich doing that, but you'll also become insanely powerful, perhaps, as well. Because, you know, you'll have access to these resources and these machines that are truly game-changing. So we also then come to difficult questions of regulation and policy-making and restraint. How do we do all of that? And again, that comes back to what I was saying earlier is that in the AI age, governance regulation has to be architected into the way in which these models evolve and the AI sort of systems evolve. So those are those are very difficult complicated questions, was something. Obviously, it's wonderful that you're exploring them through your podcast and your book, etc. But I think many more of us need to be engaged with the set of issues because otherwise, we go blindly off the cliff, not really understanding how these things will shape up and how they will impact us and our lives going forward. Yeah, indeed. And to me, one of the big questions facing us and this is what I end my book with is, will we govern AI or will AI govern us? I mean, to me, the answer is not obvious at the moment. And it just depends on how we address a bunch of questions, having to do with the kinds of restrictions we put around AI, whether we put some obligations around AI and the kind of agency we give it. All of these to me completely unanswered questions at the moment, but critical ones. Exactly, exactly right. And I'm reminded of two things here. One is, of course, what Jeffrey Hinton has said, which is the only example we have of a super powerful system being controlled by a less powerful system is a mother and a child. That nurturing, that love, that attention has to be innate in the mother for her to take care of the child. So it has to be architected. And the second thing that strikes me is a very important rule of systems, which of course, I subscribe to fully, which is its structure, drives behavior. If we want the right nurturing and protective behavior from the AI machines, we've got to embed it in their structure. And here, as a matter of famous laws of robotics, for example, come into play. So for instance, can we take those kinds of laws and embed them in the AI architecture deep inside the models so that innately and at a structural level, they operate on that basis. But those kinds of decisions, those kinds of structural and architectural decisions were saying, I'll go into happen in the next few years. That's why this is such a decisive and important time for not just the 21st century, but for humanity as such. I for one, as a technologist who spent 50 years of my life immersed in science and technology can never think and never have experienced it time when so much is changing so fast and it's so perniqued as it is right now. I mean, I guess on a couple of podcasts recently, one is a motley fool and another one was Jerry Baker's Wall Street Journal one, and I forget to ask me this, but it was, you know, to the effect, you know, machine's taking over. And my statement was that machines don't actually have to be evil to take over, right? That we could slip into a Huxleyen kind of a world through just gradual disempowerment, you know, where we gradually seed access to or seed control to the machine. And an example of this is, you apply for a job these days, you know, your CVs streamed by an AI, you know, increasingly now you're being interviewed by an AI, right? So even to get to employment, like the AI is being more gatekeeper and we've sort of, we've handed over this control without like really thinking about it because, you know, you get a million resumes, well, no one's going to look at them, it's not feasible to do it so you let the machine do it, right? It's sort of feasible. But inadvertently, we are sort of somehow disempowering ourselves, you know, and so that's one of my concerns is that we fall into a Huxleyen world without intentionally wanting to do so. Your colleague Jonathan Hade, for example, has already sort of said in the anxious generation that social media is hijacked young people's brains and of course, that's why Australia has the ban on social media for, you know, children under 16 and in a way, you know, things like TikTok and Facebook, etc. Have already taken over control, right? I mean, people have these withdrawals in terms and they can't get off social media. And so who's actually who's actually controlling our brains? Yeah, Jonathan has been incredibly successful in actually getting these issues into sort of mainstream thinking, you know, and being able to actually do something about it and having policymakers act on it. By the way, he has a new book which is supposed to be accessible to kids. Great progress. My concern actually is that what we've seen in social media, we could see that on steroids, like things go south, you know, really badly. And we've seen that already with companions over the AI companions, a couple of kids have committed suicide, you know, in one case, the AI actually, you know, even offered to write a suicide note. And it is just to me that that's a really dangerous kind of road to go down because the consequences of error are so huge that it's one of those areas where I wouldn't actually trust the machine to be actually be a companion to a 14 year old. You know, these are the easy I chat ports are now seductively powerful, right? And they will tell you what you want to hear and add to it. And now, of course, with image generation and live motion video, becoming so accurate and so photorealistic life like, you will have somebody on the other side of the screen who is an AI avatar, seeing exactly what you want to hear and being so seductive in so many different ways, it will completely overtake you and completely, you know, control you. So how do we protect against that? And it comes back again to what I was saying earlier about AI architecture and structured drives to behavior and we've got to embed these things into the models right now. We have to go back to the asm of type thinking around, you know, fundamental principles, fundamental laws that can strain and control AI. And again, as we were discussing earlier, you know, this has to be done now. We cannot do it, you know, 10, 20 years from now. Already as you can see with social media because nobody frankly anticipated its pervasiveness and its power via VFIA. Yeah, indeed. I mean, this is, this is one of the most vexing problems confronting policymakers at this time, you know, is how do you really avoid going down this slippery slope without hindering innovation? And I think it's possible. There's certain things that, if there's a credible threat, if you're causing harm, for example, post after the fact, you know, that's not preemptive regulation. But at the moment, we don't even have that, right? So it's complete Wild West. Yeah, the important thing is to have these kinds of discussions was and then for lots and lots of people to listen to your wonderful podcast so that they can educate themselves and learn about what the stakes are and where things are. headed so that they have an informed view and can take necessary action. Thanks for that giant. It's been great thanks for the time. [Music]