#19 - Lightspeed India

25:51And politics is nowhere in the world is good or great. So India is basically no better or worse than any other country. You know, the thing is, I disagree. I actually think India will be, India is going to be, I think, a dark horse in terms of it's surprisingly well run. What's one example? The votes for the recent election, they were counted in a day, right? For a billion person country. And so while the US, it took like weeks, months to get it done. Right. And that just shows that state capacities on some acts is actually flipped. Right. UPI is very, very good. other is very good. You have criticisms of these systems, but an identity system that works for a billion people, a payment system that works for a billion people, a voting system that works for a billion people, that's pretty good.

26:46Yeah. Right. I mean, ISROs, land or probe on the dark side of the moon. Yeah. We have these new trains. We have these new highways and stuff. All this kind of stuff. The airports, right? And yet it's still at the J curve. but still at the bottom of the upward trajectory. And the way I think about that, and I'd love to know your thoughts on it, is now that it's overlaunch geo, we have the cheapest 5G internet service in the world, right? And all of these things we observe in India, and that's actually quite new. You know, what else am I missing? Those are pretty good. There's a whole tech ecosystem there, right?

27:27Yeah. Good help. Awesome. Yeah, so absolutely right. you're not missing anything big. Maybe what I would add to that is India, a bit of geo thing that happened nine years ago now. I think India has one of the most digitally awake opposition in the world in terms of more limited time this fashion. So you see a few periods of ablation. Yeah, so in 2015, Vokish Ambani, who was one of the main guys at one of the biggest conglomerates in India, called Reliance, basically took this generational bet. He said that I'm going to spend 10, 15, 20 billion dollars and literally put it in the ground. He dug up places because, you know, you can't when you're that day, you can take anything you want to the government and let you do it.

28:12But he did it in a way that actually was quite strategic and really has changed the nation's trajectory. He essentially provided literally free internet to the entire country. In India, for folks that don't know, the price of 4G and now 5G is, let me tell you what I pay. I pay about 170 rupees per month. And I have about two and a half, two dollars odd, roughly. For 5G internet. For 5G internet. And I have a 3GD per day limit. So I get around that pace. I never hit that. And if I don't hit it, sometimes there are some plans that actually rolls over to the next one. Sigma will watch TV on Netflix.

29:01It's like basically 100x that amount in the US. Yeah. Yeah. So it was crazy for me. And I don't know. NGX, I don't know the entire thing. Something like that. Yeah. And Netflix in India is$699 or whatever it is in the US. I mean, so on and so forth. Sure. So digital infrastructure has become so cheap and cell phones cost maybe$1 ,100. $699 in the sense of$799 rupees? Rupees, yeah. That's about less than$10. Yeah. So when Jio essentially digitized the entire country, they became the rails on which everything else kind of got built out. Quasi-public, quasi-private infrastructure. Quasi-public, quasi-private.

29:39They had the spectrum and all that good stuff. And now we've got world's cheapest phones. And then most people have several phones. Like, you know, I know people that are in the up-and-coming part of society that actually have two phones. You drive in an already shine in India, you'll have two phones, one for Ola, one for Rupert, and you can take rides on both. And these are both phones that probably cost us in$10. So that's been very interesting to see. And on top of that, you could then, you know, when you get distribution, mass distribution, you can build interesting companies. You can experiment more quickly and learn what, you know, what works for a consumer versus not.

30:16So the target is a venture ecosystem come up and take more bets and funders coming up and going on building riskier and riskier things. And then we saw essentially this massive amount of entrepreneurial energy sort of just really coming out and just like bursting at the seams. So the last 10 years in India, I've just been there six years now, I think it's been a remarkable shift in terms of the risk appetite the nation is willing to take. In terms of the infrastructure built out, the country is doing. We spent more in infrastructure bent out in the last five to six years than we spent in the last 70 years since independence.

30:59And we are spending more than what we spent in the last seven years, the next five years. Well, the nuclear reactors are coming online. That's a really big one. Brats will be a thing of the past, hopefully, if those come. India is actually either number two after China or it's either there's a graph of nuclear reactors we can put on screen. but plain nuclear reactors, I think India is number two. Actually, on lots of graphs, for example, steel production, China's far enough by number one. But then when you delete China from the graph, India is the real distant, but real number two. So that's on steel, that's on nuclear, solar, it's on the number of unicorns, for example, it's US, China, then India is like distant to China, but it is number three, solid, right?

31:42So it is actually really rising and really, I think, underpriced in a sense and on the world stage, right? And of course, it's possible to overhype it. It's possible, but it's also possible to underhype it, right? And, you know, in a sense, I'm, I think I said to you, I'm moderately bullish in India, but extremely bullish in Indians, right? So India as a system is getting good enough to be a launch pad for the world. Yeah. And, you know, one of the things I think a lot about is, this is a very, and maybe you have some estimates. So there's about 7 million or so Indian Americans, right? I think that's a rough, rough number.

32:21And they're doing pretty well, right? Like they are, you know, producing now. Now that the second gen has kind of, you know, started to grow up a bit. You've got, you know, people in politics like, you know, Vick or Richard Vance or Cash Patel. You have people in media, like lots of, you know, like M. Night Shyamalan was one of the first, but lots of other folks. obviously tech very very gas-based finance and so on and so forth um but uh but there's like way and of course medicine lots of other professions engineering and blah blah uh but uh there's an interesting and the diaspora's done pretty well in other countries and it's certainly in Dubai and in the UAE like that's like about 88 percent Indian and that's a country that's doing extremely extremely well right at the UAE um and but the question is how much potential is still in India Because there's a degree of selective migration, right?

33:13There's not many people are coming to the US or places they have an engineering degree or do some qualifications. So many back in India are at that talent bubble. And even if it's 5%, I don't know the exact number, right? If it's 5%, that's like 70 million people out of 1.4 billion, right? That's like 10x what the Indian American population is. So that's a massive new economic disruptive force on the global stage. That's like a new Germany coming on. Yeah, right. A new Japan kind of thing, you know, like at that scale coming on. And, you know, maybe you have some thoughts on that. Yeah, I think, look, I mean, we obviously are very bullish on India.

33:53I mean, we have continued to double down in the last 10 to 15 years in India. We feel like this is where, you know, the maximum alpha is right now. And you're right. I think India feels underpriced even at this point in time. Or Indians. Indians feel underpriced. Yes. Finish your saying, I'll say something. What do I think about the potential of India? What is starting to get realized? What is still latent and will be realized? Oddly enough, I have been a lifetime skeptic of governments. And in the context of India, I came in skeptical. and one interesting thing that I realized is, there are two things.

34:41One is, if you are a government who is not proud of your country, you're not going to improve the country. Right. If you are a human who is not proud of your body, you're not going to improve the body. If you're somebody in a relationship, you're not proud of your relationship, your partner, you're not going to improve your relationship. For the longest time, I think India suffered. They didn't have a sense of national self and pride. They never had a sense of national self. And there are shades of gray in there. There's a lot of places you can slip and fall in. But I think it is an important thing that ties India and Indians together, wherever they are.

35:17And I think that sense of national self was not there for the longest time. When I was growing up, people just woke up and complained about India all the time, right? Or Peltz. But very in the 80s and 90s, yeah. 80s and 90s. It's not always working. It took two years to get a telephone in your house and for a TV. I was hanging the line somewhere, blah, blah, blah. Now I think there is a sense of rising optimism. There are pockets of pessimism for sure, but a sense of rising optimism as a whole. And so where am I going with this? I think for a country as large as India, 1.4 billion population, you have a challenge that requires a platform level change to happen.

35:55Jio may have been a platform level change, policy is another platform level change, the kind of governance that introduced... 1991, globalization in 1991. So, we see, that was a moment for people to know. From independence till roughly 91, India was, they actually wrote socialism into the constitution of India, like the socialist state of India or something like that. And so, of course, Fabian socialism arguably was the second colonialism, right? Because Fabian socialism came from Britain. and so it was something that kept india poor and it was you know all these ngos were handing out moms like the whole mother teresa thing is meant for a country that's prostrate right yeah and like basically you know it's not actually meant it's investment is more investment is a real charity in a sense because charity keeps somebody dependent and investment makes them independent right yeah so charity you know like i said if you give a man a fish you keep them independent but if you invest in their fish company, they might actually do something, right?

36:54So until basically the late 80s, early 90s, India was essentially socialist and the licensed Raj and it was impossible to do anything and talented Indians were just either like kept poor or they left, right? And then 1991, because of the crisis, you know, the Soviet Union fell and so on and socialism was clearly not the way to go, India made the right turn. Now, by the way, that wasn't guaranteed. like Cuba remained communist, North Korea remained communist, right? So it was the right leadership. We saw the right opportunity at the right time and took the right door. And India, it went at least somewhat capitalist.

37:31Yeah. And there's still a million criticisms you can have, like air conditioning is taxed. And it's, by the way, if there's one thing we should get all Indian VCs, founders, and so on, basically repeal the tags, the luxury, like the Indian air conditioning tax of all the things that's a very simple line item thing to repeal will completely transform India. Right? Do you know that tax? Yeah, at the dog you have a new thing where you cannot put your air conditioning to 19 degrees Celsius. Something stupid like this, right? But at least that one is a little harder to enforce or whatever, right? The AC tax though is just you know, Lee Kuan Yew knew air conditioning was necessary, right?

38:14It wasn't like an option like a sub, you know, like a It's like a tropical kind of environment, right? But why did I get on that, right? So, oh yeah. So after 91, capitalism hit India and started to grow, right? For China, it was 1978. So that's why China's in some metrics, roughly about 10 years, 13 years ahead of India, roughly. Other metrics, it's way ahead. Other metrics, you can argue, it's even behind in some metrics. But like China's diaspora doesn't do as well as India's diaspora. China plays a better home game and India plays a better away game and vice versa, I'd say.

38:53So that was the moment that India turned capitalist. That's why India is actually rising now. And I think over the years to come, China was always the number two communism after the Soviet Union for decades. But then the Soviet Union made a bunch of bad decisions and basically tried to ideologically reform for economic reform. And then it just blew apart in the late 80s, early 90s. And then China, in a sense, became the leading brand of communism. Right. Even if it doesn't export communism, even if it doesn't think its system is the ideal system, it became the leading brand of communism. and Vietnam and others, you know, like basically all look to China or what have you.

39:34And now I think we're on the verge of something similar happening in the Western world where the US and the Western countries, like what they call democracy is going to financially, socially, politically, physically melt down, right? The sovereign debt crisis is going to hit. And I think that if India can weather that, that it actually will keep, in a sense, the flame of democracy alive. It'll be like China was the number two that then became effectively the most pragmatically run country, whether it's actually, quote, communist or not. It preserved the name of communist. So India becomes, I think, the world's leading democracy by the late 2030s, early 2040s or thereabouts.

40:16I think really the Internet is also a big part of that because the Internet allows us to do cryptographically verifiable voting. It allows us to do, you know, like votes where you can not just, you know, your vote counts, but you can count all the votes. So the internet also is a big part in kind of guarding, quote, democracy. Because communism will be very strong. Chinese communism, I think, will be... A lot of people will think that's the way to do things after the sovereign debt crisis. And I think it's going to overcorrect in that direction. And we'll have to have a counterbalance to it. I don't know if you have any thoughts on that.

40:44Just the fact that India can count all the votes in a day, for example. It's not a small thing. It's like a detail, though, that shows that operationally it's actually able to do that. And somehow, on net, it's working within India. Look at how it was there. Yeah, I think so. I mean, back to your point, I'm going to merge the two about the potential that India has been to realize and the potential that's latent. I do think some of these big ships happen when big bets are taken and some of these big bets can be taken by large agglomerates and governments and policies and all that. And I think that's what's begun to show some promise and results in India and the kind of stuff you see in India different from 20 years ago is a result of that.

41:25and because the government and you know the globalists have done well uh it is creating a sense of optimism and is becoming a petri dish where ambition and risk taking can survive and drop 20 years ago when i was in college if we were to start a company in honor perot you know you would not get married uh not if you're on tinder i think if you're a founder you get all updates is what I hear. So that cultural shift, I think, allows ambition and it doesn't tap it down that he says, yeah, you could do this and, you know, go on that journey and there are people who back you. And if you come back, fail, you don't have to be a joke, you know, simply get patted on the back in a bar.

42:07So the failure tolerance, the cultural shift is important. Yeah. As opposed to only taking the safe job of engineer, doctor. I mean, of course, being an as a kind of engineer, but yes. Correct. So now I think that the pituit dish is somewhat ready. It still will take a decade, plus multiple decades perhaps to get it ready to the level we need. But now I think we are starting to see ambition and risk-taking happen, get back and then get results. And we've already been in the first cycle of that. And you could argue the first cycle was 2013, 2014, and now it's sort of showing up in public markets in India, where a good company is, a matter of coming.

42:41You just hung out with Bini Bala and Silverfoot Cart, right? And so we put up a clip. He was from, you know, a friend of both of ours. He's on the pod before. And basically the, you know, he was the first large exit in India, right? $18 billion for the flip car and kind of showed it was possible. Right. And, you know, now, you know, one thought you have a thesis. Why don't we get to Indian tech for a second? You have a thesis on Indian tech. I have a thesis. I'd like to hear yours. I'll give you mine. Yeah. Yeah. Thesis on India tech. Wow, it's a broad question. Yes. Teeth of India tech. Should you be mine?

43:19Yeah, pretty. Yeah, I do. Yeah. Yeah. So my view is that India itself is a market. Yes. But I'm actually more interested in India for the producer side than the consumer side. is say i'm more interested in india as a source of talent for engineering for finance for media for basically internet first basically all of the you know ai is amazing but ai needs prompting and it needs verifying that's actually the new bottleneck right because the prompting is actually writing programs in a higher language you have to be verbal enough you know to be able to do that. So I'll see Indians winning spelling bees as so you heard.

44:02Right. You have to be able to know all kinds of crazy vocabulary terms, actually. Like, for example, art history is now an applied subject. You know why? Because if you can say Cezanne or Picasso or something like that, then you can invoke that as a library call when you're using AI, right? And so every subject, more vocabulary, and not just vocabulary, the more terms, concepts, because using them correctly, you can prompt it better. You get a better result. And then verifying it is if you have expertise in there. So, yeah, it doesn't take jobs away, but what it does do is it shifts them to people who will have time to do prompting and verifying.

44:37Sorry to reject you, but I don't know if you saw that recent video by Karpathy where he does his analysis of GitHub code. Yeah. And more and more now, the GitHub repositories have English embedded in the code because now the code itself is becoming natural language more and more. Yes. So, back to your point about, yeah. That's right. And exactly like prompting, programming, assembly, like, you know, right. And India is good at English, or at least a lot of Indians are good at English. And it's in some ways, depending on the numbers, you can argue India has already actually become the number one source of English speakers on the Internet.

45:13Right. Over the US, which used to be the default. Right. And I think as more and more Indians acculturate English, that will become completely true by 2035 or 2040 or thereabouts. because they can learn languages quickly or rather. So my view is if we're sitting on this chair, this chair doesn't feel like a Chinese chair. It may be made in China, but it doesn't feel like a Chinese chair. In the same way, I think what is a thesis for Indian internet entrepreneurs? They'll build software that's just globally competitive software that happens to be made by an Indian or even in India, but it is internet first.

45:51It's not India first per se. it's internet first. It's almost a little bit like Europeans who, if they're in a small country in Europe, they have to actually think internet first if they're an entrepreneur, because the whole market is just not big enough. And so like the translation has to be there from the beginning and, you know, they have to be internet first. That's got its problems because that's a tax, if compliance and stuff you have to do, but it's got its benefits because you have a holistic view on the world, right? And so like, you know, build in India or build by Indian, sell to the internet.

46:26That's the kind of business I'm interested in investing in, which is an internet first business that happens to be founded by Indians. Now, of course, sometimes it may turn out that something like Ola, for example, is necessarily got a very local and geographical component, right? Some businesses will be built, which are, you know, necessarily going to be domain or geo locked and you have to start there. And I'm not saying India is a pretty big market. You can definitely get to a pretty large market over there. But I do think intranet first is the way to go. And I think crypto is going to help with that a lot because it means global rule of code.

46:58So an Indian may be a second class citizen in America where H1Bs and so on are becoming much more attacked and so on in the US. And they're basically, they haven't eliminated them, but they made them so much more of a pain in terms of forcing people to go back every year rather than every few years and making the renewals more longer and uncertain. That's not a small issue. That's a big issue because it means you can't, first of all, just the fact, people may not know this, but as an H-1B, you have to periodically go back to India, wait for a consul appointment, then come back. Now, if you're just working at a job, this is like, okay, I need to pause.

47:37I need to take an international flight. I need to wait for an uncertain duration and then come back. And that is a pain, even if it's like a day on both sides. If it is an uncertain multiple month thing over here, you can easily just lose your job in the US. It makes hiring an H-1B something where the bureaucratic overhead means you wouldn't do it, right? So their point is to make it, their goal is to make it disadvantaged enough where the program still exists in name, but not in substance, Right. So as that door closes and I don't think that Indians should go to America. I think if you're you, it just like a big not welcome sign is being put there.

48:14If you're a CEO, the much of the American left hates you. And if you're an Indian immigrant, a good chunk of American right hates you. So just listen, you know, to the signals there. Right. And just don't go to America anymore. And it's true for many other countries as well. Like even German tourists are getting strip searched and what happened. right? America does not want immigrant entrepreneurs to come there anymore. It just doesn't, right? So where do they go instead? Somewhere else, right? And so we have Network School. You know, we've opened that. There's Dubai. There's tons of other countries that have digital nomad programs.

48:47And of course, there's India itself. And finally, there's the internet, right? Which you can be on from anywhere. You can be a digital nomad, right? And I think that shift, I'm starting to see that shift. Somebody posted the other day that like there was two really great Indian entrepreneurs that just chose to, you know, not take their US offers just stay in India instead. And I'd love to hear your, or go to Dubai or go to Singapore, go to something like NS and so on and so forth. I'd love to hear your thoughts on that. Yeah. First of all, I do agree how hard you stayed. You came back. I came back.

49:16Yes. You're an early adopter. Yeah, that's right. Yeah, I came back before it was, I mean, there were a couple of waves in India that people have returned to India. One was in 2010, I think, 2009, 2010. It was in 2010. And the next one, probably again, around the time I moved back as well. but mine was normally a rape thing. Just move back because I want to move back. But I do remember how hard it was to get my H1B stand. I got stuck in Mexico in a place called Matamoros for six days to get my visa appointment. And it just kept getting delayed. And that place was literally bombed out a week before because people had stayed and hidden there, drug dealers had hidden there.

49:52So the Mexican police had literally had to bomb them out. So the far away hotel was actually a big circular hole in there. So we were living there. Anyway, it was an interesting time. But yeah, it is getting harder and harder to stay in the U.S. or enter the U.S. And we're all about thinking about the 5 and 10 year and 15 year horizon, right? So you make an investment today, you're thinking about where the world is going to be in 10 years. Correct. Is America going to be more or less welcoming to immigrant entrepreneurs in 10 years? I think it's going to be much less welcoming. Much less welcoming.

50:23That's what I hear. That's where you mean worse than today. Yeah. Worse, right? Yeah. So if it just is getting worse, you never want to bet on a curve that's going like this. or when a bad uncle is going like this. Yeah, it is much harder now. My friends that are living in the U.S. 10, 15 years are afraid to come back to New York to see family because they feel that when they go back to the U.S. or they legitimately stand for me, they will be stocked at the border. I mean, no internet. So I think it is getting surprisingly back. I was quite shocked to hear that just a couple of months back when I was in the U.S.

50:52It's also, the thing is, just the randomness of it deters a truck. Yeah, right. Because if you have a 10%, Like you're spending money on this plane flight, on this hotel, all this kind of stuff. And then you can be just randomly stopped at the border. It's almost like an investment. Why did you make an investment that can randomly go to zero? Why did you have a conference in America, like a tech conference, if you can't have international guests, right? Tech is an international thing. You can base it there, right? Yeah. So it has lots of knock-on effects. Yeah, I mean, the debasing of the institutionalization of the institutional powerhouses of the U.S.

51:29is quite sad to see because I kind of really benefited from it. It was probably one of the things that made a superpower to begin with. So, yeah, I think it's sad to see. And I don't know if the trend line is ever going to reverse, but it seems to be accelerating and going to the end. Now, coming back to your question on India and tech thesis on India. And again, it's a hard question for me to answer because at the end of the day, we are in the people business and I'm not in the business of predicting what future needs to exist. I'm in the business of just understanding the motivations and intentions of a founder that's trying to create a future of a certain kind.

52:03And as well as anybody else, we are primarily in the people business. So my intention is to keep the best people in India, stay in India, and build in India. Even the best people from around the world to move them back to India somehow. That's my primary interest. Interesting. Okay. So now what's interesting about India is... Do you only fund entrepreneurs in India itself? There has to be some connection. For the India fund, we have the background entrepreneurs from Southeast Asia and India. Okay, got it. from a viable fund, you know, different regions are different funds for us. So what is interesting about India, you mentioned this earlier, that India actually lives in the cloud a lot more because it's much cheaper for us to go live in the cloud than any other country.

52:48Everybody literally lives in the cloud. India has a lot of free time. And they watch the same movies as the Americans do and they play the same games as, you know, all around the world. And, you know, they are all sitting on Reddit and shitposting everywhere on Twitter and all that stuff is happening. So the more unplugged you become from the physical infrastructure, the more plugged in you are to the digital infrastructure. That is a very fungible thing that you can move people around with. So they are much more at home because they speak English. They are culturally easy to assimilate in any other culture.

53:23And they can move around. So what you might be seeing, and I would love to hear how you're seeing in that school, I think you began with the idea that you can literally put together a set of people that believe in a similar vision, that have similar culture, and the culture is not regional related. It's more about a digital culture that exists in the clouds and sorts. You can institutionalize that and create a movement around it. And I think Indians are actually a really strong and a great perfect fit for this new world that is emerging, which is going to be this network state of sorts where they are the native citizens already.

54:04Yes. And they will have the least friction to get a passport and get citizenship in this new world that you are creating. And what I mean by that, by the way, is it's a little bit like people understand fiat currency and cryptocurrency now, that there's like an internet native version, there's a land native version. And I think the next step, which we've kind of already in the middle of, is fiat identity and crypto identity. Because, for example, your Google login, right, is in a sense a digital identity, right? A very important one, in fact, right? Your crypto wallet is another one. Your pass card to your hotel room, your Tesla digital lock that opens it where they use a phone app, and your 2FA, right?

54:47All of these kinds of things are pieces of digital identity that eventually I think will get consolidated into like effectively a crypto passport, or digital passport that gets you into online and offline property seamlessly. Right. And India's part of the way there with Adr and other things, right? Like that actually can log you into online services, but it also gets you offline, right? Yeah. And they have, what's the thing at airports where it's just like face recognition? You get a lot with DigiAltra. Yeah, DigiAltra. Does that work? Yeah, it works beautifully. Yeah, I don't have to carry my...

55:21It's absolutely not matter. So Nidhi Atra is this digital entity that, you know, it essentially scans your face at the airport. So then now you are, you know, you are in the system and then you don't have to carry your passport. Well, for international travel, you still have to. But domestic travel, you don't have to carry any entity to prove. So what you do is you just stand in front of the screen and then it takes your, it scans your face and you are, the door opens and you get into the airport. And it used to be, you know, a few years back and it still is there for majority of the airports.

55:49in india here to show it i need you take your wallet out and you carry if you forget your i need a card and you gotta go back home and pick it up and today i don't really carry anything to card with me i actually don't really carry my what i do really remember to carry my wallet when i'm in singapore right now or the us and i tend to forget you don't need a credit card debit card uh easy saw digital upi and qr code base you don't need to carry an idiot card i have this thing called digital locker which keeps my ditty and it's a legal as a legal uh way to show your duty to people. You cannot be denied entry if you show that and so on and so forth.

56:23So now when you combine this digitally native cloud native population of very, very young people, really hungry, really ambitious, and you combine that with the infrastructure the government is providing, combine them with a bunch of money that is then to take risks with very young and unproven founders. And also pretty capital efficient there. very capital efficient yes right uh then you get this in that's where the future of india is and on top of that you have this fast growing uh midi floss with higher and higher disposable income in frantic india has obviously multiple pots there is a uh really really rich part of india uh has very high per capita gdb there is this middle part that pays differently and there is a bottom the bottom is really quickly coming out of poverty and below poverty line it's coming out of poverty uh some disposable income and then they are very you know we are spying to be somewhere the one or the middle uh what does that mean uh what it means is they want to go take, you know, take their girlfriends of boyfriends to a movie once in a month, once in six months.

57:38Great. Take them out to a lunch or dinner and a small red straw. Consumption is starting to increase, yeah. Yeah. If you look at the number of iPhones India buys every year, it's quite massive. It's surprising to us. That's another important stat that I was surprised at seeing recently. India's risen from 1 % to 20 % of iPhone manufacturing. Which is, again, it's like, it's rare that something keeps surprising me to the upside. Yeah. You know, like as you said, like governments, things tend to, you know, somebody said something like, India disappoints both the pessimists and optimists. And optimists, yes.

58:13It's actually right. It's true. It's true for, it's been true for India for the longest time. It probably won't be true forever. But that's what's interesting about India.

58:25Now, why is that happening? Look, 10, 20 years ago, my parents really saved up their entire savings to buy a house or for somebody's marriage. or they would buy gold or they would buy a vehicle. India, the new India doesn't care about that. Nobody wants to buy a house. The young kids don't want to buy a house. Interesting. They don't care. They don't want to have possessions. They want to rent. Yeah. Now to wear gold in public is tacky for the young people. So they don't want to buy gold and keep it in their locker and where the money go. They don't want to buy a vehicle because they want to just take Ubers everywhere.

59:03Like who cares? So they are featured in it first. They would be prodding to that. Yeah, they want to plug in, plug out of infrastructures, banking infrastructure, vehicle infrastructure. So the lifestyle of sort of the San Francisco tech guy is actually the default lifestyle of it, a much lower price point. But actually, comparable quality, honestly. It's like maybe 10 or 100x in some cases cheaper than California. But it's comparable or better quality in many cases, It's like food delivery, like Uber, all that stuff exists, right? You know, you can be completely productive with great coffee, MacBook, air conditioning, everything there.

59:43And it's not like super expensive or anything. It's like accessible to a very large number of people. Yeah, and you can get all of this stuff in 10 minutes in the big cities. Yeah. And pretty soon in the K2K cities in 10 minutes. Yeah. Because economics doesn't work out. Now, the one thing I will say is when people go to India, they'll find places and there are plenty of places that are still, you know, filthy or, you know, those like terrible streets and something like that. And that's absolutely true. And the way I've actually started to express it is India is like California in reverse. And what I mean by that is California is fractally getting worse.

1:00:17Right. So everything that's old is like beautiful and it's got like amazing, you know, Spanish roof architecture or its beaches. but everything new is like in california it's like a homeless encampment it's like a burned downtown it's like a way most set on fire or something like that it's like the new stuff is bad most of the buildings that are new are bad because they're not buildings they're destroyings really right and in india it's opposite the older it is with some exceptions like the taj mahal which are really really old right but in general the older stuff is like it's like a slum or it's like a you know bad street though poverty is radically reduced in india like the rate of extreme poverty has dropped off a lot, right?

1:00:53So it's like, you know, it's like a messy street or something. But the new stuff, it's like a shiny new Apple store or it's a new restaurant or something like that. And importantly, a lot of the people are very, especially English speakers, and of course not everybody speaks English and so and so, but a lot of people do, are very au courant with internet memes, you know? They're extremely conversant with everything that's happening on the English internet, right? and so that's surprising to people if they haven't had you know time to they may you know because people type without an accent yeah right you type without an accent and it's their complete conversant in written english and there's like a slight accent in within india but even then like you know because indians learn languages recently fast so it's fractally getting better that's kind of how i think about it let me know yeah yeah is that because you've also lived in California and India.

1:01:46Absolutely. I think in India, you can visit three to four countries in a day if you wanted to. Go to Mumbai. Tier 1, Tier 2, Tier 3. Yeah, talk about that. Pretty much, right? So you go to... Zythonol, Shahs, I think, from a mission. I think you mentioned that perhaps, and then our friends at Bloom Ventures, Nusajit's... Nusajit has that interview. Sure. I mentioned that in his recent, I think it was last year, a couple of years back in his publication, he does a really good job of that. It's called Indus Valley. And I think he's got this really, really interesting observation that, you know, in a single day, I don't know whether you said this or not, but it's my interpretation of it.

1:02:23In a single day, you can, if you go to Mumbai or Pagetur, you can land in an airport that looks like a seven-star hotel, drive on a highway that looks like absolutely brand new and flowers and trees on both sides and feels like you've come to Singapore. So you've been down in Singapore, you can then reach a city that feels like Mexico in the middle of a sweaty afternoon. And then you could pass by something that looks like a Brazilian favela. I probably find a bunch of countries in a single statement. But that's interesting in India. This is what, you know, there is a pocket of massive improvement and there is the messy middle.

1:03:02and there is a, yeah. But if we look at the graph over time, and like, so there was a tweet on Indian infrastructure that I put up that actually Prime Minister Modi retweeted that showed the, like, you know, for example, illumination, like lights or highways or, you know, plumbing, all this stuff just radically improved. Yeah, over a 10-year cycle, it's completely different. Yes. The 10-year cycle, the connectivity, the electricity, you know, consumption, If you look at those graphs, you look at the road network that you put this up. I remember seeing that a train network, the flight networks, the airport infrastructure.

1:03:39We look at a 10 to 15 year graph. It looks like it's a very different country. Yes. And you can still go. You can find whatever you want to find in India. The optimists will go find the Singapore in India. The pessimists are going to go find something else. Yes. But the pessimists have, if you look at the graph over time, it is very hard to be pessimistic on it. Right. And that I think is, unless you're calibrated on the trend line, right? It's kind of like people can take a beautiful photo. There's still beautiful areas in California. But the trend line from, you know, because I was there almost 25 years, right?

1:04:15From 1997 to 2020, 2019, so 22 years, something like that, right? The trend line was just definitely like this. Like it hadn't yet hit zero. But, and how long were you there? You were there from 15 years, so 2003, 2018, right? So you definitely saw it towards the end. It was like much, much, much filthier. It's a little different world, right? So it's funny because we're in a time where I think previously these sort of civilizational level changes happened over decades, but now they're happening over years. Yeah. Or they're, they haven't over centuries, now they're happening over decades, you know?

1:04:52The rate of change is so ridiculously fast. Yeah. And we can quantify that actually. Have you seen the graph of, for example, I think it was internet adoption or actually no, it was e-commerce. E-commerce spend had been growing as a percentual total spend X percent a year until COVID. And then it actually literally increased 10X in a year. So that's like the Lennon thing. There's years where decades happen. That actually was a year where a decade happened, where the rate of change just went like that. right things that people had known were happening just went vertical like this yeah now covid in india actually i mean interesting macro thing covid in america after that like the society has just felt worse off like people say people often post about this like um maintenance quality has dropped off you know polarization has increased even further service quality like you know it's like everything in terms of maintenance all this stuff is kind of broken down and um whereas in china for example somebody went there and they said post-covid and you know china was really ruthless during cover they locked up in rooms and so on yeah but they also a lot of the things that people used to do in china like you know spitting on the street and stuff like that like a lot of that is actually dramatically being curtailed and china feels more japanese um like there's a greater degree of public you know cleanliness and so on and so perhaps as the system got pushed into a new state as a function of COVID.

1:06:18And I actually think India on balance, because of the internetification on balance, it actually feels stronger after COVID, like the sites come together more and so and so forth. I don't know if you have any thoughts on that. Yeah. It's a funny tweet somebody put out a few days back that the world has not been the same since somebody ate a bat. So it actually appears to have come out of the wet lab or the, you know, as it weren't more than what okay because there's enough data on the nih funding of on the covet you know like the 19 labs that it's more than incidental there's too many different connections for it to be i think too constructed yeah i could go through all the genetics of it but it looks it looks positive so yeah i think look india and the covet was brutal time uh it's very hard to compare which country had the worst of it india certainly didn't do well and the first way was fine the second way It was brutal.

1:07:11Yes. People did come together quite nicely. The society, which was really great to see. Many of us, and I remember like being part of a lot of this COVID rise in the line, we would call off hospital and figure out events. And I think, yeah, it was a tough time. It does feel like it did not change society in a way that you could pinpoint pre and post and say we are a different world now, like what you're saying with China or perhaps, you know, the US. it doesn't feel like it's very different. So I think we did come away not getting too deeply impacted culturally or economically on COVID. It could be because India itself is a little bit conservative across, you know, personal policy and monetary policy.

1:07:52So we kind of... We didn't do the big print in the same way. We didn't do the same thing that we did. And the healthcare infrastructure, you know, it's not as massive as it is in China, but it, you know, it gets the job done. Yeah. And India did have its own vaccine. You did have its own vaccine. and i took the Iranian vaccine not the uh Pfizer one and uh and you know felt like you know i think we came out of it okay i'm glad i'm glad to see that it did lead to a lot of like the demonetization wave made sure that people don't really care about carrying money anymore i think it did lead to a lot of the digitization of services you know i wouldn't have a company you know like you could get people to come your house and clean up and you know you could get But, you know, quick commerce picked up during that time and deliveries of all kinds of things.

1:08:38And I caught COVID in the U.S. and it was so hard to get medicines delivered to you. You just couldn't get medicines delivered to you, COVID. And any IE just comes to your doorstep. Even during COVID, you won't come to your doorstep. So I think those things start to get adopted. And one of the funny things that happens during COVID is, you know, even grandparents had to learn all the stuff because many times they were living by themselves. or if their kids were ill or sick, then they had to take care of them. They learned how to use digital services at 70, 80 years of age. So it led to this mass coaching of the entire country to learn to take classes online, to transact online, to get services online, and so on and so forth.

1:09:22So I think that led to a behavior that has stuck around a lot more. So all the whole, it seems like it did accelerate our digitization journey, which is already kind of really getting there, but it's really accelerated. I think the digital part is what really accelerated after that. I think if you looked at a few graphs, so society is at least the same, if not better, but really doesn't seem to be badly hurt. The digital part radically. And we have a much younger population, so it's a big deal. That's right. So it hit us hard. That's important. And frankly, India. That's a good point, actually.

1:09:57the demographic aspect of it is underappreciated. Yeah. Yeah. And frankly, India doesn't listen. Like, it is don't listen. Right. It wasn't like people are still walking around and getting beaten up by the police. Like, a bunch of funny tweets coming out. But people just began to mingle. Like, in other countries, people protesting against during making masks. So, part of it was that India are both docile and don't listen. And severe obligation to tell people to wear a mask and many people as well. So that was good. And then sometimes people just don't care. There's a million of them anyway. And both of it kind of worked out in the way that the country came out of it.

1:10:36Okay. It was quite weird to see. Interesting. So, okay, let's see other things. I mean, you basically, you left in 2018 for India. And so, when did you come back to India? When I come back to India, you came back in like January 2018? No, I came back in August of 2018. Okay. So it's only a year and a half before COVID. Yeah. Yeah. Yeah. So then, yeah. And so what was the, like the venture environment at that time was, I mean, 2021, just talking about that for a second, right? A lot of deals that got done in 2021 in the U.S. had this giant, I mean, like markets are just going like bananas at that time, valuations are out of control, all kinds of stuff.

1:11:22And then inflation obviously went vertical towards the end. And finally, they acknowledged it was doing so. And then they had to hike rates and not had to. I actually think both the printing and the hiking were both disastrous policies, but they felt they had to be fine. How did that, like to what extent was India's economy also connected to like the Western gyrations of markets versus how much was it kind of its own thing? Not very much. Yeah. So if you look at the, if you kind of go back and maybe you can show that in 2020, when the COVID shock happened, the market's going to go up in the U.S.

1:11:59by what, 20, 30 percent or less a few months. In India, it was probably more like 7 or 8 percent. So the shock, because the first rate wasn't as bad in India. Also, we are not as connected to the U.S. stock market, you know, as perhaps other countries might be. so um so the shock was there but it was it was less uh and and then it just continued on uh uh so yeah on the venture side there was a massive gold rush even in india like seed rounds were getting done at you know 20 30 post money and the six months ago they were happening at 10 15 post money so really double overnight uh cdjs were happening 100 200 post money half a million dollars of revenue, especially in SaaS.

1:12:46So SaaS category kind of really blew up massively in India. And majority of the deals that happened there were probably SaaS and software kind of deals. And so the snap up into COVID frenzy, lots and lots of capital went into different companies. I think at that, maybe don't quote me on this, but about 30 to$40 billion of capital went into India across early mid and grow. So what's funny is that's a large amount for BC, but it's a small amount in the grand scheme of things, you know, like because so many hundreds of billions of dollars. But yes, this is a big chunk for BC. It's a good chunk of money.

1:13:24And for Hawaii, you were the world's second largest pool of venture capital going into the market. And China was third or fourth at the time. So I think now we've kind of stabilized. And the snapback out of COVID was also very brutal. Many companies, you know, going into COVID, especially if they were offline, really struggled. And then they used the times, the good ones, the good management teams used the time to really get leaner and get more hungry and powerful. On the way out of Kobe, all the companies that were building into the COVID way, what people used to call pandemic market fit, they also lost that and lost a lot of capital and lost the thesis and they eventually died.

1:14:03So all that happened in India in two years. And I had a few of my companies just blow up to a billion dollar and then blow down to zero again in like two years time. It was crazy. At the same time, crypto was happening. You probably know this. And then tons of tons of companies and the NFT was picking up. It was a frenzy, a feeding frenzy. And DeFi was picking up and the Sendian Africa in India. So we saw a lot of that. Lots and lots of gold rushes all happening simultaneously. It was really hard to keep a straight head and figure out what kind of entrepreneur you want to back and what company will wait to go back and what will sustain will unlock.

1:14:45So, we will be looking at the best we could. And I think most investors around the world struggle with that period. We'll see what that vintage looks like in the next 10 years. But it'll be challenging if you have been a bit more indiscriminating deploying capital within that time and didn't care about valuation and resume the days of stay. And I think they lose a lot of money. Yeah, interesting. I think in general, the one thing that is good about those boot times is, you know, the Carlo Perez thing that we talk about a lot of these things. You want to talk about that? Go ahead. All right. Please go ahead.

1:15:20Yeah. Yeah. It's like, in a sense, the dot-com bubble was a global search to find Google and Amazon. Right? So all the capital that went into it, it did more than that. But if it just did those two, you can also add Netscape and a few other things, which obviously Netscape is very important. If it just did that, it's worth it. Like all the capital. And so if that 30-40 bill, if it generates$1 trillion company, even$100 billion company, you can argue it's worth it. Right. And I mean, it probably will probably will do quite a few like, you know, I think kind of of course you responded in that, but that will probably be there.

1:16:01Right.

1:16:04So and, you know, it's because the thing is that entrepreneurship is hard and rare. And so you kind of need to take the whole thing and, you know, toss it on the table. Another example is you remember the chatbots boom of the mid 2010s. this is around the time of tay and microsoft and so and so there's like 2015 ish 2016 like floor type stuff like a like a program at a do this step one step yeah if else check that's right yes yes that's right and and there was microsoft tay which was like an r like a b0 ai chat bot and i remember seeing all that and i remember i'm not sure if i uh wrote it as an a-signcy memo or something.

1:16:46But I remember writing something like, for people to make this work, what they're actually, what they really want, as people are talking about with chatbots, I'm like, what they're asking for is iterated Google. Like what they want in its full generality is to say something in natural language, get back a response that uses the entire internet to make the response and then do another query, right? That is iterated Google. So if you can do that, of course, it's amazing. It's more valuable than Google, but that's the scale. The problem we're talking about is like beat Google, right? And I remember writing that.

1:17:16And that's actually what happened. Like ChatGPT has actually iterated Google and it's actually better than Google and it's actually taking share away from Google. But that was the scale of the challenge. And actually, Greg Brockman actually got interested in AI because and part of the chatbot boom. He's written about that, right? So in a sense, all the capital that went into that, if all it did was get Brockman interested in chatbots, it paid for itself. Yeah, right? So that is like a really right tail way of thinking about some of these bubbles or what have you know. It's crazy how powerful the power law is.

1:17:51Yeah, exactly. Or actually, Docker Bubble, what else it is, it did PayPal, right? So it did Netscape, it did PayPal, it did Epinions, which is a real company, but it did Google, it did Amazon, right? So PayPal gave Teal and then LinkedIn and everything else from that, right? And YouTube and blah, blah, blah. Then there's Google. So just that search for those entrepreneurs was kind of worth it. Yes, right? And that's so funny because that's not the way that people normally think about it. They want to score every investment independently. And of course, that's worth doing. But it's sort of like no bubble actually goes to waste in the long run.

1:18:25I shouldn't say no bubble, but you know what I mean. Many of these ostensible bubbles actually are a search function, which actually pay dividends in five or 10 years. So, man. Just a way to look at it, yeah. I agree. I think if I were to redesign the venture world, I would sort of say it's going to become a Vesheva-Bilogel where the extreme right are these absolute wool shots. You know, no shard in hand working, but if it does work, it's like, you know, you cure cancer. Yeah, do a lot of stuff. On the extreme left is, I don't know, roll-up companies. Guaranteed performance, but, you know, 15 % coupon year-over-year.

1:19:01And everything in the middle may not be as interesting. That's one way to think about that. That's interesting. I mean, the thing is, I think it's hard to do the left-hand side company nowadays. It is very hard. And the reason I think it is hard to do that is like the argument, you know, for example, the so-called lifestyle business argument that 37 signals made for a long time. The problem I always have with that is if you don't intend to scale a business, then you intend to run the business for the rest of your life. So the lifestyle business is actually the rest of your life. Whereas it's actually easier to build a business that is a gigantic thing because then you can sell a piece of it at some point.

1:19:46And sometimes it's a project you do want to run for the rest of your life, right? But unless you want to run it for the rest of your life, you want to actually build it as a company that you might be able to sell and then do something else in life if you want to do something else in life. Number two is a lifestyle business is not necessarily more competitive in the market than a scale business because it is attracting almost definitionally less ambitious people. That's the bigger problem, I think. That's right. And so the ambitious people who want to conquer the world will optimize the heck out of every screen and every flow and so on and so forth.

1:20:17And even something we think of as simple, like, I don't know, nowadays people, because something is easy to them, they sometimes, unless they're experienced, they think it's easy to build. Like WhatsApp or Discord, you don't necessarily feel like you're doing something very sophisticated when you're using it. You're just typing in some chat stuff to somebody on the other side. It seems basic. When you want to implement it and actually make it work at the scale that it works, it's actually quite challenging, quite hard. Like just video uploading from, I don't know, Brazil and having it viewed in India and America and Japan all at the same time seamlessly with replies coming back and, you know, deliver at least one semantics and 500 other kinds of things is hard, right?

1:21:00So the lifestyle business kind of thing isn't thinking about polish, I think, right? So that's my counter argument is I'm not sure the low risk business exists as much anymore. You know, you're right. Low capital though, to edit what you're saying. Let me take part of what you're saying, which I do agree with. I think the small business exists. The small business with a few people that with AI and with crypto and with internet distribution can do really well. Like Telegram is only 30 people, right? And WhatsApp is only 55. Instagram was only like sub 20 people. That business definitely exists and that business will continue to exist.

1:21:43So it's just an edit on what you're saying, Krause. Not low risk, but low capital. Yeah, and then I guess the thesis that some people have, and I don't know how this is going to play out, but you're absolutely right. I think what you're saying is, quote, unquote, you put 10 B plus people and put them together, they could be 10 companies, less ambition, less finesse or whatever. The combination becomes a C minus automatically. It doesn't become 10B plus. It does not even stay B plus anymore. It should become less. Exactly. I totally get that. I think you're also right that if the job of a business is to suck in$1 and spit out$3 through productivity gains and high margin work or higher value add, with AI, you suck in half a dollar and come out with five.

1:22:30And it may be possible because all this sprinkled magic of productivity and quality and scalability now finally gets solved with AI. So instead of one coming out with four, you have half coming out with five. And we'll see. I think it's showing promise at small scale. The challenge is that when you add 10, 20 of these 0.5 capital suckers and gold 50 come out of the other end, we'll see. Yeah, I mean, the big thing about AI is, I mean, there's a lot of big things, but the fact that it has scalability and consistency and low price, even if it does make errors, is still so insanely valuable because you're just hyper deflating the cost of the legal consult from a lot of people.

1:23:22So, yeah, I think, I don't know. And you have a lot more resilient P &L. What happens today, if you're building some of these old school businesses, you are more people heavy. So if you don't control demand, which small businesses don't have a lot of control on. Actually, go ahead. Finish what you're saying. I'll say something. So you have a fixed cost. You have people sitting on the bench and there's nothing to do, you know, figuratively speaking. But the AI and if you have agents doing the job, then at one point it'll be able to do pretty much end-to-end of a variety of things. Okay, I'll come back to it.

1:23:56Okay, good. And then what happens is you can just purely focus on scaling demand because supply can scale infinitely with demand. So you shift your fixed cost into an operating cost. You lose operating leverage, but you are essentially scaling a business with the exact same repeatable, scalable, predictable cash flow profile. The moment you get 10x demand, you can copy paste 10x agents with the same quality of service. And the cost will become valuable because you're paying more for infrastructure for 10 times more agents. But you are essentially in lockstep with demand. The demand goes out. The customer says, you're not good.

1:24:36I'm gone. You can kill relations and you're back to lockstep with the demand. So your revenue and margin profile stays pretty much in lockstep. But you are a lot more resilient through the cycles, which is important for a small business. A lot of businesses obviously control all of the stuff. But anyway, that's what we're starting to see now happen because of it. Yeah, please. Yeah, so I think the only part I'd be describing on that is the end-to-end part. I think it depends on the task. So if it's a robot placing things, that is something where the verifying of whether it's placed is easy. But many other tasks, it's actually middle-to-middle.

1:25:16So you will have to prompt it and you have to verify it. I see. But AI is middle-to-middle and then you take care of the ends. And you can't scare that everybody's saying. Yeah. So basically, depending on how much the middle part takes in any process, the part after prompting and before verifying, that now compresses to prompting and verifying. But it means, for example, as a coder, you have to specify it and then you have to review the code at the other end and then integrate it. And that's like the job of an engineering manager is actually higher skill than job in some ways of an engineer. And so I'm not saying it's not a productivity winner.

1:25:46It is. But it is amplified intelligence in many ways rather than fully agentic artificial intelligence. So the smarter you are, the smarter it is. But I will give one analogy. Ugar, you say something. I was saying that verify, prompt and verify loop, I think it's a shifting window where things you verify, prompt and verify enough times, will get auto-verified and auto-confused. So the question, I mean, in the last two and a half years of AI, since the chat shifting moment, I don't feel that's true yet. According, it's bidding to be true. I think the verifying part is still hard because, actually, I want to be more precise.

1:26:25On the front end, for images and for video and for front end code, that you can verify with a different thing, which is you verify with the GPUs from your eye for your brain. Or you can instantly see something is wrong because it's like two hands or it's got a messed up button or something like that. So that's where verifying is cheap. but on the backend for logic, for database code, for crypto codes, curie code, that is actually where verifying is expensive. Maybe we can turn it into something which is visual. And so you can visualize almost like, for example, an audio file, you can turn into a spectrogram and you can see if there's some giant spikes of amplitude and that looks broken.

1:27:05You can quickly see it with your eye, but you can't quickly listen with your ear, right? So it's possible we might be able to, for example, generate some visualization of the code so you can quickly visually verify it. That generation itself has to be deterministic code rather than probabilistic. But that's why I divide. I'd say the front end stuff, the images, videos, that's what AI is amazing for because the verifying is cheap. But the backend code verifying is expensive, which is a non-obvious division, I think. I tweet about that, but I don't think people have made exactly that division. I give another analogy, which is many businesses for a long time, until basically the 80s, they were doing their accounting with pieces of paper, not spreadsheets.

1:27:48So instruction and spreadsheet compressed that whole middle step of adding up all the numbers. They used to have a mailroom at companies rather than email. And then the instruction of email just basically tipped that completely away and it pushed it to other things. And as an example of running a clinical lab, because we started it in such a way that it actually had a database at the core, I could run a query and it could just give out all kinds of analytics and stuff like, you know, which variants were present with which other variants and which insurance companies and all this kind of stuff that it cost millions of dollars for a traditional clinical lab to try to find those analytics on some process.

1:28:25For example, bottleneck identification worked as follows. Because every station in the clinical lab, when you create a record in Postgres, you get a created at timestamp for free. And that meant that later when we wanted to improve turnaround time, we could take every station, station one through station, let's say 20. And these were in a complicated order because sometimes you have to revert back to a station. You could take a sample and you take its created at timestamp at all these stations and you could make a graph of the sample moving through stations. then every edge between them you have the time that it took and you can start making the histogram of like oh did this have a long delay at times right or does this have a high variance so you can make the edges wiggly if they're high variance you can make them red if they've got a long delay and so on and so forth and this fashion on a big monitor you can instantly visualize all the valets and you're like oh right there is the highest variance and slowest step and we need to add three more pico greens or something to paralyze that, right?

1:29:30So that kind of analytics you couldn't get. And so that's how I think of AI becoming the brain of many companies, right? Lots of processes that were previously as manual as a mailroom or previously as manual as counting at numbers. AI just takes care of that middle part. But then you still have to kind of do the end caps. Interesting. Yeah, so I'm just going to look at it. You've been thinking about this quite a lot. And I think of, again, you guys, that I'm not diving into AI, which is fantastic. you're familiar with formal theory, right? It's almost like... Like formal verification. Formal verification, right?

1:30:02So you almost can verify the output within the system itself so you know it's actually accurate. And I think it's an NP-hard problem because you don't know what other accuracies are there that you're not covering because you're not programmed to measure it. But I think of AI systems, because they're probabilistic, it is much harder to measure accuracy and harder you measure accuracy for a poem that you write as Shakespeare. You don't know if it's accurate or not. It's a feel, it's a why. But I think the next revolution of AI will happen when you are able to formally verify output and quality of output in a way that closes the loop on the prop that begins the production of the tokens that you see as photos, videos, or whatever.

1:30:51What does that mean? Most AI systems today are actually open loop systems. They're not closed loop systems. take for example K.ar fantastic company many of our companies use K.ar it takes inputs from a CRM Sixth Sense or a Zoom Info it'll do some research it can figure out this person there's a VC here each whatever gender whatever living here or there it'll write me a beautiful customized email to me it'll write a million copies for a million people and then pushes it into Upspot and Upspot does the programming and has the Yes, because the problem is now I don't want to read those emails. That's another problem to solve.

1:31:31There's AI to sort of fix that. But where does the loop get open? Is that now these emails have gone to our spot. Our spot puts a tracker in there. It puts a headline in there. No, I know, I know. But what I'm saying, though, is when AI increases the generation of stuff to such an extent, it will break that channel that was working. And I don't even read email anymore. And then I kind of don't already. Basically, someone's like, oh, I sent you, you know, I'll go and look or whatever. And so instead, you go to other channels that are high trust channels that people don't have the number to or something like that.

1:32:03Right. And that's where actually where your true messaging is. Now, I can't get in there. The whole point is not an AI. Right. Yeah. So so I think that there's. There's there's a window right now. Yes. Where generated stuff from me, I will overwhelm all kinds of missing channels. And there'll be a reaction against it, which I'm already seeing, where people are like... But the reaction is where the loop is close. Okay, go ahead. Because when you send this out to Upspawn and you say, I don't care, I will never open an email that looks by the other email, that information is not coming back to Clay today.

1:32:37It's staying in WhatsApp. Sure, sure, sure. Yes. They have put the trackers in. They don't know what kind of subject lines people open and what times when you send it, people open or do you have it in HTML or with NVIDIA embedded in it. And HubSpot knows. So, Klai does not know how to improve its AI, how to make it better so you open it tomorrow, the after one year later. That loop never gets close. And these are software loops that are easily... Fair, fair. That's true. So one part of that I do agree with, which is the closing of the loop in terms of open rates and so on. But the other part is just that systemically rates will trend down when as people are increasing the number of their AI, you know, like emails getting sent out.

1:33:17And I think you're going to start. So the complement to that is going to be crypto because you can charge cryptocurrency for emails. Yeah. Right. And we could do that. And I had a company. And I wanted to use email on a company. How is that? Well, so I want to do the V2 of that. Right. Basically, because I think tasking is going to become a big thing. But the reason for that is I think AI makes everything fake. Crypto makes it real again. Crypto sets up hard digital borders that AI can't cross. Like AI cannot fake the private keys for a cryptocurrency. AI cannot pose as something with a crypto identity.

1:33:56AI cannot manipulate on-chain data. For all of that, it's like a hard deal border that these AI agents can't bust. So that's what you can count on on the internet are blockchains that are unhackable. And then because it's like a hard surface, that actually becomes the verifiable internet, which is Web3, or the verifiable web. And then Web1 and Web2, Web1 is like the worldwide web. Web2 is like, let's say, the social web. Web 2 is becoming closed because X and other things need to stop AI scraping. And then Web 1 is becoming spam. And so you get spam or paywalled closed. And then you have Web 3 be the open web, the cryptographic signed web.

1:34:39And I think that's where we're going to go. What's the intersection of AI in Web 3 that you see, like a use case section? So there's a few. Well, there's some that we're working on. Right. But so the one that's out there in public that everybody knows is agents using crypto to pay for things because the crypto is currency of the Internet. Right. Since it works across borders and an agent can boot up an address, pay another agent who can receive it. Right. So just that alone is like clearly a big thing. Any agents can sign smart contracts with other agents. So any kind of automated trading, you know, or transactions, cryptocurrency will be the way that AI is, you know, going by.

1:35:18So an intelligent agent, which people talked about in the late 90s, you need both AI and crypto to be at scale. You need crypto so that everybody could accept the transactions across borders on the internet. You need AI agents to be able to go and swim the web and find those things. So that's one piece. A second piece is, and I think this is most interesting for me, is on-chain data as the thing that AI can't falsify. So like the verification step and so on.

1:35:47so that to me is what I call like the ledger of record. Like rather than the paper of record, just argument from assertion, you have cryptographic verification. And right now that only works for financial facts because financial facts are going on chain, but the blockchain is actually the crypto more generally is the number four market in the world. Do you know that? Like number one, NYC, number two, NASDAQ, number three, Shenzhen, number four, crypto, right? And it's quickly gained. It's going to become number one, like intranet capital markets will also be Indian capital markets. Like the way I look at this, and this is a five or 10 year thing, but it's not more than that.

1:36:21We now finally have the on-chain dollar, on-chain fiat. That's going to mean you also have on-chain, not just currency, but on-chain equity and on-chain fund interests. So very soon, every internet company will have an internet equity. For almost 10 years, this was pathologized because of the SEC and their jihad against crypto and so forth, Gensler and all these guys. and they'd all be like, is there a security? And so on, they'd like, you know. But the obvious point is you wanted a legal pathway to put a security on chain. Duh, right? And that's not dishonest or anything like that. It's like saying, don't record your securities in an Excel spreadsheet.

1:36:59You need to do it on a piece of paper. So you went from, you know, a piece of paper to a cap table, right? Why can't you go to a coin table? Obviously you should because that is kind of like, you know, the intermediate step is like Carta or AngelList, which are great. great things, right? So you go, it's almost like the ascent of man, you know? The caveman is like, actually the caveman doesn't have a company at all, right? Then you've got a piece of paper or, you know, then you've got a cap table and a spreadsheet. Then you've got Carta, right? Or Angel, which are great companies. And then you've got Bianchi and Cointable.

1:37:28The advantage of this is, you know, something I didn't actually realize. Arguably until crypto, international law, international trade did not exist. You know why I say that? So let's say you've got, you're an Indian company and there's a, there's like a Brazilian company that you want to acquire, right? How many people speak both Portuguese and Hindi? Not that many, right? Okay. Maybe speak English on both sides, but how many people have done Indian Brazilian acquisitions or Indian Paraguay or something like that? Not too many, right? So what people end up doing is they use a U.S. company as the adapter between the two, because U.S.

1:38:04did enough business with both of them. They use U.S. law. And so this way they, they, they make it work. So the U.S. actually turned out to be the central hub, they reduced it to a hub and spoke problem rather than a peer problem. The other country that increasingly you could not, one could, but Chinese people could use is China, where they can use that because that also does enough business with other countries that there's some bilateral thing. A Chinese company can buy this one and sell this one or something along those lines, right? But now with crypto, you have global rule of code. so eventually that just like when you came to india you had global mail right because your mail just worked right and you had global movies and you had um global media your twitter just worked everything else just worked you will have global money and global entity and global instance exactly because if you think about what is an entity first you have bitcoin then you have stable coins and so you have all the transactions on chain and then you have all the transactions on chain of a company Once you have all the transactions, you can generate financial statements.

1:39:06Once you can generate all the transactions, you can also have the cap table there. So the cap table plus all the transactions plus all the smart contracts, that is what a company is. So you have an on-chain entity and it actually lives on-chain and that floats above and you can port it between jurisdictions. Just like you can port everything else. the reason that's harder to do than all the other things we've done before is it is like a group of people it's like law it's money it's like high stakes stuff but we're getting there right and i think that's going to be one of the most important areas for uh indians but actually anybody who's not in the u.s or china to really invest in actually frankly lots of americans and chinese will benefit from as well but especially if you're not in a quote in a in a country that is contending for being the global number one.

1:39:54You want crypto because it protects your rights online. You know what I mean? Right. And it protects your rights abroad. It gives you like currency that can't be seized, gives you contracts that can't be unilaterally abrogated. You're like a first class citizen on the internet, even if you're a second class citizen in America. That's like a really deep point. Like, you know, people can mess with your H1B visa. They can make it hard for you to start a company or something like that in the US. They cannot make it hard for you to move your Bitcoin. Yeah, you were saying I think at some point in time, your wallet is your passport.

1:40:25That's right. Yeah, look at it. I think it's starting to really play out. What's exciting you in crypto today? I mean, we keep hearing about stables a lot. We'll see. So you know what? You know I launched USTC at Coinbase? Alright, well done. Yeah, you want to say something? We'll put that on screen or whatever.

1:40:44We're launching support for a new stablecoin called USD Coin. We call it a stablecoin because one USD coin can reliably and stably be exchanged for exactly one US dollar. So October 23rd, 2018, like we put in the 25 mil, we collateralized the consortium, we launched the, like that whole thing doesn't exist without old Balaji. Wasn't that like a massive bull period in 2018? Oh, that was a bear period. That was a bear period. 2017 was a bull period. This was extreme bear period. Yes, yes. This is what, this is what crypto had fallen. Bitcoin had fallen down to like 5 ,000 or something from 20 ,000.

1:41:18It was down 75%. But, you know, I'm an old weather. You know, I believed in this, you know, through multiple bubbles and cycles and so on and so forth. And so, yeah. And basically, the graph on that, that was October 23rd, 2018. And I'll show you the graph, actually. And was it called $17 ,000 or something that I increased? $17 ,000? I remember I was lost in... Yeah, $20 ,000. I've got the price graph It's a hard thing. It's like a tat. I don't get any tats, but I do have that as a tat in my brain. So here is like, so you used to see. Let me see if I can show the exact area.

1:41:58October 23rd, 2018, zero. Wow. Right. Is that cool? That's amazing. And now it's at$60 billion. $10 billion. Crazy. Crazy. Right. So mind going. Mind going. Right. That's only seven years ago and it's now like a few percent of transactions for the dollar globally. It'll become everything in a few years. So why don't we just wrap up lightning round and so forth. I'll just go through a bunch of different areas, give you thoughts on, just give you one word and you can be in a few.

1:42:30So let's see, we did a bit of it. So biotech, have you thought about that at all since your Google Cloud days? a little bit of it a little bit what is interesting to me in biotech now is is it a long answer you want a short answer give me the medium i've heard yeah okay so in the context of india what i didn't know is that if you think about imagining new molecules that we cure diseases to building chemistries for it to testing it and then deploying and launching and scaling it if you think about the thinking of new molecules as a head india has the body it doesn't have the legs yet the body is turning those uh ideas into chemistries trying it out and then doing animal tests and then first layer of human tests and then pushing it back and they push it back to a pfizer or a buddhina or whatever in the u.s and that they they cover legs they have the brains they tell an indian company what to build and they have five or ten cooking chemistries you should build and then go test out and then it's near, you know, second stage human trial or the approval stage, then it goes back to the US.

1:43:37They finish the process. And for cancer, it'll take five years. For something simple, it'll take a few months. And then they go prioritize, package, sell, manufacturing, sell. With LLMs, now, if you're starting to see it, there's a recent paper that came out of Stanford two months back, is that they tried out 75, 76 different modalities and figured out, you know, I don't know, combining efficiencies for each of those with certain diseases and all that. And then they came up with two chemistries in six weeks, what used to take two to three years. And then they actually had successful lab rat tests.

1:44:11So what's interesting to me right on biotech is that India did not have a head, but it had the body, it didn't have the legs. Now we have a head. So you could then build a fully integrated biotech company of India that does not yet know how to personalize and sell and package it. But what they would do is what used to be called mostly like contract CROs. Now you can take the CRO of that. You can design your own drugs, and as you put it, push it back to the US. This is similar to how China started as just like a low-end thing, and then eventually went up the value chain and actually the CEO around China.

1:44:54So correct. That's right. That's interesting. Okay, that's good. So media. So one of my theses is India becomes a media superpower, but in the following sense, not in the sense that Bollywood takes over the world, but rather that Indians are very good at content creation and become a larger and larger and larger percentage of content created online. Let me hear your thoughts on that. Yeah, they're starting to see it. We actually am back to a few companies that are essentially calling creators, young kids creating content and then publicizing it. Um, so India is so big to the native that, uh, many people are choosing to live to as a primary job.

1:45:31YouTuber is becoming, you know, a cultural phenomenon. Uh, I do think, like you said earlier, Indians are the best exports. So we, I'm trying to see Indians show up in movies and TV series and never have I ever and blah, blah, blah. And that was not normal five, 10 years back. Now we're popping up in big movies. so this familiarity with the culture and sustaining the face I think is quite interesting and now the content that is getting created in India actually can potentially get vital in the US. You're seeing it in gaming as well you know they are culturally relevant games about Mahabharata and Ramayana they are getting popular in India the question is can like the way Tanka is popular everywhere in the world and we play games with North Scott themes and so on can that ever go expand beyond India and become a popular theme in the US or in Japan or Korea.

1:46:22I mean, that could be interesting to see. Same with comic books. India has a really nice, nice comic culture. It's kind of died out over the last 10, 15 years, but it's picking up again. The F2-like company is coming up, and AI is flipping that wave, and then they're all very culturally Indian. I mean, can you find, can you export that to the world, right? Indian stories. You have a company called Pocket FM, and the guy essentially was traveling in Indian railways, and then Indian railways have, you know, these are all seven, eight, ten-hour journeys. You would have these Thilawas, you know, guys that sell pocketbooks, pocket pavillas.

1:46:54Spicy stories, romance, horror, love, adult fiction, endemic adult fiction. So he was reading and realized that Shiloh could convert this stuff to audiobooks. And he converted these to short form but long episodic audiobooks. And he just took off. It's a company called Pocket FM. And they make$70 million,$250 million revenue now 70 percent of it is in the u.s indian company building designing new short form novellas you know spicy stories and it's basically though at global quality it's not it's not like indian accents it's not indian only that's the key thing it's yes for for audio yep and now they're laying on e.i because you know initially you would have to have humans sitting in a podcasting room like that to capture their voice and tone.

1:47:43And those 40, 50 % gross margin takes a long while. There are contracts that you have to do. Over time, you just get to 90 % gross margin by just using AI to create different synthetic voices in any language. And then the interesting part of that is you can experiment in India because the population is large. And interestingly, if something spikes in India as a story, it actually spikes in Germany, except then you convert to Germany. You get to Germany. But it's the same seven or eight stories in the world that keep playing out. Love, loss, grief, ambition. So it's culturally similar enough. See, the interesting thing about that is I do feel China is a Galapagos Islands where the animals are different there.

1:48:24The ecosystem is different, right? For example, Meituan is like a huge, the equivalent of Groupon in the US became a much bigger company in China. So they had almost like a different evolutionary tree. Have you ever been to like Socotra? No. Skotra is like this island in the Middle East. Yemen, right? Yeah, near Yemen. Yeah. That has a totally different set of trees and the Balwab trees and all that. Yeah, exactly. To anywhere else in the world, right? And so it's like a different use. So China's like that where stories that take off from China or vice versa don't necessarily happen in the rest of the world and et cetera.

1:48:54It's relatively rare there's a story. But India's connected enough because it's English that it is correlated with, let's say, Germany. Yeah, I mean, it's a 95. Yeah, that's another thing that's happening in India. It's quite interesting. Okay. All right. Let me do two or three more. Space. ISRO has actually done surprisingly well. Low-cost launches are a thing. And so tell me, you know, have you thought about space? Actually, Adnicole, me and Sriram and several other people invest in that. You know that one, like a space. I don't know. We know it very well. Fantastic company, Adnicole Skyrode.

1:49:23We invested in a company called Texas Space to raise about$90,$95 million. You're also able to have high-to-spector satellites. Sorry, I didn't know if they were a competitor with all you color satellites. Oh, no, they're not. They might use Adnicole. They used to go launch because they don't do the launch stuff. They do the satellite and payload stuff. But yeah, very cutting edge. And we were the first institutional investors in any space company four or five years ago in 2019. And now I think we have hundreds of companies building space. And again, if you think about hundreds of companies, hundreds of companies building in space across launch vehicles, reusable launch vehicles.

1:49:54There's a company called Ethereal X that does the Ukraine. Incredible stuff. There is propulsion systems. There's drone station. There is telematics. there's satellite building. So this is a place that you, this is a sector you invest to fair up in there? Or at least where is? Yeah, we are very much, we are really, I don't think I could rattle off all of those different subsections like you just said. Yeah, so we, yeah, we have a pretty deep and sophisticated thesis around space tech because the ISRO talent base is so strong and then the young entrepreneurs coming out of the woodwork and join hands and then, you know, you know how much any government base is for?

1:50:25It's peanuts. So, you know, once they have done their duty, they've learned their lessons, they've learned their skills, they are very happy to go work by private company now. Interesting. And yes, you know, the ISRO, as an institution, is also very supportive. They are offering labs for free to test out, do turbo testing, vibration testing, space fitness testing for satellites. It is quite amazing to see. So, okay, that's interesting. I learned something. And I knew ISRO was doing well. And when I know that Hunter's coming, that's really, really cool. On drones. Drones. Yeah, it's a very interesting space.

1:51:00The company you should check out is called Airbound.com. They're building VTOL blended wing body. So it's a single blended body. There's no moving parts. And it takes VTOL. It takes off vertically and then gyroscopically navigates itself to a horizontal glide path with wings. And it can cover about 60 to 70 kilometers. This is a company that the ambassador and the preceded this raise from budget-tier investors in the U.S. large seed round. They're doing, I think, Dode is an interesting space in India because, you know, traffic is just such a pain in the ass in India. If you really want to offer things of different varieties and quick commerce space, you have to take the aerial route.

1:51:49The road routes are just too slow and too expensive. And regulation has been, you know, generally quite positive in India. What doesn't exist yet is a technology that can scalably, reliably go deliver things that are not 500 grams. You know, you've got to get to three to five kgs for it to cover 40, 50, 60 percent of the overall volume of deliveries you want to do to consumers. So 40, 50 percent of consumer deliveries are some three kgs. So this company right now is building a drone, one of the most advanced drones in the world that has an LD ratio almost matching the best aircraft in the world.

1:52:31And what he wants to do is to just build the best aircraft in the world and just use it for drone deliveries. So the goal is to build the best aircraft, the best lick and drag ratio, the best. You mean like aircraft aircraft or drone delivery aircraft? He's building the best aircraft that is shaped like a drone, the size of a drone for delivery. For delivery. He's got a custom bill for a custom of the output air rules. But the goal is to build the absolute best aircraft. And is drone delivery, I know drone delivery is actually live in China now. Yeah. Is it there in India? Yeah. He's doing pilots in India right now with a large healthcare system.

1:53:06What is happening in India that people don't offer tech-wise? India is not coming to a point where people are building real world-class, first-in-the-world object policies. The drone was an example of that. The satellite companies, rocket companies are examples of that. And there are many more that are coming. Cineconductors and stuff that we're looking at seems like it's very cutting edge. so there is there is absolutely intent and there's capital available and there's talent that we believe is going to land India in a place where we are to know what we're doing, to know innovation that if it works it can actually really reshape industries.

1:53:55That wasn't true 5-10 years ago and I would argue we still are a minuscule

1:54:04minuscule percentage of what i think the contribution to global ideas that needs to change and i would love for that to accelerate and i'm trying to do some things to make that much faster but there are small pockets a few pockets so there are very deep pockets of innovation that are absolutely mandatory but this company that we just talked about you're coming with people that are that are uh leaders and you know senior developers at tesla at spacex Google's wing which is let's try to build exactly this 10 years ago and couldn't annual and it was drawn team leaders with us and this one at you know a bunch of so this stuff is really cutting edge so I think there are more and more pockets of that with young first time founders this kid is in 19 now when the ambassador was 17 there are more people like that coming out of the woodwork I think there's quite people don't know this and I would love for that volume of people to be much more and back them more.

1:55:04And I would love for more people to know that there is stuff like this coming out of India. You're not a software automation. We have a back office anymore. We're not just a consumer, mass consumer internet and e-commerce country. All of that. And that's also deepening and broadening, but we're also this now. That's really awesome. I think, I don't think people, again, I think it's still underpriced. I think we, that, that story, like this kind of stuff is awesome. That's like world class, right? And I think the world class stuff, people kind of knew Indians individually were doing that kind of stuff, you know, but I don't think they realize how much of a, for lack of a term, bench isn't even the right term.

1:55:43How much in reserve there was for the Indian. Yeah. Okay. Awesome. Maybe with that by me. Awesome. Great. Cool.