The Vinod Khosla Podcast

Hey guys, welcome to the Superteam Podcast. Today's episode is very special. You know, Akshay, every couple of months we have one of those episodes where we're like, I can't believe this is happening. And today is one of those episodes. Why don't you tell the viewers why that is so?

Yeah, I mean, we've been up since 5 AM for this, so you know it's special. Welcome to the one and only legendary OG of the Indian entrepreneurship scene before any of us were born. Mr. Vinod Khosla, welcome, sir.

Thank you. Great to be here. And thanks for having me.

All the folks, young folks watching, you know, sir co-founded Sun Microsystems back in 1982 when he was 27 years old. And then of course went on to have a legendary venture capital career at Kleiner Perkins. And now is the founder of Khosla Ventures, which has produced a bunch of hits on its own over the last few years, including Oscar Health, Stripe, Opendoor, Anchorage, Instacart, DoorDash, and Square. And we couldn't be more excited to talk to you about your entrepreneurial career, your venture capital career, and kind of like what you're doing now, especially with that beautiful background you have going on, sir.

Thank you.

Feels like we've caught you on vacation, sir.

No, I'm actually at home.

You know, Akshay introduced you, you know, Vinod Khosla Ventures, but we were reading up about it and we saw that you introduced yourself as, you know, it mentions the term venture assistant. You prefer that to venture capitalist. Why is that so?

I always think of what I do and what I really, really enjoy is help entrepreneurs build bigger companies. So it's not just the capital because lots of people can give you capital and we've seen an abundance of capital. But most people who go on boards today of venture companies have, in my opinion, not earned the right to advise an entrepreneur because they've never been an entrepreneur themselves. So you earn that right. You don't get it by getting an MBA. You don't get it by joining a venture firm and then investing in some startup. And most of the time, they give really poor advice because they don't have empathy for the entrepreneur. Entrepreneurship is hard. The highs are high in entrepreneurship and the press writes about it, but the lows are very low. It's, it's, it can be pretty depressing at times being an entrepreneur. And unless you've gone through all these cycles, You're not going to do a good job of advising an entrepreneur. You have to understand their concept. It's not just a business. It's somebody's life. It's somebody's mission. You know, one of the words we don't even use in our partnership, I hate it when anybody uses it, we're investing in this deal. It's not a deal. It's somebody's life. And so, I take a very, very different view. Of entrepreneurship. And I think our job, what we like to do, is beyond capital. Most of the value I hope I add is in making a company a bigger, better company, a higher probability of success, being empathetic to the founder where they go through ups and downs. As you know, it's well known mental health issues among founders is pretty common. Common because it's such a highly variable, ambiguous, uh, and up-and-down kind of business. So, uh, I, I view my role as more mentor and assistant to entrepreneurs. I say I'm much more assistant to entrepreneurs in building their business than an investor in entrepreneurs.

You've been an entrepreneur, uh, yourself, and since the '80s and '90s, like, over time, you must have seen some of the more common problems that entrepreneurs come to you with. Um, have you seen any differences though, that, hey, through the '90s or 2000s or early 2000s, you know, a lot of folks would have a type of problem, but now the problems have differed? Like, have you seen any commonalities or differences over the last 3 decades?

One big thing, there's a lot more support for entrepreneurs. You can all read the same publications on the internet. There was no internet publications, so you didn't know what other entrepreneurs were going through. You didn't see those blogs, you didn't see those forums, those communities form. There were no Discord session to join. You know, the world was very different. But also the rest of the world was much less accepting of entrepreneurs. You know, today it's okay to be an entrepreneur and it's okay to fail. And then it doesn't mean the end of your career. In those days, people really— nobody respectful would join a startup, which was a real advantage for startups because you didn't want a lot of protocol and respect and big titles in startups. They generally hurt a startup. But in those days, it wasn't credible to do that. So I do think the environment is much more friendly to entrepreneurs. Much more supportive, and there's much more to read and learn from, from Reddit, from Discord, from every kind of community forum you can think of, or every Medium post you can read. And it's available worldwide. Back when I was in India, you had to take a bus to New Delhi to find an old magazine to read.

Yeah, that's how you started though.

It Yeah, then.

So you betrayed your current interests, by the way, by mentioning Discord. And we'd be curious what you're thinking of lately, because that's really the medium of communication and interaction in crypto. Is that something you've been spending a lot of time on?

Well, crypto is interesting, right? I divide the crypto world into two very distinct worlds. And most people in the crypto field don't think about it this way, but I have my own perspective. And anything you do in crypto has to align with your values. Uh, there's crypto for the crypto world, and then there's crypto for the real world, and those are very different things. In crypto for the crypto world, you really like the fact of anonymity, but in crypto for the real world, if you're doing banking You want to be regulatory compliant. You don't want people to shut you down because you're not regulatory compliant. You get to make a choice. And crypto, for the crypto world, has been a set of values. Bitcoin's a great store of value. I think of it as a better gold. That's what crypto is, except you can't wear it as jewelry, so it doesn't have that use. But I guess you can buy Bored Apes as a badge instead of jewelry. But it is a good store of value, and especially in certain parts of the world, it's very valuable to hold that than your Argentinian currency. So there is good use of crypto in the crypto world. There's tokens that are supposed to have value without intrinsic value drivers behind it. Now gold is like that. We've tended not to do things outside the regulatory regime for reputational reasons. So unlike if Balaji was here, he would sort of say, let's start our own country. I'm not trying to do that. But there is lots and lots of room for disruption in the real world. With crypto that are very, very interesting possibilities. You know, a good example would be we invested 4 years ago in, in Helium. Helium's building a real physical wireless IoT network using Helium tokens to incentivize people to join the network. So it's really creating network effects. And now there's pretty universal coverage in most major cities in the United States because of the Helium token. And that's a very good use of fiscal, uh, of crypto disrupting the real world. They have already announced they're working in building a 5G network the same way. Now those are really large, long-lasting businesses, and I don't think an AT&T could compete If somebody built a 5G network based on crypto tokens and incentivize the early participants, which is the nature of tokens. If you get in early, you benefit. And so, but you're also contributing early in building a network. I would love for somebody to say, how do I disrupt JP Morgan and Bank of America using the crypto approach? So we have an investment in a new company. It's not been announced. That would be a formal registered bank in the United States, but be very crypto-friendly and crypto-compliant, but with KYC/AML. And that prevents illegal activity in the network. One thing we haven't done is anything that leads principally to illegal activity, whether it's terrorism or sex trades or gun trades, or— we've just avoided that, or things that are strictly speculation. We've not bought tokens early and then sold them after hyping them up in an ICO. It's not something we'd like to do, but building real businesses that will be around 10 years later and taking the advantages of crypto, it can be frictionless. It's not in a lot of areas today, but it can be frictionless.

And so do you think that the— that this idea of bootstrapping a two-sided marketplace using tokens will just become more common. And it's just a better way to build a two-sided marketplace than, say, investing in performance ads or just paying Google and Facebook venture dollars to do that, especially as that orchard is getting overpicked. You know, I used to, back in the day, work in a junior role in a tech company where we did a lot of this.

Well, it's not a good thing for the big companies, but it is definitely a great thing for startups that don't have natural scale day one, and they want to achieve scale. If they produce a great product and get users to join because the joining early has benefits, then it's a really good way to attain critical mass. You know, take— there are many such applications. So the Noodle IoT network is another great example of a crypto-based network Uh, you know, the old style way of doing it never caught on, which was buying a Tile. If you know the company Tile, you relied on enough people to install the app on their iPhone so they could detect your Tile. Well, Noodle is a much better way to do Tile than Tile was. International money transfers is a really good application of crypto, of course, especially since there's so much less expensive, and frictionless. Sweatcoin is a great example if you want to talk about one. You're putting a community together of people who are interested in losing weight. It incents them to lose weight, and it gives an advertiser a forum to talk to them on. You know, Juno in banking, recycling businesses— nobody recycles old bottles, but crypto could be a great way. A favorite example that nobody's done, I hope somebody does, would be something like a loyalty program. Why do I have to tie it to airline miles? Why not have a universal loyalty program that's token-based and a fungible token that I can use in lots of different places? So there are lots of really good examples of very traditional businesses done much more efficiently in the crypto world. But there's— because of the speculative value and people trying to make a quick buck, there's not been as much focus. Now, I do think the next 5, 10 years we'll see a lot more company creation with real businesses over the long haul. I think as a store of value, as a gold equivalent in the crypto world, Bitcoin is probably going to be the currency people use.

So if I may ask, who introduced you to Bitcoin, or did you, did you just happen to stumble upon it yourself?

Well, I was on the board of Square when Square first jumped on the Bitcoin bandwagon. It was one of the first major companies to seamlessly go between your debit card and crypto. Like, literally, it was less than 15 seconds to purchase your first crypto. If you had a Square Cash wallet. And it was, as you might imagine, a big debate back then.

I was hoping you'd say you know Satoshi personally.

I wish.

So what, if anything, did you get wrong about crypto and that you have changed your mind about, say, compared to 5 years ago?

Well, I definitely didn't— estimate the rate at which it would be adopted. I think it got adapted faster than I would have thought. And frankly, uh, the reason for that is I figured crypto would be used to build real businesses that take time to build. And so I estimated a long-term rate of adoption that was different based on building real businesses with crypto. Which I still think is the larger opportunity long term, to disrupt real life with a much better system. Just like e-commerce was a much better system than retail, crypto can create a lot of businesses. Uh, so I think I got that very wrong, and the rate of adoption was higher mostly because the speculation piece took off faster. And look, there were countries where it really made a lot of sense. In Argentina, in China, it made a lot of sense to put your money in crypto, not in the local currency. And, and that was probably something I got very wrong. But I've never regretted the fact that I— we didn't get into the speculation business. I don't want SEC investigations because I was part of an ICO. No, we looked at the Telegram ICO long, long time ago, very carefully, when every other venture fund in Silicon Valley was looking at it and participated. We were one of the few funds that said too much reputational risk, too much risk with the SEC, and we want to comply.

So, sir, do you think speculation is, you know, when I think the— this reminds us at least a lot of the media business, right, where When creators first came onto YouTube, they were making content that no rationally thinking producer would let be on TV. And so they would come out and film themselves having breakfast or do vlogs or just content that would otherwise be thought of as inane or silly. But the fact that a peer-to-peer media network emerged where anybody could publish anything meant that quality would be both simultaneously much higher and much lower. And so do you think it's the similar thing where, you know, nobody, no rationally thinking investor would allow, say, a picture of a dog to be financialized into an asset that's worth $40 billion. But for every dog, dog coin that we get, we also get Bitcoin. And so do you think that's like a necessary evil that we have to endure, or do you feel like regulations are going to change that?

Look, this depends on your value system. Again, the question I ask, you know, generating content for YouTube or Twitter, that, that market already existed. You know, TikTok's built entirely on user contributions, right? Uh, so there's nothing wrong with that. When you get into the speculation business, you have to ask the question If things go wrong, who gets hurt? If the wrong people get hurt because they followed a trend, then one has to be careful, and that's why regulation exists. It's not for the average rich investor. It's for the simple people who get caught up in it and then lose their life savings. If you go on Twitter and look at the number of people who said they lost their life savings. They had $300,000 savings over the last 15 years, they put it all in crypto and they've lost most of it because the currency collapsed. That's very sad and harmful and not something I want to participate in. In a video creation example, if, if things go wrong Yeah, somebody doesn't get to monetize their video, but it's not their whole life.

Yeah, so that's, that's one of the reasons, by the way, on this podcast we'd never discuss prices or tokens. And, uh, I'll use this opportunity to plug the job board. That's the only thing that we do, which is, uh, people should just, if they want to get exposure to the space, find a way to do something meaningful and find a job. And so we'll also link to Helium and their job board.

But, you know, look, creating networks is real value-add. There's plenty of real value-add businesses to be built. Could somebody create an e-commerce Amazon competitor using crypto? I think they can. Nobody has attempted it yet, but it can be done. And you know, you look at something like a Sweatcoin, it's a niche approach to an e-commerce site. Whether it's e-commerce and advertising, Amazon is both. That's really what it is. But could you create an Amazon competitor? You know, I think you can. Could you create a competitor to home internet service? Absolutely. It's one of my favorite projects I'd like to take on. If you remember the beginning of cable in India, it wasn't that long ago and people just, individual entrepreneurs just strung cable in the area and had a video hand and provided cable. The crypto version of that is very possible today over the air because everybody already has the physical internet connection. So lots of new things are possible. You know, gaming is another area crypto makes a lot of sense, and there's areas in which DAOs make a lot of sense. I was talking today to the CEO of a major game network And they said, we do community governance, but I'd really like to pass it on to the community, not have my corporate parent do that, make those rules. So a DAO, a governance DAO would be perfect in that example.

Sir, I may indulge one last question on this. If you were to turn Khosla Ventures into a functional DAO, knowing what you do, now about how effective decision-making works and leadership, how would that, would you, how would you structure that and how do you think about that?

Well, so here's the question people fail to ask often in the crypto world. What are you trying to do? So one of the fundamental things was where do you need distributed trust, right? That was the original. You don't have to trust anybody. It's decentralized. Well, for most things you don't need distributed trust. But if you have, say, a supply chain, a global supply chain, I do think you need distributed trust. You can't trust where a drug came from and where it was adulterated or people had, you know, diluted the drug or whatever. So there are situations, but in most situations you don't need distributed trust. Do you need distributed governance? Absolutely, if you want to establish trust with the community, because we've seen so many violations of trust when it's centralized with a corporation. That's another good example. So, or, or take something like network effects. Creating network effects is very hard for a startup. It's a great use of tokens and tokenization. So I, I would say you have to say, what are you trying to achieve? And if I were to turn Vinod Khosla into a DAO, what would it be? I'd probably— from a scout network, if I had a scouting network and turned that into a DAO, I could see value in that. But I wouldn't sort of just do it for the sake of doing a DAO.

Yeah, I mean, eventually you could have a scout network of thousands of people who have followed your career, or— you know, who are interested. That's right.

Or if you wanted to create a pay-it-forward network, that'd be one. Or a loyalty network is another one.

And maybe if a scout earns a certain number of tokens because they've referred some great deals, they can be deemed partner and things like this within the community.

That's a good example.

Yeah.

Let's move on from crypto to AI, because I've seen a bunch of your talks with Sam. I know you're really interested in it. Have you thought about, you know, what are some of the applications of AI that you're excited about and anything that you think would be applicable to India especially?

You know, I'm very interested in medicine. My son has a startup trying to build an AI primary care physician. I want primary care physician available to every resident of the world, every Indian, every African for near free costs. That only happens if you can approve an app as a fine primary care physician that's equivalent to a primary care physician in performance. Even human physicians make errors, and it should be 24/7. So if you wake up with a problem at 3:00 AM, you shouldn't have to rush to an emergency center. You should really just be able to consult a physician. AliveCor has done that with cardiology. With a simple device that fits in your wallet, you can get ECGs at home and diagnosis at home without invoking a physician. That's a pretty great example of AI. So all of AI, we are doing a lot in, in medicine, but nothing more valuable than, say, Having a primary care physician available 24/7 for free for every person on the planet. I also think AI can do a print— AI tutor for every kid on the planet. I wrote about both these things 10 years ago before AI was what it is. And my wife has a nonprofit ck12.org It's a nonprofit open source where we are trying to build now an AI tutor so every kid can have personal tutoring. If you look at rural India, 40% of teachers don't show up for school any given day because they have a second job. You know, maybe they have to harvest. Those numbers are true of rural India. Well, those kids don't get a fair education. And so That should be a nonprofit free service. AI has lots of applications. On the content side, amazing applications. If you Google AI imagining Lord of the Rings, there's a Medium blog post. This person imagined what GPT-3 would draw if it read the book Lord of the Rings and didn't have any visuals. The drawings it did of the characters are almost like the movie, characters. It's stunning how similar they are. It blew my mind.

Wow, this is amazing. I'm just looking through this, this is like giving me goosebumps. This is pretty incredible.

So all the creativity of the best Hollywood producers can sit in the AI. Uh, we have a small application on Roblox, uh, called Splash. It's AI composing music composing the instruments, playing the instruments, singing the songs. AI does end-to-end music generation. It is mind-blowing, and it's caught on. It's the biggest music game on Roblox and strictly an AI play. Makes every 13-year-old girl feel like they're a great musician playing music. My point is, this content is being created by AI, whether it's a picture. We've seen movies or videos being created by an AI for marketing purposes. So I saw a real estate agent able to send out 2,000 videos, which they could never afford to produce, on one house being sold because the AI generated all the videos. And the Real estate agents just entered 1 or 2 line prompts about the message they want to customize to a user. You were looking for X, or maybe even just give a list. Tanmay, you were looking for a house. Here's one that fits your requirement. This is mind-blowing stuff. So AI will go a long ways. AI can machine machine tools. It can draw architectural drawings. All of that is happening now. It's such a rich, vibrant space that I'm very, very excited. And if you can generate this much content, then maybe you can create network effects with the crypto world and be in a whole different sort of ecosystem. These technologies compound each other.

Have you, have you thought about speaking to Sam and getting them to build something for Khosla Ventures where AI could determine if they should invest in the startup or no?

So we are the, we are one of the, we are the first investor and probably the only venture investor in OpenAI. So we are large investors and I talk to Sam regularly.

Can you, can you get him to build something where the AI decides if you should make, do a deal or if you should fund a startup or no? Or do you think we are far away from that?

Look, it can be done. You know, Look, the ecosystem of startups is becoming large enough. You need a large enough dataset to train on. I think it hasn't been large enough, but it's quickly becoming large enough. But there's too many specifics that are not being captured in the data. Like what's the personality of the founder? What's the risk profile of the founder? What kind of team quality did they assemble? But you know, 10 years later, you can see how successful one startup's team was in other things in life and judge how, how good quality the team was 10 years ago. For example, like that data generation has to be very rich. AI does well when it's high-dimensional models.

Hmm. You need something that can map like market data and then also map data from the founder's childhood and then make like a prediction of where the startup could end up?

Well, there's lots and lots of really high-value things AI can do. Medicine being a great example. I don't love AI used in marketing because you're hacking people's brains mostly. You know, you're still trying to sell more things to more people, not something that I feel great about, um, but that means it's probably going to be one of the first use cases, unfortunately. But let me be clear, now it's feasible to build an AI advisor for every person, sort of their network proxy that can filter everything and say, don't buy this, or this is a better product, or so The anti-marketing tool for individual consumers would be a great AI project. I'm surprised nobody is doing it. I've asked a number of people to do it. So I'm announcing here, hopefully one of your audience decides to do it.

I mean, in some ways, I guess tokens are the opposite of marketing or advertising in that the payments are directly to the users. And so there may be an AI crypto project in the works here, sir.

Will, I just want to mention that about 3 million jobs were threatened in the last 5 minutes. Just want to point that out.

It's personal, sir. Tanmay is a YouTuber in one of his jobs and he has 4 million subscribers and you're now threatening that part of the job.

The first question to ask is how those people jump on the AI bandwagon. And really leverage content generation instead of trying to fight it.

Also in India, we're shipping like half a million doctors every year. We're already working on every person should have their own primary caregiver.

Uh, well, but that's not enough doctors for a billion people.

It's still not, it's still not enough doctors. That's true.

And we will never have enough doctors and they will never be low enough cost where everybody can use them every day.

And if we have too many doctors, I guess everyone can get a second opinion then, because—

Yes, I, I look, I think there'll always be value for expertise, but expertise can be cloned in AI systems generally. Take cancer. Very, very few people in India, definitely in rural area, will ever see an oncologist if they get cancer. It's just not available. That expertise should be in an AI app so that we can see— provide that service to anybody.

Sir, do you think we're close to figuring out how to live forever?

No, I don't think we are. Um, I think the first goal is not live forever. There's a lot of hype around anti-aging, and nobody's discovered if there is an inbuilt aging limit. So, half the people in aging would say about 120 years is maximum human lifespan. But what is very practical to do, and a lot of people working on, is How you not— don't worry about extending lifespan, expand— extend healthspan. So at 70, somebody should feel like they do a 40-year-old. They don't get, you know, if you look at Alzheimer's, if you look at diabetes, you look at cardiac disease, they are all diseases of aging. The probability of the disease goes up as you age. And if you can make it where people don't get those diseases, so the lifespan they have is all healthy, that's the first goal of anti-aging therapies. It's to delay the onset of chronic diseases. Chronic diseases, the most expensive diseases in healthcare. They also afflict many more people than individual things. And solving— avoiding those diseases by changing the probability is very, very possible in the next 10, 15 years, in my view. And that's a very good first goal to shoot for instead of this dream of Larry Ellison wants to live forever.

Is there— is there— are there any things that you follow that have been that have held you in good stead that you would recommend to folks who are especially entrepreneurs, high-performance individuals, you know, who are career-focused?

Well, there's a lot of literature on staying healthy longer and avoiding disease. So, as you know, with my Indian genetics and my family was very prone to diabetes, my doctor told me when 27 years ago, after I— before I turned 40, I'd been on insulin within 2 or 3 years. I'm still not on insulin at age 67, and my A1C, which is a measure of diabetes, is lower than it was at age 40. I've been able to reverse it through management, aggressive management. And I think you can do that. Anybody can do that. These things are available. There's lots of literature and I highly recommend people go on the web. I don't want to give any medical advice to anybody. That'd be illegal. But lots of options available.

What you said sets up for the question we had here, which is first, do no harm is bad policy. Medicine does not teach doctors risk-reward tradeoffs mathematically. Rapamycin is a no-brainer beyond a certain age based on the health status of the patient. Can we double-click on that and talk more about—

So, rapamycin is a drug that's approved for organ transplants to suppress the immune system, right, at a particular dosage. At very low dosage, rapamycin does a few things. One of the causes of aging— so, there's a couple of mechanisms of aging: autophagy, senescence, couple of things like that. Rapamycin addresses some of those things at low dosage. But we don't have 30 years of history of somebody taking it for 30 years. So, somebody at 50, would I recommend rapamycin? Not at all. Somebody in their 70s, would I recommend rapamycin? It's an absolute no-brainer. In between, it depends on your risk profile. And your family health status. Like, if your father got Parkinson's, I'd take it at 60. If you didn't and there's no family history, then I'd treat it differently. So this is a probability of harm and versus a probability of gain kind of equation. It's an expected value calculation for those of you who are more mathematically inclined. Now, statins in your 30s, especially with Indian genetics, makes a lot of sense. One, it's cheap, so it doesn't cost very much, especially since the genetic version, and you can even take it as supplements. Red rice yeast is a statin, and you can take it as a supplement. You don't even need a prescription if your cholesterol is high or LDL is high. Should you start taking it in your 30s? Absolutely. Because it's been tested on over a billion people over a long period of time, many decades. Would I say the same about rapamycin? No. If I was 30 or 40, I'd wait for 20 years of results before I took rapamycin. But if I'm 75 and dealing with Alzheimer's, I'd be taking rapamycin aggressively. So, that's sort of mathematically how you compute. Doctors don't do that in drug approval processes. So, you know, it's funny to me, every doctor, at least in the U.S., takes the Hippocratic Oath. I think it's true of doctors everywhere in the world. First, do no harm. And that's just mathematically incorrect. If you take a risk and it has Take a new drug. If it kills 10 people but might save 1,000 lives per year, you should approve the drug. Today's regimen of first do no harm means the drug will never get approved because it killed 10 people. And societally, it's much better for society that we can save 1,000 lives or 990 lives. It's bad at the individual level, but you don't know which 10. So that's a good example of this opinion. It's a very scientific opinion. It's first principles thinking.

Is there— so have you come upon a framework or a playbook for health tech startups to deal with regulators and big pharma or policymakers in the government? Is there a playbook that you tell founders to follow?

You know, look, every country is different. And you have to be regulatory compliant in every country if you're trying to build a long-term company. Short-term, you can get away by bypassing regulation till somebody catches you or stops you or something like that. But long-term, you can't. And the more success you have, the more regulatory scrutiny you get. So if you're building for the long-term, you want to be within that. Within that, there's lots of ways to be creative.

Right?

So, you can do things like, for example, my son's primary care company has an AI doing diagnosis. Then they have Indian physicians in India who are not licensed to practice in the US. Do some of the verification, but in the end, a US doctor, prac— licensed to practice in the state they're practicing in, has to say this is the diagnosis and this is the prescription. But their workload is 80% lower or 90% lower. So huge cost savings. That's an example of complying with all the regulation, having human oversight, especially in critical things. AI is very important, but early AI will always be error-prone. So it has to be— have human oversight till it gets good enough or better than doctors.

AI being better than human doctors, is an— do you believe that's an inevitability?

Absolutely. You know, I wrote about this 10 years ago in 20— in 2012 in TechCrunch, I wrote a blog called "Do We Need Doctors? I Think." Right? It was soundly criticized and of course huge amounts of criticism and public, like, deriding it. Last year, the Journal of the American Medical Association said, in reading an image, MRI, X-ray, CT scan, they will no longer accept any publications that say an AI is superior to human beings because it's a well-established fact. They stopped even accepting articles. Now that's in image-based stuff, but that'll happen in other areas also.

So you had the last laugh, sir.

I did. Look, what entrepreneurs do is they have to imagine a future that nobody believes. If everybody believes it, everybody would be doing it. So you have to imagine a future that you believe, be passionate about that mission, then try and make it happen. Now, there was no way General Motors or Volkswagen could have created an electric car. They tried before Tesla, right? So when I look back, and this is a stunning fact, in 40 years of innovation, there is not one example, not one in 40 years, of a large innovation that came from a large institution. Like, there's no way GE, General Electric, or Siemens can innovate in energy. No way. It's not possible. Every innovation has come from a startup. Car rental companies or the car companies couldn't do Uber. You know, Walmart and the big retailers couldn't do what Amazon did. You know, media. It wasn't CBS or Fox News that did media innovation. It was Twitter and Facebook and YouTube. People who didn't even know they were in the media business. So no matter where I look, if it's a large innovation, big companies can do incremental innovation like Intel going from one kind of semiconductor process to the next generation because it's obvious, uh, and it's low risk. But large risk, radical changes on business models— there is not one example of 40 years that I've been doing innovation where a incumbent or a large player did it. So in fact, most of the examples I gave you, right, the people didn't even know the business they were entering. Elon Musk knew nothing about cars, and to me that was a huge advantage because it let him think from first principles and say, how should a car be built? He just didn't assume away what everything that the auto experts said couldn't be done. Same thing in space, whether it was— we invest in Rocket Labs, he did SpaceX. It wasn't Boeing or Lockheed or Airbus that did space launches.

How does an Airbnb prevent itself from becoming a Hilton, like, or is that inevitable?

Well, I, I find founder-led companies do better than non-founder-led companies. For a long time, right? You know, AWS didn't come from a computer company. It came from a retailer. If I said to you in 2000 that the most important computing company will be a retailer, you'd laugh at me, right? But it was founder-led and he thought from first principles, why not? IBM didn't do that. Digital Equipment Corporation, or DEC, didn't do that. Which computer companies survived those transitions? So it is founder-led companies, it's founder vision and founder's ability to think outside the box. I mean, take crypto. Very, very unlikely the large innovations will come from a JP Morgan or a Goldman Sachs or bank of a bank, traditional bank. Can you see SBI doing a crypto innovation? It won't happen. So it depends on your audience to bring crypto to disrupt the real world. And that's the big opportunity for your audience.

I can't believe Mr. Vinod Khosla said the sentence, can you see SBI launching its own token? SBI needs to figure out how to not, you know, take a 2-hour lunch break first. I think that's the first innovation they need to advance.

So your website says you prefer brutal honesty or hypocritical correctness. I watched your talk in Stanford, which we will link to, as, uh, and which is that, you know, brutal honesty is one of the luxuries that you indulge in as from having some financial success. And you can say some things that other people who may have to risk their jobs to say will not. So, can you talk a little bit more about that? And what are the 2, 3 brutal truths that you would say today about the tech industry that nobody else would?

What's important is take the VC business. Most VCs are nice to their entrepreneurs. That's a little bit like having kids and saying yes to everything they ask you, right? They will be completely spoiled and it won't help them. Right? What you really want is measured— them knowing you care about their success and giving them honest advice. So, um, it's very, very important that you give honest advice to people. Many years ago, I was on the board of a company, and there were 3 important VCs. Nobody thought the business would succeed, but nobody told the company. Right, they kept saying polite things, and probably 40 engineers wasted 4 years of their life because nobody would give them honest feedback. And when I did, uh, the founder didn't pay attention because the other VCs were saying nice things to him. So I think it's really important you give honest feedback so people can do something about it. You know, there's a very good book, since book recommendations are always part of podcasts, Willful Blindness by a woman called Margaret Heffernan out of the UK. And she talks about lots of people know lots of things, whether it's in a startup or in other situations, but people hate to go against the group thinking. So you end up with groupthink as opposed to people If they have an inkling something's wrong, they should speak up. And especially in startups, if your business model is wrong, one of your executives knows it, but they're afraid to speak up or disagree with the whole group. So, Willful Blindness is a book I recommend on how to encourage a culture where people speak up and give you the news you may not want to hear. And that's where brutal honesty helps because an entrepreneur can do something about it. As opposed to politeness, and you discover the problem too late.

So I wanted to ask you something for the Indian startup ecosystem. You know, we see Indian CEOs in, in the Valley fairly regularly now, uh, yet we haven't seen any Indian companies really dominate at a global scale. What do you think they can do to really make it huge globally?

You know, the first, first thing is to set a large ambition. And what I found missing is global ambitions when you start. If you look on Twitter, I like to say there's a huge difference between a $0 million company or $0 billion company. $0 million in revenue is the same as $0 billion in revenue is the same as zero revenue. But the quality of the thinking and what you're thinking about is very different between those companies. And I think people seldom think about finding the co-founders to go build a $0 billion company. If you just find your friends, you may not build the biggest company. You should say, what will help me avoid all the risks or ascend sort of the biggest mountains? You build a team for that, for example. How you build a team is very different between a zero million revenue company and a zero billion revenue company. That's just one example. You know, enterprise SaaS companies in India are starting to think globally. So things like Freshworks and things like, there's examples of it starting to happen. More people need to think globally from the very beginning. Even if they only enter the Indian market initially. The analogy I like to give is if you're climbing Mount Everest, you have to get to base camp first, but you pick your base camp so it can lead to Camp 1, which can lead to Camp 2, which leads to Camp 3, which leads to the Everest summit. It's a carefully planned path early. If you're just trying to get some revenue, you might reach a base camp that doesn't lead to further ascent. And this is a mistake many boards make also. Uh, Everest is clearly planned. Even the first place you go to, base camp at low elevation, is strategic for where you want to get to long term. And that kind of thing is important to build globally important companies. It's very early.

So you said you prepared a monthly report on how you spend your time and to ensure that you're budgeting time to the things that you love the most and that are important to you. Is there a sample section of that you could share with us that would help entrepreneurs?

I haven't done it for 20 years, but when my kids were young, I definitely did it every month to make sure I spent, you know, I had an important rule. I had to have 25 dinners with my kids a month. And that meant few business dinners, few out-of-town flights, or mostly take red eyes when, when I had to. So I used to do that when my kids were young, and frankly, I got so good at managing my time very precisely and quantitatively that I don't need to do it anymore. But I still every weekend stare at my calendar and Monday morning come in with cancel these meetings, create these slots and things like that.

I mean, I'm glad we made it.

On that note of managing your time, sir, thank you so much for giving us some of your time. Really appreciate it. And we hope everybody who's listening took something great away with it. We're going to link to the job board for Helium in the description and some of the other companies that sir mentioned. Do check them out. And thank you so much, sir.

Great, thank you.