Vinod Khosla sits with Sebastian Thrun, Chairman and Co-founder of Udacity, for a wide ranging discussion on the state of Silicon Valley, the future of energy and sustainability, and the role AI will play in the healthcare, education, legal sectors.
Sebastian Thrun:
Hello to our Udacity Thought Leader Series. My name is Sebastian. Today's a phenomenally great day for me because my really good friend and mentor, Vinod Khosla, is with me. Vinod, as you might know, grew up in India a few years ago, moved to Carnegie Mellon I just learned, as a grad student, got a Masters to get into Stanford. And then was one of the very early Entrepreneurs in the Valley working on Sun Microsystems as a founder, joined later on Kleiner as John Doerr trusted partner in Venture Capital, and now runs his own firm Khosla Ventures, with massive investments in energy and education in many other fields in the biomedical systems, medicine. I would say if you rank the people who know something about the Valley in the world, he would be among the top three people. Absolutely, yes. He's seen the world in so many ways and he's so wise and he's also always controversial, which I always like, he's a straight thinker, logical thinker who doesn't believe in conventional wisdom, if conventional wisdom is wrong. Welcome Vinod.
Vinod Khosla:
Well, it's exciting to be here.
Sebastian Thrun:
You've seen the Valley for such a long time all the way from Sun Microsystems, which was of course acquired by Oracle and from the Venture side, is the Valley changing?
Vinod Khosla:
Well, the level of activity is clearly changing and what is changing dramatically is till about 10 years ago from the big thing of the Valley in the early '70s the Valley was about technology affecting technology businesses. I think over the last decade, what we've seen is this radical transformation where the Valley technology affects every industry, not just technology industries, so we could sort of dink around in our narrow area of enterprise or mobile or software, but it wasn't about transportation. It wasn't about housing, it wasn't about media. Now the Valley is about every industry and quite surprisingly to all the Valley nerds like myself, it has become much more of a global issue than it ever was, because it's truly affecting society in a very large way, and we can come back to whether that's likely to continue in the next 10 years. I think it'll be even more explosive the next 10 or 20 years.
Sebastian Thrun:
You've heavily invested in medical technology and human health services. Give us your vision of what's going to happen there.
Vinod Khosla:
Well, one thing is very, very clear. What's not clear to start with is timelines, but what's very clear is any notion of expertise in any area and definitely in medicine will be embodied in AI systems. And so the notion of humans having expertise that systems don't, is just a quaint note. It'll look quaint in 40 years.
Sebastian Thrun:
But it's more than that, right? Because you can also now, bring this expertise to places where it isn't today.
Vinod Khosla:
Okay. So let me actually go there. About a year ago I published a document. I was 63, I'd been thinking since I was 60 about what I want to work on for the next 20 years. And I wrote this document because I wanted to be clear for myself what I wanted to work on. And it's titled, Reinventing Societal Infrastructure With Technology. The subtitle is: A Call for Entrepreneurs, because no big company in my view can reinvent infrastructure, only startups, entrepreneurs, the Valley can.
And I found very surprisingly, there was no large area of GDP, while sitting at my ranch, I couldn't come up with a technological solution which were multipliers from 100% to 1000% on resources and that's really, really exciting. Back to medicine. And really I framed this document which is about 50 pages long in the context of 700 million people on the planet have a rich lifestyle. It's rich in energy, it's rich in education, it's rich in healthcare, it's rich in living space, it's rich in transportation. We have nice fancy cars, 7 billion people want it, can we do a 10x increase in provision of goods and services to the 7 billion people to bring them up to the level of the top 10% without destroying the planet? And I was shocked how easy it was to imagine.
In healthcare for example, it was very, very clear you couldn't start enough medical schools in India to train enough students. You couldn't hire the professors to start enough medical school. Even with an infinite budget, you couldn't bring the doctor patient ratio conventionally to where it is in the U.S. today, which is not adequate. So if you sort of said, “well, is there another way?” an AI doctor would absolutely scale this. And so I do think a 24/7 personal Physician for every person on the planet at close to zero cost per month is feasible. And so I've got my son working on that now.
But it's also true, it's silly to not have an AI Oncologist. It's such a deterministic field.
Sebastian Thrun:
Yeah, that's very true.
Vinod Khosla:
And in all this, we are doing this in cardiology, we are looking at that. We are looking at it in mental health. There's hardly a specialty that isn't subject to this kind of package, that expertise into AI systems, which will then learn across all the patients in the rate of learning will be faster than for human experts.
Sebastian Thrun:
When you say these things, many people will hear you say, jobs will go away. Sam Altman's building OpenAI and he says it’s the last company ever built because once he succeeds, it's going to be all the other companies for him.
Vinod Khosla:
Yeah.
Sebastian Thrun:
What's your take on that one?
Vinod Khosla:
Well, I am one of the large investors in OpenAI. So there's a reason we invested. I do believe AI is powerful. I would say first, one caveat, when I talk about expertise, it doesn't take away the human element of care, but I'd rather have a great nurse than a great physician. And funnily, I recommended to the Dean of Harvard Medical School about three or four years ago that they stopped selecting for IQ, because you have to be so smart to get into Harvard Medical School, you have no EQ. And I suggested he look at the USC film school as a criteria for admission into the medical school at Harvard and the more empathetic-
Sebastian Thrun:
Good looking
Vinod Khosla:
And more empathetic and more mirror neurons. But I do believe we will get to a point looking at traditional economic metrics, GDP growth, productivity growth. We will do very, very well when these systems kick in. And I don't believe they've started to kick in. We will have great abundance. The only thing we will have unless we do something about it will be economic disparity. And in 2004 I wrote a piece in Forbes on great GDP growth. AI will drive great GDP growth, great productivity growth, great economic metrics, and great abundance with increasing disparity. So I do think our social problem will be disparity, but we will have way more resources than we can imagine today to address disparity. So even in 2014 I proposed that some form of universal basic income will be essential. I still subscribe to that view. Bill Gates has proposed a robot tax. So the issue of more goods and services for an economy is separate from how we distribute it.
And I do agree the system is unfair and biased towards the haves and not the have nots. Which is why, by the way, I think for those of you listening out there, our primary care doctor and oncologist and cardiologist for every single person will increase that. I also, one of my favorite projects is to do an AI lawyer so every person can have a personal lawyer to enforce their rights. One of the sad facts for most people who need to file for bankruptcy today in the U.S., is they don't have enough money to hire a lawyer to file for bankruptcy and lawyers don't let you bypass them.
Sebastian Thrun:
I'm looking forward to the day where your judge and your lawyer, both sides is an AI system, so you just cable them together and the verdict comes out.
Vinod Khosla:
I've actually proposed that in most arbitration, companies and employees and others agree to AI lawyers, AI judges. Where an AI knows when to triage an issue up to a human judge for now when the issue is ambiguous. It would save a lot of money and costs.
Sebastian Thrun:
It would draw a crazy line that an AI system could judge over you and fine you as a human being. There would certainly be a step further to give up to the machines.
Vinod Khosla:
Well, great jumps in technology, lead to great jumps in uncertainty. And we just have to realize we will have to deal with them, and if we award them we will give up a lot of upside and benefit for humanity. The better approach is to recognize that these things will happen, that we will be surprised with unintended consequences that we have to be able to address. I am not scared of AI as a sentient being. I do think we can harness it for mankind's purpose.
Sebastian Thrun:
I'm with you. You have invested in energy. I remember this, when I was just entering the Valley as a junior professor, there was incredible focus on energy in many funds. What happened to it?
Vinod Khosla:
Well, so energy's interesting. It's timelines are very different, they do take a long time.
Sebastian Thrun:
Like decades.
Vinod Khosla:
Right. We invested very early so we were one of the first funds. We did not chase it in the later stages, which many funds did, and in the Venture business you rely on other investors to come in and keep funding a company you started funding.
Sebastian Thrun:
I know.
Vinod Khosla:
It's this pyramid, almost pyramid like scheme, but everybody along the way gets benefit. I think whenever hype cycles happen that investment valuations go through the roof and then when they come crashing down the funding dries up.
Sebastian Thrun:
Let me ask you, are we in a hype cycle right now with self driving cars? There's so many companies now and valuations reaching multiple billions before there's even a product at market.
Vinod Khosla:
Yeah. I do believe the market is large enough. It's hard for me to judge whether we are in the hype cycle or not, but we can come back to that. Finishing on energy. We actually decided when the market change to triage fairly aggressively on which companies to support and which companies not to. So in our portfolio from them. One, the fund has done much better than most people think, our energy fund double digit returns, which actually is not bad, it'd be top of the class in energy, but you take large transformative technologies. I think some of those will result in that. Tesla is a great example, it transformed the auto industry. Quantumscape here most likely will change the battery industry, but it takes seven, eight years to develop that kind of a product for automotive standards.
Sebastian Thrun:
That's true.
Vinod Khosla:
So, it is already a unicorn if you look at the last valuation and they have plenty of money. LanzaTech is doing steel mill exhaust gases taking carbon dioxide and partially combusted carbon monoxide and the company is doing very well. I'm almost certain it'll be a unicorn, doing extremely well. View Glass just got a valuation of $3 billion all from the same cohort. I could go on and on. I think a company like Calera which has lost visibility will revolutionize cement production, so cement transportation, steel, glass and buildings. These are massively transformative technologies. You add to that other sustainability technologies like Impossible Foods. It's fun. This is the fun part of the Valley. I think we invested in hamburgers and rocket launches at Rocket Lab about the same time and they launched about the same time.
Sebastian Thrun:
Which one is more important?
Vinod Khosla:
Well, hard to say which one's more important. Right?
Sebastian Thrun:
Both are. Impossible Foods is a great company.
Vinod Khosla:
Impossible Foods is a great company, but it can transform the planet. The goal is to free up 30% of the planet's land area. It's not for the faint of heart to be that ambitious and the founder, Pat Brown is just amazingly, I have no question it is on path to be... It's already a unicorn, but it is on path to be a very large company if they execute well. And there is always a if.
Sebastian Thrun:
Tell me. So-
Vinod Khosla:
Look, cleantech was not a failure as people view it. We started four battery companies. I think three of the four will make us money and one of the four, Quantumscape, will make a lot of money. We started three solar companies and probably maybe two of the four will make money for us. So we clearly had some spectacular failures in biofuels, but LanzaTech will make up for many of them because it's doing so well. So this is the nature of the venture business.
Sebastian Thrun:
Let's talk about other fields that are not yet disrupted. I would say banking, fintech. There's this constant threat of fintech companies fizzle into banking, but I haven't seen the massive change yet. What's your prediction with that?
Vinod Khosla:
You haven't, but you take a small product that nobody knows. Like Square Cash.
Sebastian Thrun:
Yes. Which I know about.
Vinod Khosla:
We invested in the series A of Square. The company's worth about $25 billion or something now, but more importantly in the last two or three years, if you look on Twitter, they don't call it Square Cash. They call it Cash App. People are using it as a bank account. There are many, many people depositing their paychecks into Square Cash. The writing is on the wall, another three years.
Sebastian Thrun:
So what you're saying is, it's happening right now. Just open your eyes, Sebastian.
Vinod Khosla:
Yeah.
Sebastian Thrun:
All these little pieces. Put these pieces together and tell me the next 10 years of the future of banking is.
Vinod Khosla:
Four years from now, five years from now, I think somebody new, whether it's Square or Apple or PayPal will be the largest consumer bank in the United States.
Sebastian Thrun:
Vinod Khosla:
Or Facebook possibly. Yes. So entirely possible. That's what will happen in banking. That, by the way, then leads to this capital and banking is 60% mortgage lending for all the consumer banks. That will be deployed much more efficiently and without the overhead of traditional banks.
Sebastian Thrun:
Which brings me to regulations, right? In medicine, in banking, there's this incredible strong piece of regulatory oversight that forces on banks, really bad balance sheets, incredible compliance. In medicines, even worse, approval of medication, reimbursement codes, moving at glacial speeds. How does this come together?
Vinod Khosla:
So there's many ways. In banking, wallets don't have all the compliance requirements, but if you, for example, used AI for compliance, not 10,000 back-office workers in India or the Philippines, which is what the banks do, you’re going to completely eliminate compliance overhead. Compliance is for a good reason. I'm not saying it shouldn't be there. A lot of fraud happens when you don't have compliance rules, but that compliance can absolutely be made much better and very low cost with AI. So when people talk about “it has all this overhead”, yes, because they got 10,000 people doing what is now being addressed with what's basically back office processes in outsource legacy branches that shouldn't exist. All this infrastructure will go away and I think we can keep all the compliance and yet have a new style of banking with very low costs.
So no question about it. Lending becomes much more unbiased. When lending happens today, it is extremely biased by the biases of the humans doing the lending based on their prior experience. That will change.
Sebastian Thrun:
Better, more efficient. Higher returns.
Vinod Khosla:
Better, cheaper, more accessible and more efficient. Absolutely. That's true in banking. In medicine, if primary care for a month costs less than the cost of copay, the world is a different place.
Sebastian Thrun:
That is very true.
Vinod Khosla:
My bet is what is called copay today is more than enough to provide the whole service in many, many cases. Not all.
Sebastian Thrun:
Wow. So no health insurance. That's your future for health insurance. You don't need it to make it.
Vinod Khosla:
Well, you will need catastrophic health insurance if you need a liver transplant or a cardiac surgery, that's a whole different thing. But could you maintain cardiac health for less than $10 a month? Absolutely. And have much better 24/7 AI chatting with you and advising you, right. You can do that in primary care. You can do it in mental health.
Sebastian Thrun:
Almost everything, I agree. Tell me, as an investor, you're an early stage investor, so you've worked with lots of small founding teams. What makes your winning pitch or winning founding team you want to support?
Vinod Khosla:
First, it's unreasonable founders. Unreasonable founders who don't take no for an answer, persist through lots of down cycles because startups are about up and down cycles and have a great presentation. I first wrote up in 1986 called The Entrepreneurial Roller Coaster: High Highs & Low Lows.
Sebastian Thrun:
It's like Ben Horowitz's book.
Vinod Khosla:
Yeah, and I literally did this in 1986, the same presentation. It's now translated to the new version of PowerPoint, but other than that, it hasn't changed much. It's founders with persistence, clarity of thinking and a large mission.
Sebastian Thrun:
How much is experience important for you?
Vinod Khosla:
Not very. In fact, I would argue that in the startup world, the more experience you have, the more biases you have and the less likely you are to innovate in any area.
Sebastian Thrun:
I, as a Stanford professor, should not be running Udacity?
Vinod Khosla:
Let me put it this way. I look back at the last 40 years, and I could not find one large example of innovation coming from people who had experience in their industry. Genentech was started in Kleiner Perkins offices with no pharmaceutical experience. All of biotechnology happened because of an intern at Kleiner Perkins in the '70s. Twitter and Facebook and YouTube innovated media. It wasn't Fox and CBS and NBC. It wasn't Boeing and Airbus or Lockheed that did space, it was SpaceX and Rocket Lab. It wasn't General Motors and Volkswagen.
Sebastian Thrun:
Even self driving cars, it's Google.
Vinod Khosla:
Self driving cars, it's Google, and on electric cars, it's Tesla. In banking, it's Square and PayPal, not Visa and Bank of America. I couldn't find one large innovation, incremental innovations, yes, that came from somebody who knew an area. And this is why. When you join an area, if you bring all your biases, you are likely going to end up with the same answers. When you think from first principles and clearly in six months you know as much about the field as people with 10 years of experience, but you have a fresh look and you'd cast every assumption.
Sebastian Thrun:
So let's say your Bank of America and you're listening to this and look at this, a little bit shivering with your hands because your experienced, a bit older. What would your advice be, for say, the CEO of Bank of America? What should he do if they really wanted to -
Vinod Khosla:
So let me give you a flip example. They'll start an innovation group. Any large financial group, it's some people on the side and they can't do certain things because it affects their major businesses.
Sebastian Thrun:
So you almost want to invest then?
Vinod Khosla:
Here's the other thing, right? So let's say an innovative project isn't going well. What happens in Bank of America? The people transfer to a good area. First, they're not struggling with getting funding every year, and so they take the long view and they don't test in a very rapid way all their assumptions, they just assume their assumptions to be true. When things run into trouble, they find other jobs laterally within Bank of America. What does an entrepreneur do? They know their life depends on it. They have no option. They're testing every assumption. They're looking at all the ways in which they may be wrong and they have no way out. So necessity is the mother of invention. That's why it's very hard. So should they? Absolutely, they should invest. GM buying Cruise as an example. There's no way GM could have done that on their own. Right?
Sebastian Thrun:
So then the lesson would then be, if you bid on Lockheed Martin or Google X, make sure that people, they are desperate. Have little resources and are forced to stick with it even if things don't go well.
Vinod Khosla:
The less resources you have, the more your probability of success.
Sebastian Thrun:
That's very true.
Vinod Khosla:
Because you test your assumptions more, you do lots of incremental things and you get evolutionary exponentials, not planned big leaps, which never work. That's like the Russian economy version of startups.
Sebastian Thrun:
Let's go back to your youth. You grew up far away from here in India. Udacity has a huge following in India of people and you in many ways are an incredible role model for so many people there. What advice would you give yourself to a younger self still in India? What to do well, what to do differently?
Vinod Khosla:
Well, I think I was very, very lucky. I used to read very old issues of Electronic Engineering Time. So year or two after they were current here, they got shipped to India and I could go to a place in Old Delhi and rent those magazines. So in Old Delhi I could go rent old magazines and I'd read them and I read about Andy Grove starting up Intel, and I've told Andy Grove the story when he was alive. He was the inspiration for me. And I said, here's this Hungarian immigrant coming to Silicon Valley and starting a company. Why shouldn't I? Why can't I do that? And that became real motivation. So motivation matters, and I think I was right. I was naive about the fact that I was taking a lot of risks, but I had this arrogant confidence in myself that first I didn't need much more than McDonald's to get food and a $200 a month apartment.
That's what I felt I needed back then. And so I felt I could always get a job. So if something failed or I got fired or whatever, I wasn't ever worried about the downside and every person has to make their own trade off. I'd rather take the risk of failing often. But if I succeed, being very successful. I like to say most people, whether it's people, managers, businesses, reduce risk to the point where their probability of success goes up, but the consequences of success become inconsequential. I choose to make the other trade off. I tell young people, be explicit. Now, to be honest I don't want to understand why anybody, especially young people would make the trade off in terms of conservatism, it makes no sense at all because it affects all of the rest of your life and even when you fail, your rate of learning is so great when you take risks that, that's what every young person should be doing.
Sebastian Thrun:
I have to say to a friend, life without risk is boring, just boring.
Vinod Khosla:
Yeah. It is boring without risk.
Sebastian Thrun:
Forget the upside. It's much more fun to experience new stuff. So I guess you're then fearless, how do you make other people fearless?
Vinod Khosla:
Humans respond to storytelling and so examples, talking about it like we are doing here, role models really help. Andy Grove was a role model for me. It's most rewarding for me when I get these emails and get them like, “You are a model for me,” some 25 year old, and that's exciting. It is scary because sometimes they say they gave up this good job to take this risk just because they heard some talk by you.
I think I did a talk in 2015 at Stanford business school and I told the audience of 400 MBAs that I wasn't speaking to most of them, just two or three of them who would change their trajectory based on this hour I was going to spend. I've gotten hundreds, I still get regular emails from people who still hear that talk and say, "Hey, I'm changing my career, my path." I'm going-
Sebastian Thrun:
What was the message in that talk?
Vinod Khosla:
The message was, be internally driven, don't let society determine what you should be or what's expected of you. Every Indian, their parents expect them to be a doctor or an engineer. Don't follow that rule. Be internal, be driven by what you believe in, but also merge that with what's pragmatic. So follow your dream is not good advice. Follow your dream in a way that's pragmatic if you're successful is much better advice.
Sebastian Thrun:
When you look around the Valley from people like Larry Page, Elon Musk and Mark Zuckerberg, they seem to have the same distance from people emotionally. Is that a feature that makes a good entrepreneur, to be less connected to people?
Vinod Khosla:
I think not necessarily. I think what's important is not to be overly influenced, and I talked a while ago about first principles thinking. If your first principles thinking, you can take lots and lots of input and advise everybody to get lots of input but then not follow it but synthesize your own path based on the input. There's a very distinct line between getting lots of input and following lots of input versus synthesizing that into your own trajectory and most people don’t do that. I've always done that. It came naturally to me.
Sebastian Thrun:
You are one of the very few, I think, that talks about it. I think so does Elon Musk, so does Larry, but most people don't think this way. That's really interesting. It's something I would love to share with the audience honestly because that's what makes all great innovation happen. Just think about it logically, like why do people die of cancer, they shouldn't.
Vinod Khosla:
They shouldn't.
Sebastian Thrun:
It's totally preventable. Good. Last question for you, Vinod, as I love talking to you. Silicon Valley is a bit of a bubble in some sense. You get these rebels here who think logically and very big businesses. I'm sure you've been asked this question many, many times, but what stops this kind of DNA from engulfing the entire world? Why is this not as a Silicon Valley in India and in Europe perhaps in most places of China?
Vinod Khosla:
Yeah. So let me say first, give a word of caution. There is no appropriate time to be a rebel. So I don't, for example, recommend people drop out of college. There's a base level-
Sebastian Thrun:
I do.
Vinod Khosla:
Which is fine.
Sebastian Thrun:
Come to Udacity. You can learn much better skills at a much lower rate and we are done faster. Sorry, I have to do this. Go ahead.
Vinod Khosla:
I understand. But learning base skills so that you have any skills, any way you do it, where you can always make a comfortable life for yourself initially, your family later, really, really important. But then once you reach that level of knowledge, you want to increase your rate of learning by taking lots of risks, experimenting a lot, trying new things, and that's a culture. I like to say Silicon Valley isn't a place, it's a culture and it's this culture of, it's okay.
If I go to New Delhi today and somebody is recruiting, I say, "Oh, that person has 20 years of experience at Cisco or IBM or GE” that's treated with reverence. If I see that resume, I'm not going to talk to that person. I'm not hiring anybody who's worked for 20 years at one of these companies. If they had any self-respect, any risk taking, they'd be out of there so fast. It's a disqualification in my world. That culture here for an entrepreneur for a 25 year old is supported by investors who do that. Other executives who are willing to take a job at a startup instead of at Google, which is safe.
So whether it's hiring more management to help you, whether it's getting investors to take the long view or take this ridiculous view of, oh, you're worth $1 billion with $1 million in revenue and the ecosystem of people who'd be embarrassed to say don't take risks. That's the culture of the Valley that needs to be replicated and frankly, if it is replicated in other parts of the world, we are going to see a lot more innovation. There is lots more headroom for innovation. Let me end there because people always think there's only a few places where innovation can happen.
When we did the Impossible Foods, the Impossible Burger, people said, "You're innovating in hamburgers. What's there to innovate?" 90% reduction in land area needed to produce meat, 80% reduction in carbon emissions or water use. Those are stunning innovations. Stunning. In fact, more than anything that's happened in software, we are 3D printing, experimenting with 3D printing houses now. We have three construction startups, who'd have thought construction, transportation, food would be areas for innovation. There is no area that is not open to radical innovation in my view. It just, people's mindset limits them. People always find it easier to explain why something will fail than to say, I'll figure out a way around that failure, so that's a good place to end.
Sebastian Thrun:
As a great place to end. I 100% agree. I would challenge anybody watching this go and innovate tonight, get a glass of beer and think about something crazy. I would say everything crazy can be done. The people who say can't be done are always wrong. You know it.
Vinod Khosla:
Pardon me, we just funded a seed project to do a Mach-5 passenger jet.
Sebastian Thrun:
Good. Which-
Vinod Khosla:
By the way, London to New York in an hour and a half, the world will be closer together.
Sebastian Thrun:
I agree and it should be done and it can be done. There's no physical reason why it shouldn't be done.
Vinod Khosla:
Absolutely.
Sebastian Thrun:
Great business case. Thank you so much. I could speak hours with you. I've learned something from you.
Vinod Khosla:
Thank you.
Sebastian Thrun:
He's one of my heroes and role models and everything I say, as you say. I mean, it's just very logical.
Vinod Khosla:
Oh, it's great to always chat with you, so it's always fun. Thank you.