2040s: A Lookback | Vinod Khosla

I hope all of you had a good day yesterday. Of course, you won't say it was a bad day, so hopefully I'll get some brutal honesty, hopefully in the surveys. They're anonymous. So what I want to do today is look back at the early 2020s from 20 or 25 years in the future. And I want to sort of add my perspective and why I might be optimistic. But let's start with how the 2020s started. Of course, the coronavirus. Big event. Pretty horrible event, to be honest. Ukraine. And, you know, these were two— and Ukraine is ongoing— really catastrophically bad events, probably bigger than we might have imagined possible. And it's very sad because a lot of people got hurt by both the war and the coronavirus. But I'm an optimist. Those of you who know me, I'm fundamentally an optimist. So I look on the bright side, and I'm going to argue that if we look back from the 2040s, 20 years from now or 25 years from now, As sad as these events were, they may be the two most positive events that happened on the planet. And let me argue that that's the case. So what did COVID do? It focused us on China's control of the supply chains. Almost all the steel is made there. All the supply chain got disrupted there. Rare earths, all the manufacturing. Uh, we were hostage to a lot of Chinese policy globally, and identifying a problem is the first step in solving it. The status quo, all of you know this, It is amazing how much we depended on this and didn't worry about it. I do believe this creates brand new opportunities for entrepreneurs, and it focuses us on the real long-term danger of concentrated point sources. We also proved we could do a set of things that weren't possible before or thought impossible. How fast could vaccines or drugs like Paxlovid could be developed? Nobody believed telehealth was possible. I won't go through all of them, but you get the idea. So many areas where we innovated where we didn't think something was possible or a market didn't exist. Ukraine did the same thing. I couldn't have 2 years ago imagined Germany would say no dependence on Russian gas. The world desperately needed an energy transition because of climate. But we weren't anywhere near on the path to be serious about that transition. We never realized how geopolitically and environmentally disastrous our dependence on fossil fuels were. My view, there's a cascade towards energy independence, and more importantly, an energy transition, because I view climate as one of the larger dangers. In a way, we will never wake— go back. It was a bigger wake-up call than anybody could have imagined. And now we are on this new path. We are on this well beyond in many, many areas, not just oil and gas, beyond what we imagined a few years ago. And it'll give us the next 20 years as a really great time for innovation. We got policy passed that got us moving towards this transition. Lots of dollars, and the Europeans have their own equivalent of this. Yes, we— the Ukraine led to global repolarization, but it was an important wake-up call again, which creates an opportunity for all of you. I wrote about this in July of '21. Climate's a large problem, but I believe it only takes a dozen or so people to affect that transition. To be instigators of change. You know, frankly, whether you like him or not, Elon Musk caused us to believe in electric vehicles. Um, I have a slide I'll maybe get to. Pat Brown coined the term plant proteins. It wasn't really a word or a category. And then of course Bob Mumgaard really got the world interested in fusion. A few individuals in major areas got the, got the world to focus on the change that is possible and convince them that's possible. I think there's only a dozen areas to solve the climate crisis, and it takes people like this and some of you to solve these problems. Some transitions are already in process. Some there's, I think, fairly good solutions for that are still being de-risked. And then of course, areas we don't have solutions yet. But these few areas are what are both necessary and mostly sufficient to solve the climate crisis. It doesn't take thousands or tens of thousands of people. And then we had this other, uh, game changer, which is very recent, but the progress has been going on exponentially for quite some time. The explosion of innovation we are seeing now is hard to imagine. Nobody can predict where this is going. But being optimistic is a good way to be, or panic if you're in the old business. Um, and panic is a good thing if you can use it to transition your business. Lots of axes of innovation that help AI. I won't go through them. Plenty of older innovations. That are going to help in this, that will all leverage the AI trend. I'm very, very optimistic. Yesterday you heard about physician AI, that everybody in the world can have a personal physician, a personal oncologist if they need it. Every kid can have a personal tutor if they need it. No matter who you are, you can have a personal financial advisor, even if you make— you're at the bottom of the economic rung. So imagine all these things that are possible. Personal media, personal songs. I personally believe 80% of 80% of all jobs, so that, that's 64% of all jobs, will transition and be capable of being done by an AI. And I say capable of because we have the choice whether to let these AIs do those jobs or not. And I believe some countries will, others will be laggards for lots of other reasons. The other thing I want to say, because this is a room for entrepreneurs, I've never seen large innovation, and I've been in the innovation and only the innovation business for the last 40-some years. I've never seen it come from a large entity, an institution, a policy. Um, and I can go through, All the innovations, the large innovations, small innovations can come from bigger companies. What I'm saying to you is the world's innovation is reliant on all of you. It's you who will make this change or not make it happen. And so there is a burden on this crowd that we talked about very briefly last night at dinner. There is no example I've seen in 40-some years I've been doing this where the large innovation came from somebody you'd expect it to— GE or Siemens in energy, for example, or Pfizer in biotech. Of course, I believe we will get to a place in the not-too-distant future of I believe before 2050, if we are aggressive about leveraging these technologies and leveraging them carefully, and all of you aggressive about innovating in these areas, the need to work in human society will actually disappear. People will work because they want to work. I wrote about this about 6, 7 years ago. Article in Forbes where I said all the economic metrics will be great. Productivity will go up, GDP growth will go up, abundance will go up, but I taglined it with increasing income inequality, and I'll come back to that in just a minute. There are other physical things we need. And those are harder problems, but I believe technology is being used to find better metals, minerals, sources of ores, better refining techniques. My dream is all the lithium comes from the ocean. That's entirely possible. And we've started to see plans to do that. We've started to see biomining instead of mining, much cleaner. And in this kind of timeframe, all that will be realizable if some one entrepreneur takes it on them to change this world. And I said this 6 or 7 years ago in this article in Forbes, that UBI will be affordable. And I briefly referred to this last night. The world is on trajectory for 2% GDP growth. You project it, at least in the US, for the next 50 years. GDP will grow from $65,000 per capita— or sorry, not GDP, per capita income will grow from $65,000 to $175,000. If this room does its job, we can accelerate GDP growth to 4%. With all that productivity gain? If we do, average per capita income in this country would be $475,000. Such a substantial increase in per capita income that UBI, or Universal Basic Income, will become affordable. And I referred to this about, uh, in this article 7 years ago. But there are some caveats. There's scaremongering around AI. We've all seen that. Personally, when people ask me about sentient AI or AI killing us, I find it such a silly debate. I'm way more worried with strong AI in China's hands doing— in bad actors. And I believe we are in a techno-economic war with China. I've spoken a lot about it this year. And whoever wins that war, and I call it a war because it's a war for economic dominance, for this abundance I've talked about. And if we win it, we will win global political influence. And I believe what is at stake is what values rule the world. I do honestly believe that China believes they have a superior political system dominated by the Chinese Communist Party. They believe that, and I can't argue it's not, but I much prefer Western values. And I think we are in a race for whether Western values dominate Africa, Latin America, Southeast Asia. And so this scaremongering around AI worries me. There will be job displacement and that can slow down all this technology deployment. You know, truckers out of jobs, physicians out of jobs, all of that will happen, but there's a much bigger prize at the other end. Of course, politics can derail all this and we have to be very, very cautious. We can miss this opportunity, and the worst would be if we miss it and our adversaries in this techno-economic war, which is mostly China, wins it or deploys it. And I've referred to this inequality question. I think there's solutions, especially if GDP growth and per capita growth happens as I've projected. And I don't think 2% to 4% is that big a change, especially with the technology at our disposal, with the kind of innovation engine we have in the Western world. And I refer to the Western world, not just the US. There was an article my son sent me that I'd done an interview with Laura Holson in the New York Times about 20 years ago, and I talked about What it Means to Be Human in the Age of AI. This was around 2000. We will still have tribalism, racism, American versus German. We will have income inequality. And I do believe one of the larger problems will be the source of human motivation. I actually believe it's wrong that that motivation comes from working an assembly line job. That's the wrong kind of job to get your motivation from. It doesn't lead to happiness. So I do believe these are real issues we will have to address, all addressable. So I do think we have a very good time ahead of us, especially as innovators, but really as a society. Imagine some of this. Almost anything that we think humans can do, AI will be able to do it for us, and we will do it if we want to do it because we are passionate about doing it, not because we need to do it. The other question I get, uh, challenge with is change doesn't happen very fast. I disagree. This is New York City, Fifth Avenue, in the year 1900. There's one car in this picture. Rajesh found this set of pictures and I just love them. This was New York City only 13 years later. One horse. Old cars, Fifth Avenue. I can't imagine at the beginning of the 20th century change can happen that fast. This is my favorite. This was a Department of Energy projection in the year 2010. They actually made this in 2011 for the number of cars, electric cars in the US. In 2035. And this tells us an important story. They talked to all the experts and extra experts only extrapolate the past. So if you ask General Motors or Volkswagen or Ford how many cars there would be, the DOE came up with a projection of, uh, 2,300 cars. In 2014, Tesla shipped about 80,000 cars alone. In 4 years, they were wrong by 20 years. That's how fast change can happen. And why is the case? If we had relied on General Motors and Volkswagen, we would be on trajectory in 2035 to have 2,300 electric cars. Fewer there than there are in Palo Alto alone today. But it takes this instigator I talked about to not let the change happen, to make the change happen, to make fusion happen, to make plant proteins happen. Of course, I could go on. I won't spend too much time on this. Can employment change by 90% in a short period? In 1900, 40% of US employment was in agriculture. By 1970, it was 4%, a 90% shift in employment. In 70 years back, a long time ago, and remember this involved people moving from cities, from rural areas into cities, very large changes. In today's age, this is a third of that time. What can we achieve in 20, 25 years? Scaremongering, I talked about it. This is a cartoon from 1989. They thought electricity would kill people, very much like AI will kill people. This is one of my favorites. This is a phrase I first adapted in early 1980s. This is all of you. This is why experts say it's not possible. Lord Kelvin, who was the president of the Royal Society, maybe the most informed person in science in the early 1900s said in about 1906, heavier-than-air flying machines were impossible, just before the Wright brothers flew their plane. Of course, Time magazine was the expert on e-commerce. Don't rely on experts, don't reply on on what the world tells you is possible. Create the world you want. I say there's two types of people: people who extrapolate the past— lots of those— and people who invent the world they want. That's all of you. Thank you very much. I'm gonna I'm going to just show you a smither of what's possible. So last weekend was my daughter's wedding, and I did my little speech, and I wrote it as a human being, no ChatGPT. She would have hated it. But then I asked my AI friend to first turn it into lyrics and then actually rap the whole speech. And so can we cue that up? I just want to illustrate how broad a range— this is entirely generated. We can stop it. You get the idea. That was a clone my voice trained only on 20 minutes of my speaking since I don't sing. Imagine what could be done with much better systems. I think all of this— and this is courtesy of Splash AI. Stefan is here. Is Stefan here in the room? No, I guess he's sleeping. Let me open it up to questions, to challenge, hopefully. Um, but you see, I just want to give you a flavor of what's possible. You heard about physician AI, and later today you'll hear about other kinds of innovations, non-AI-based innovations, how we can do a number of things. But let me not spoil that, uh, session. Any questions? Yes.

Nabih.

Let's go there first.

That was an awesome presentation. Yesterday, Neil's overview of CURE AI, we talked about how you have to sort of disclaim medical recommendations AI make. Why are we willing to accept errors and misdiagnoses and poor treatment from actual humans, in the medical space, but it is unacceptable to have a poor recommendation from AI.

You know, this is an interesting question Neil and I have debated extensively, and I have said to him the most important system, because diagnosis is often ambiguous, he has to develop is a triage system. When you're in doubt, when there's a possibility of error, pass it on to a triage, to a human being to make that mistake. Much more acceptable to society. And, and I'll come back to this issue. This is really, really important. And I think if humans make mistakes, we find them more acceptable than if systems do. We will get past that. Elon Musk has now convinced the world his self-driving causes less, fewer accidents and less death. But he has hundreds of millions of miles driven now in Tesla cars with their clunky, well, deeply criticized self-driving. But it's still not okay. In fact, Kai-Fu Lee, who's one of the prominent AI researchers in China, said to me a couple of years ago, your disadvantage is if a self-driving car kills a human being, You're going to delay self-driving cars by a few years. In China, the government will say that's the cost of winning a race. So we do have a disadvantage. Um, so systems will make that mistake and we have to learn how to triage them. That's why the term copilot is such a beautiful term. Um, you know, most people don't recognize just drug interactions. The same person given two drugs that don't work well together causes more deaths than breast cancer in this country every year. Misdiagnosis causes more deaths than breast cancer. But we accept it from humans. We just have to live with social phenomena and ease our way into it. Thank you.

Thank you, Vinod Khosla. It was super inspiring. I wanted to ask you more about the geopolitical crisis that's going on internationally. Having lived in many countries growing up, I think about it a lot. And when I'm reading news of China taking over airports in different parts of Africa, what's happening in Taiwan, as an instigator and very focused and passionate about the work I'm doing, I'm I draw a lot of motivation to do my work better, but nobody's taking care of those things, and it's very concerning.

There are different degrees of actionability of these large problems. I would say that's why COVID was so great. The world is now sensitized to China's designs and their view. And of course they've invested heavily in Africa, other parts of the world, um, and locked up resources like mines and other things. If we invent much better ways of doing those things, that's up to all of us to do. So it's not everything's actionable. Vinod Khosla. You know, the world isn't all rosy rosy. It was in 1986, I did a presentation called The Entrepreneurial Roller Coaster. The piece that people miss is the tagline where the highs are high and the lows are low. And all of you know the lows of entrepreneurship. That's life. But more is possible than not is sort of my optimistic view.

Yes.

I was hoping that 2040s looking back you would talk about what happened to the world after China invaded Taiwan. Do you have a point of view on what happens? Look, China's— Taiwan is an important tactical question. It is complex and I don't fully understand all the dynamics around Taiwan. I do know China is extremely focused on the chip industry in Taiwan. That's the whole reason in this country we got the CHIPS Act. This is the wake-up call that I think was good. How we handle it depends on who gets elected, who's president, the Western policy, what happens in Europe. What are the major current issues? It's very hard to predict those dynamics, but I'm sort of optimistic. Well, last question. I can't hear you.

Okay.

Thank you for an excellent presentation. So you were referring to UPI as sort of a positive end goal. One of the historical facts of Keynesian economics was that his predictions on what's going to happen with increased productivity really didn't pan out the same way, right? Keynes said that people are going to use the free time to grow spiritually, engage in the arts, and I think it's enough to see what happened during COVID to see that that's not necessarily what happens when people have more free time. So I guess my question is, to what extent your optimism hinges on the idea of UBI. And if not, right, if that's not necessarily the best outcome, what is?

Well, I do believe the US political process and mostly Europe is a complex, messy process, but we in the end migrate towards the right tradeoffs. And so I do— I'm optimistic politics will migrate there. And, um, uh, you know, if, if we are wrong for one 4-year cycle, we fix it in the next. So we have a way in Western democracy to reboot systems that go wrong, to fix things. So I'm optimistic. I'm, I'm pretty sure it won't be a straight line. It'll be 3 steps forward, 2 steps back. It'll be messy, but I do think we will get to the right place. Thank you all very much.