‘There Will Be A Lot Fewer Jobs In 20 Years…’: Billionaire Entrepreneur Vinod Khosla On Impact Of AI

So by the time somebody who's 20 or 25 today is in their 40s, there will be a lot fewer jobs. And there's no two ways about it in any competitive capitalist economy. 2030s will be very, very disruptive and different countries will adapt different policies. To have more or less impact from AI. IT services will disappear. BPO services will disappear almost certainly within the next 5 years. So the 250 million young people in India or so that are students today, when, as they start to appear on the job market, they should be selling AI-based products products and services to the rest of the world. That's where the economic opportunity has to go. That's where policy has to go. Almost no economist is thinking about it.

India will benefit somewhat from the AI revolution, but if it doesn't catch up to the United States and China, it will fundamentally be much weaker than those two countries.

An Indian company can dominate the American scene if they did it right and built the right kinds of applications. So I would summarize by saying everything's up for grabs.

Hello and welcome back to the Hindustan Times podcast. I am your host, Shashank Mathur, reporting to you from Washington, D.C. There's really two stories coming out of America right now that are changing the world. The first is the rising tide of artificial intelligence, which promises to transform the global economy. There's also the rise of President Donald Trump, who is transforming the world in a very, very different way. Joining me today is a very special guest, someone who has more than made his mark in both technology and politics— Vinod Khosla, the billionaire co-founder of Sun Microsystems, who now leads Khosla Ventures, also a prominent political donor who has been at times a trenchant critic of President Trump. Mr. Khosla, it's a pleasure to have you here on the Hindustan Times podcast. Um, karein, begin for the benefit of the people who don't know you, although they must be very few. You were born in India, studied in IIT Delhi, you moved to America in the 1970s, founded a hugely successful company Sun Microsystems, you became an even more successful venture capitalist. It has been a long journey for you in Silicon Valley. Could you begin by telling our viewers a little bit about yourself and how you came to be Vinod Khosla, who you are today?

Well, I'm very much interested in what's interesting to me. I don't worry about careers. I don't worry about titles. I don't worry about essentially fitting into an organization. My focus from the day I graduated from IIT was to do a startup of my own, because I'd heard about a Hungarian immigrant, Andy Grove, starting a company, Intel, in Silicon Valley. Right. And so that was my dream. And I found my way to Silicon Valley and started my first company, which was also quite successful, though people don't know about it. It went public. And then I started Sun Microsystems right after that. And I've been in the business of enabling entrepreneurs, helping entrepreneurs, assisting entrepreneurs, uh, my whole life, helping create large impactful companies and social impact. So that's sort of been my area of interest and how I got where I did. It's always trying to make things happen that you want to happen.

Mm-hmm. Vinod Khosla, you got to America, I think in the 1970s, you did your MBA at Stanford. You must have been one of the very few Indian entrepreneurs in tech in Silicon Valley, completely in contrast to what we see today. What was that like? You were a bit of a pathbreaker for Indian entrepreneurs more generally. I think you were the first generation of Indian entrepreneurs to make it to America and make it big there. Could you tell us a little bit about that?

Well, the startup ecosystem back then in 1980, when I started my first company, was very, very different. It wasn't a 20-year-old starting a company. It was establishment people starting companies because only they could get large amounts of funding to start a company. So my approach was very simple: just do it. Don't worry about the rules, the regulations, the biases, all that. Just ignore it and just do as— go start a company, which is what I did. And I mostly just ignored everything else.

Right. You know, could I ask you a little bit about Vinod Khosla, you've been writing a lot over the last couple of weeks and months. I've been reading a lot of your work. You're talking about artificial intelligence and how that's going to change the world economy. That's something that you're particularly focused on. Talk to us a little bit about that, because I remember reading something you wrote, which is about how you were around for microprocessors, which was what Sun Microsystems was so successful at building. You were there for the internet revolution. You were there for the smartphone revolution. But you say that artificial intelligence is something completely different. To what we've seen in the past. Talk to us a little bit about that. Why are you so bullish on this?

Well, all those were really massive changes in platform technology, and each enabled the class of applications and uses and companies. And so each of those has been platform shifts that then enabled hundreds of new kinds of companies to be built. Uh, AI is Another large technology change of the same ilk where you create a platform and that platform, uh, and the last one was the mobile phone, enables a whole set of companies. So the internet enabled a class of companies like Google and Amazon. The mobile phone enabled the class of companies like Airbnb and DoorDash and Uber. Uh, AI is enabling a new class of companies. What's interesting about AI is previous technologies helped do various jobs. AI is creating human intelligence, uh, where it's very, very likely in the next 5 years AI will be better than most humans at most things.

Mm-hmm.

So there's very little, uh, where humans will be substantially better. I'm not including areas like art and literature and song, and those will all be opportunity areas for humans.

Mm-hmm.

Uh, but in terms of economic functions, AI will be able to do that. So it's a massive transformation of the US economy, but also the global economy. I have just written and not published a piece on what AI can do in India. In September 2024, I wrote a large piece on what AI can do for the global economy. I think it can— it will cause a lot of productivity gain the next 5 years, which all sounds nice.

In fact, that's what I wanted to talk to you about because I think it was a 13,000-word piece long. And I must say that I read the 3,000-word version, so I, I'm I must admit I am weak and wasn't able to read the whole thing. But if I could ask you about what you wrote there, because you talked about how— one example that stuck out to me was you were talking about bipedal robots that are enabled with artificial intelligence coming in, doing manufacturing, doing a lot of things that humans do right now. And that would cause immediate disruption in terms of jobs, as technological revolutions have in the past. So could I ask you this? For people watching right now, what does the world economy enabled by AI look like in 10 to 15 years? How does How do things around you that exist right now transform completely in terms of just your daily living?

Okay. So there's two parts of it. One is, and this is all detailed in this large 13,000-word essay or blog, uh, what is technology capable of and will be capable of in 23rd— today, 2030, and 2035. There's a separate discussion of what are the economic implications of that, and they're very, very large. So very simply, the first step is almost all expertise will be in AI, and AI workers will be able to do accounting. You won't— they'll do accounting better than accountants. An AI worker can do— be a physician, a doctor, an oncologist, a mental health therapist, a physical therapist. Um, AI will be chip designers, AI will be salespeople, AI will be, um, building designers and architects. So AI capability will allow you to hire AI workers that do all those jobs by 2030. Now, before then, there'll be AI workers who can work as a junior assistant or an intern to a human architect or a human physician or a human chip designer or a human, uh, pick your favorite function, accountant. So starting almost today, AI assistants or interns are available to many, many professions. And by 2030, there'll be professional workers available in most areas. And we are doing that kind of thing. I recently counted 30-some professional workers we are building. They'll all start out as interns and assistants to humans, and then we'll, uh, move on to being full-fledged workers, experienced workers, over 5 years, which is what human workers do. You hire a fresh college graduate to be an accountant, and then 5 years later they're a senior accountant. So that's the model of expertise, right? The next wave of larger revolution coming is robotics. Uh, so the first phase replaces human intellect, uh, what you learn, what you learn in college, what you learn in medical school, what you learn in architecture school. The second wave is work as a human worker. So most labor— it doesn't matter whether it's cleaning the dishes in your kitchen or working on an assembly line or working as a farm worker— robots will be doing those jobs. And I think they will take 5 years to mature. They're about 5 years behind intellectual work or thinking work. And so you'll start to see it pretty extensively in the next year or two, and by 2030, next 5 years, they'll be doing most manual work fairly well. Maybe not yet cardiac surgeries, so, or brain surgery, uh, that'll take an extra 10 years. But in the next 5 years, for most work, whether it's assembly line workers, uh, or, uh, home workers, you'll start to see AI do those jobs as robotics, right? Uh, and then over time it'll get better and better at precision, like a fine tailor stitching, or a cardiac surgeon doing heart surgery, that'll take a little longer. Now, what I would say to you is, in San Francisco, there's probably as many, uh, self-driving cars, um, in the, uh, high drive ride-hailing business. So there's more Waymos than Lyfts and Ubers, uh, in terms of ride capability. Yeah, not quite true today, but it'll be true by next year, right? Today, uh, Waymo driverless cars, which are robots, human robots, uh, replacing, uh, or sorry, robots replacing the human driver, are pervasive in San Francisco. It's hard to not notice them. In fact, it's to the point where nobody notices them anymore because there's so many of them. Right now. And if you haven't ridden on one, it's a magical experience, feels much better than getting an Uber. So this will start to pervade. The question is, what impact does it have for the economy? The biggest impact will be jobs. Start to see that in the next few years. But because these things are so much easier to do and so much cheaper to do, you know, robotic labor will cost $2 or $3 an hour, not $20 an hour, right? And if labor is 10 times cheaper, if a doctor is 10 times cheaper or 100 times cheaper, um, if an accountant is 100 times cheaper Uh, it has implications for the economy. So my bet is you will see reduced inflation and then going to huge deflationary economy by 2035 globally. Now let me spend a few minutes on this piece. Uh, it's a new piece I've just finished writing on.

Actually, I, I do want to talk to you about that because I do want to talk to you about the impact of AI on India, but just a very quick question because I was reading something by a Nobel Prize-winning economist, Daron Acemoglu. He wrote that, you know, much of the economic benefits of automation over the last 30, 40 years have been cornered by people who are in the elite, the 1% of the 1%, you know, the business owners. You've written about how you also see the fact that the introduction of artificial intelligence will mean that capital— people who have money, who have resources, who back businesses— will have much more power relative to labor, people who work on the factory floor, for example, people who work ordinary 9-to-5 jobs. Could I ask you a little bit about what that kind of disruption does to everyday politics and society? Because, you know, we would— I would love to talk to you about President Trump, but his political rise and the rise of a lot of authoritarians across the world is because of the fact that people have lost their jobs, people have lost their livelihoods. Does AI pave the way for something like that to happen globally? Because people are going to lose their jobs, they're going to be angry, there's going to be backlash.. And that is not always very good for democracies and political systems. I would love for you to talk a little bit about that before we get into the India piece of this.

So I covered this fairly extensively because the question, obvious question is what will people do when AI is doing all these jobs? Um, it's very clear over 25 years. So by the time somebody who's 20 to or 25 today is in their 40s, uh, there will be a lot fewer jobs, and there's no two ways about it in any comparative capitalist economy. Um, now the good news— and there's good news and bad news— I think much of policy will drive how the benefits accrue. So different economic policies will lead to lead to different outcomes. Different political systems will lead to different outcomes also. I think democracy may slow down the adoption of AI in some countries, accelerate it— for example, robotics will accelerate in Japan because they don't have enough labor and definitely don't have enough people for elder care, for taking care of elders. So robotics is moving very rapidly in that part of the world. Uh, also in China, because China has an aging population. Um, China's, uh, administration or China's, uh, politics allow them to be dictatorial. So they are exercising that to make sure first they win the race for global robotics and global AI. Uh, and it's a very serious race. Uh, but to your question, uh, and I was getting to this, I describe it in my long 13,000-word paper, the 2030s will be very, very disruptive, and different countries will adapt different policies to have more or less impact from AI. Mhm. Now the opportunity, and this is why I've just finished a new piece, um, I will be posting it very shortly on our website and I'll be happy to share it with, uh, you and others. Uh, AI can do two things. Before 2030, AI can make a doctor available to every Indian for almost no cost. AI can make a tutor available to every Indian child and adult for adult education, uh, to every adult for almost no cost. AI can make a lawyer available to every Indian so they can have their legal rights asserted, have access to the law. So there's a lot of good things. AI can make entertainment almost free in a country like India. And these are all hugely deflationary kinds of efforts. Now, the benefits of this AI need to be redistributed by government. So redistribution is a bad word, right? But if there's infinite production of goods and services like medical care, like education, like transportation with self-driving cars, then there aren't jobs, but these services become essentially free, and governments will be able to provide a much higher minimum standard of living, right? There's a few hard areas, housing being the most important one, uh, food is the other big one, uh, in America The bottom half of the population spends about 40% of their income on food and housing and healthcare.

Right.

So we do have to address that. And I do think AI can play a large role in dramatically reducing the cost of housing. But those are some of the areas where we will have to redistribute income to create more housing.

Right.

And a minimum standard of living for every citizen by 2040. That's entirely possible. Some of these services can be free if the Indian government does it by 2030. And I'm looking forward to the India AI Summit in February. Uh, hopefully this message is talked about and the policies that are required to make sure that every Indian has a higher standard of living. And that is government policy, economic policy. That has less to do with technology, more to do with politics. And policy. But it's a massive opportunity for India to completely change the trajectory of services available to its population.

Just two questions on that, because the first question is about the fact that what India wants to do, of course, now is manufacture more, create more jobs that way, because of course, for billions of people, tens of millions entering the workforce every year, and the fact is that there aren't enough jobs at this point in time for all of them. Youth unemployment in India is high. And what you've said no holds barred, which I appreciate the honesty, is that a lot of those jobs are not going to be around. And so giving people services, ensuring that healthcare is good, and so on and so forth is going to be great. But I think any Indian government is going to ask, what about the jobs? Because if we can't create tens of millions of jobs, and in fact, we're going to be losing jobs because of AI, what does that do to us? Because, you know, it's going to be hard for them to go to the election polls and tell their people that, yeah, they're losing jobs. I'm sorry. Um, and it might be—

this is why I said politics and policy will make a massive difference on how AI is deployed, right? There are countries who will slow that down. Germany is a classic example, the funniest case. They banned robots from working in retail on Sundays because humans can't work in retail on Sundays. It's, it's a stupid policy. And surely the German economy will decline because of that. Um, but to your question, uh, I think the focus on job creation may not be the right answer. The focus on education for the purpose of getting a job may not be the right answer because AI will do it better. Now, there will be a transition period where I would focus on when humans and AI working together, and that will happen during a decade of transition, or decade and a half of transition. India becomes the largest exporter of AI-based goods and services to the rest of the world. Not IT services. IT services will disappear. BPO services will disappear almost certainly within the next 5 Wow. Next 5 years, BPO, all of BPO services and IT services will, will essentially start to decline rapidly. I would guess much faster than 10% a year, maybe even 20% a year declines in the market. So those are large dislocations, but they will be, at least for the transition period over the next 15, 20 years, room to export AI-based services and have lots of humans work in conjunction with AI and exporting it to the rest of the world, right? So the, the 250 million young people in India, um, or so that are students today, uh, when— as they start to appear on the job market, they should be selling AI-based products and services to the rest of the world. The 7 or 8 billion people on the planet. I think that's where the economic opportunity has to go. That's where policy has to go. Almost no economist is thinking about it. No business manager is thinking about the adoption of AI-based business correctly. Um, and that's a danger. It's a huge danger, and we can't ignore it. Clearly, China is not ignoring it.

Right.

There's hundreds of robotic companies in China. And they can be more dictatorial. They don't have to worry about the vote. I think India has to take a different approach. But it is entirely possible.

And this is the question I wanted to talk to you about. Because I'm glad that you're also acknowledging that there is going to be disruption in India, as well as a lot of other countries that rely on labor. Are going to have to figure this out. But the question I also had was also on the geopolitics of this, because at this point in time, it's very clear that this is a race between the United States and China. And what you're saying is that AI is going to fundamentally reshape the global economy, and the countries that dominate the development of artificial intelligence, that make the chips, that have the data centers, have the most compute, are the ones that will rush ahead. Right now, it very much seems that the United States and China are in a class all by their own when it comes to developing the future of AI. Does this lead to a world in 15, 20, 30 years where the United States and China, having dominated the future of AI, have accrued so much economic and therefore political power that everyone else is just swept to the wayside? You know, can't you see, you know, Europe has economic problems, those get worse. India will benefit somewhat from the AI revolution, but if it doesn't catch up to the United States and China, it will fundamentally be much weaker than those two countries. So a lot of countries are going to look at AI and say,— oh my God, is this a technology that if the US and China dominate, then we're going to be dominated geopolitically as well? Your view on that?

So a couple of things. I think that's a real danger. And it would be bad if that technology was very narrowly focused on two countries, coming from two countries. India does have the talent. It has sufficient talent to develop these technologies. Today's model of AI is very focused on large language models, lots and lots of compute, ChatGPT and so on and so forth, right? Lots of people. I don't think that's the final chapter. Other models of AI are possible and will start to appear in the next few years. And India can get ahead of the curve. First, there are two areas. One, developing the AI. Um, that's a harder question, but I do think India should try and win that race. There's the question of applying AI and applications of AI. It's like you don't have to develop the internet to have applications of the internet, like delivery. Uh, so, uh, or build, uh, you know, you can build a Flipkart on somebody else's internet. Mm-hmm. Um, so, or Zomato, or you pick your favorite application, they will be applications on AI that India should absolutely have a serious effort on. And that does not need that large amount of compute. I'm very interested in this because we have a company called Sarvam building Indian AI model. And the one of the reasons we did that is countries have to have sovereign models, definitely for national defense, but also for other purposes. We have a similar company in Japan doing a Japanese model using kanji script called, uh, Sakana. So I do believe in the idea of sovereign models, and it should be pushed. But applications of AI is wide open, and an Indian company can dominate the American scene. If they did it right and, and built the right kinds of applications. So I would summarize by saying everything's up for grabs, everything's up for rethinking, and none of the older assumptions will hold in 10 years, right? And that has serious implications, right? Look, China was well ahead of the curve about 4 or 5 years ago when they adapted their last 5-year plan. They made winning in AI as an objective of the 5-year plan. And they are executing on it. And India should be thinking in the same terms. So I hope India takes it seriously.

Right. And the question I wanted to now pivot to, away from technology to politics, as I warned you, was a lot of Indians are looking at the second story coming out of America, which is the Trump administration. You have been active in politics for quite a while now. And it just fundamentally seems that American politics has changed in a way that very, very few people recognize. You know, Indian students, Indian professionals are increasingly feeling less welcome because of the visa changes, because what's happening, for example, with the H-1B visa. Could I ask you, as someone who's been in the United States for something around 50 years now, is it the same country that you came to when you were an immigrant in the 1970s? Is it the same country or has something fundamentally transformed in American politics and life?

Well, America goes through waves, okay? In the '70s and '80s, when I first came here, you know, you were still a foreigner, and you've heard all the stories of bias against foreigners and others. Um, but then it went through a phase in the '90s and early 2000s where Immigrants were accepted, most of them. You know, there was such a large frequency of Indian doctors when people went to the doc to see the doctor, Indian engineers, Indian CEOs in big companies. But the left in India— in the US carried the idea of rights, or what's called DEI— diversity, equity, and inclusion— way too far to the extremist, extreme levels. I think they did a lot of damage, and I'm a real opponent of the DEI movement, right? I'm an independent, not a Democrat like people think. I used to be a Republican, but I believe climate is a big enough issue over which I became an independent. I don't have left or right politics. I tend to be in the middle. I think the left did so much damage and continues to take that approach. Mamdani is a good example in New York, will cause a large backlash. Trump's ascendance is the backlash. That too will go away. I don't think it will last for very long. He's an unethical person who lies and cheats and will do anything. He'll invade the countries, you know, I can't imagine, I would not have imagined anybody threatening to invade Greenland, right? A peaceful little country on the side, not harming anybody, a great global citizen. But Trump has no rules. But that too will pass. I think you'll see a backlash and more movement towards the middle. I hope that happens.

Right. And could I ask you this, this, these rules that have come in on H-1B visas, on making it harder for students to come to the United States, do you see this as fundamentally damaging to the United States? Or do you understand where the administration is coming from in terms of trying to limit skilled immigration to the US? Because they're arguing that these H-1B visas are taking away jobs for American technology workers.

Yeah.

What's your sense?

This is very silly, naive understanding of what really happens, how value creation and innovation happens. Look, last 25 years in America, the percentage of GDP growth coming from innovation has been very, very large. I've seen estimates of 40% of all growth comes from innovation and the innovation economy, as it should be. Uh, that is driven by H-1B visas, the interchange. In the '70s, we used to call it brain drain from India to the US. I was guilty of that. Um, recently you've seen people migrating back to India, and that's good for India to have that expertise and knowledge brought back to India. This interchange is very healthy for the global economy, uh, but Trump has stopped that. He doesn't understand that it's about creating jobs through the innovation economy, not— and creating innovation and value to the innovation economy. So I do think Trump's policies are going to do long-term damage. He's mixed legal immigration with illegal immigrants. Or immigration. Illegal immigration is not good for the country. Legal immigration is very good for the country. But he's bundled those together into political slogans, if you might, without differentiating and saying one is good and one is really bad. I think it will hurt the US very significantly not to have the two-way flow of talent between countries like India and here.

Right. Could I ask you one question about this? Because when it comes to immigration and what's happening with Indian Americans, there's been this rise, especially online on X, particularly where you're very active, of this rising anti-Indian sentiment. A lot of hatred, a lot of racial hatred has been directed towards Indian Americans. And their place in America, a lot of it is actually coming from this Make America Great Again movement and people who supported sort of these nativists, as you were. So could you talk to us a little bit about that? Because when people look at that, when they go online, they do think that America has changed because this is the tribe that's currently in power. So talk to us a little bit about that. How do you—

Yeah, it is the tribe in power. Like I said, these things will change over time. Uh, but X is not representative. I will never dialogue on X. I respond— I post on X, but I never respond to the MAGA extremists because it's mostly MAGA. So in any country, you can find 10, 20, maybe even 30% of the population in various extreme groups. Left, right, Hindu, Muslim, yeah, it doesn't matter. There are tribal constituencies in every large political society, and X represents a very, very disproportionate number of MAGA extremists. There's almost no reasonable people left on X, so X isn't what America is about. It's like you're reading a religious zealot constituency. Right.

And could I ask you, you know, just a sense, because you are, although you say you are, of course, a middle-of-the-way thinker on politics, you've also been a major donor. And of course, money talks in US politics. Could I get your sense of where you see this going, where you see the country going in terms of its politics in the next 5, 10 years. Because, you know, what does happen on your side of the pond, or your side of the ocean, shall we say, does impact what happens in India. For example, one example is these 50% tariffs that have been levied on India that has caused, you know, significant concern in India about its exports to the United States. There's a sense that America has become so unpredictable, so unreliable. Talk to us a little bit about where you see politics going in America in the next 5, 10 years, let's say?

Yeah, look, I'm not smart enough to predict the future of politics. Uh, I just don't feel like I'm qualified to predict it. And to be honest, I don't think anybody is qualified. How the dynamics of politics will lead to phenomena, and it's phenomena that take over, amplified by things like the internet. Um, so I, I would say it's very hard to tell. Uh, it's very likely even in, in this year's elections and the elections in 2028, we will see more balance between the left and the right, hopefully much more in the middle emerge over time. And have sensible policies and sensible— and some sense of ethics in politics, which Trump doesn't have.

Vinod Khosla, my last question to you. We talked about two stories, the artificial intelligence story, story about emerging technologies. And we talked about the second story, which is President Trump, both extremely disruptive to the world as it exists today. —people are going to be seeing this, people who are mid-levels in their career, 40-something, people are going to be younger, just starting out with their jobs, just starting out with university, perhaps even in school. How do you prepare yourself for the world that's coming? Is it skilling? Is it better education? If on an individual level, to people who are watching this, how do they prepare for these disruptions?

At the individual level, And I think the situation in India may be a little different than America. In the world of AI, where AI is good at everything, I think people have to have a much more general education and the ability to think and participate in, in, in thinking. So I tell university students, Don't do just electrical engineering. Do electrical engineering and neuroscience and some medicine or medical courses, uh, some language courses. Do a diversity of things in your education because in the world of uncertainties, the people with diverse capabilities will be much more able to adjust what they're doing and be much more dynamic about their careers. Even the notion of careers, it's not clear to me, will remain 15 years or 20 years from now. So diversity and of learning and education and the learning how to learn are probably the most important skills in a world that's becoming more, much more dynamic and with much less predictability. Right.

Vinod Khosla of Khosla Ventures, also formerly of Sun Microsystems, talking to us about AI technology, how it's going to change the world economy, and also the rise of President Donald Trump and his politics. Mr. Khosla, thank you so much for joining us here on The Hindustan Times podcast. Thank you very much.