#13: AI, Work, and the Future of Life: Lessons on Success, Resilience, and Reinvention
52:16 Share

The jobs we need to do will be done but don't want to do will be done by an AI. And then humans will do jobs they want to do or pursue other things they want to do. So I think it will be truly empowering for human beings after a chaotic period. Welcome to Divot, a community for people trying to make their mark in the world where each week I'm interviewing some of the best creatives in business, tech, sports and entertainment to learn how they've made their mark.

You can watch all the episodes on YouTube, Apple Podcasts, Spotify, X, or go to divit.org. - Divit. - This episode is brought to you by Salesforce. - Today, we're back on Sand Hill Road in Menlo Park, California. We're meeting with entrepreneur, technologist, and futurist, Vinod Khosla. Vinod is the founder of Khosla Ventures,

It's a firm focus on assisting entrepreneurs to build impactful technology-based companies. He's also the co-founder of Daisy Systems and Sun Microsystems, as well as being a venture capitalist for almost 40 years. Vinod has not only backed some of the most transformative companies of our time, but he's also been an advocate for big bets like AI, healthcare, and climate tech.

In this conversation, we're going to talk about AI, how to maintain excellence for decades, and what he'll be doing in 2050. Hope you enjoy it. When did you first start thinking about AI as something that would have a big impact on the world?

Well, I have an interview with the New York Times, I think it's in the year 2000 or so, somewhere around then, when I said AI will have us redefine what it means to be human. I was purposely not very specific, only because it sounded so weird. And these things are outliers, so it was weird. But I was pretty convinced AI was going to have a large impact on society.

But it was hard to define back 24 years ago, 25 years ago. Yeah. You made your first investment in deep learning around 2014. It didn't become the biggest company in the world. What pushed you to keep investing in space in spite of that? You know, you either believe in things or you don't. And I'm a person who has beliefs and like to invest.

Whether a particular company is successful or not, it can be unsuccessful for many reasons. And Jetpack was the company, I think. I'm forgetting the name. But it was inspired by ImageNet in the image library. It wasn't successful. It got sold. So we didn't make much money on it. But it didn't change my belief, which I had held for a while, that deep learning was going to be important.

In late 2018, you invested in OpenAI. As you look back on that decision, would you put it more as a team investment, a technology investment, or a market investment? Well, there was a lot of reasons, but a little bit of history. I talked about my 2000 belief system interview in the New York Times. But in 2012, January of 2012, I wrote two blogs. One was called Do We Need Doctors? Another one was called Do We Need Teachers?

And those are on TechCrunch. And they were clear indications, I believed by then, that AI would clearly do the functions that most humans do broadly. And so when we invested in deep learning, it didn't work out. It didn't change my belief. It said that particular company didn't work, but AI will work.

So when OpenAI came around in 2018, I was pretty excited because the team was right. It was a fast moving entrepreneurial company, high velocity of innovation. All the things were right. And they were going after building AGI, which sounded ridiculous to everybody except me.

I had this relationship with Sam Altman going back to that same period where you were investing in YC companies and participated in a lot of the batch funding there. Is that where you met Sam? Did you meet him before that with his first company? I met Sam before he went to YC when he was still doing Loopt, which wasn't very successful, but I was really impressed with Sam.

I thought I'd like to invest with him. We kept up a strong relationship when he was at YC, but when OpenAI came around, it was the kind of really ambitious, big swing project that I personally love. Technology driving real innovation, real disruption, and doing a lot of good. And I think we were both aligned on the mission of technology benefiting humanity.

And so that was sort of meeting of the minds. Do you think AI development is moving too fast, too slow or at the right speed? Nothing is too fast in AI development. I wish it moved faster. Why do you say that? There's no reason for it to go slower. Now there's different parts of it. The technology should move as fast as we can move it. And I believe, and this is surprising to most people,

ChatGPT is just about two years old. In the next two years, we'll see more progress than we've seen in the last two years or the previous four years. So we'll see an acceleration of the rate of development of AR capability. Then there's the separate question of how it gets applied to areas. In there, practically applying it, we'll have to figure a lot of things out.

Most people who are using AI today don't even know how it works and what it's good for or not. They're trying to apply it and experimenting with it. And experimentation is a good thing with new technologies. I would say in corporate America, more than half the people who are talking about AI don't know what AI is.

And by that I mean they can't tell the difference between the old machine learning data analytics technique from what's a real AI technique. So very dramatic differences.

in level of understanding of where AI is, how to use it, what works, what doesn't work yet, what else do you have to do to make it useful? I think that's a big discovery process the next three, four, five years. You've said the rate of technology development in AI is going to speed up over the next two years versus how quickly it's moved over the last two years. There's some evidence that the rate of improvement of LLMs

Can't possibly keep up from where it's been. Do you think that improvement comes outside of that space just in the on the edges around it? Or do you think that the actual models and what the models can do are still have a long ways to go to improve? So the nonsense you hear comes from the press's need for headlines.

And every time somebody has a belief system, the press wants to write about something that's contrarian, otherwise it doesn't make good headlines. So I would ignore what the press says. They take snippets of data and report it as a major headline. The fact is, even since that particular headline started last August, September, there's been massive progress. That's not to say it's always the same technique.

new things keep getting added on and will keep getting added on. So, you will see rapid acceleration.

in this and I'm pretty excited about that. Does AI accelerate robotics adoption in a way that we haven't seen in the past? Does it enable things that haven't been possible? Absolutely. AI, you should think of as general intelligence and that's sort of relatively soon. Whether it's in one year, two years or three years, it's not five or seven, which is shocking to most people. Most people don't believe it.

And I define AGI as being able to do 80% of all jobs humans do. It won't get adapted that quickly, but the capability will be there to have an AI accountant. But not every accountant in the world will be replaced by an AI in that timeframe. That's a great startup idea. We are doing that in one of our companies.

just an example, or AI structural engineer, or an AI oncologist. All these capabilities will be possible, and then you'll have to put the frills around it. So that's the meat and potatoes, and the gravy comes iteratively, slowly, as you adapt it to real-world situations. Sometimes, for example, the workflow around AI may take more time, but is critical to making AI productive.

Most people who used AI and haven't seen results is because they haven't gotten the rest of the system, like workflow right. Most of the people who aren't getting productivity, and you hear these stories, are because they're doing it wrong. I also think there's a fundamental difference in the world of AI. Mostly you hear about co-pilots, which means AI helping humans do a job.

I think the much more interesting and radical approach is, and that doesn't work very well because the humans use it. In software development with technologies like Cursor, people are seeing productivity improvement. But they're 20% improvements, not 200% improvements or 500% improvements. When you have to modify human behavior from the way they've done things, you're going to see small improvements, still huge improvements.

implications for productivity. But when you do the work that these humans are doing, instead of helping them do the work, so instead of helping an accountant do their work, if you just build an AI accountant, you're going to see hundreds of percent improvement in productivity. And that's why you start to hear about some of the AI companies that are doing the whole work.

charge not $20 a month or $30 a month or $100 a month, but $1,000 a month or $2,000. And yesterday there was rumors that things open AI may charge $20,000 a month for because you're replacing a PhD worker

that has PhD degrees in physics, chemistry, math, everything in one entity. Well, they're really, really valuable. And so I think doing the work will be a much faster acceleration of AI adoption than helping humans do the work they are. We may go through a safety period where all of the AI's work is checked by a human being.

a human accountant. So in a CPA firm, that's still the truth. They hire a fresh grad out of a local college and they make him a junior accountant, but then there's a senior accountant who oversees all the work they do. For example, when they're auditing a company or closing the books of a company. Now think of the AI model as initially, and for the next five years, this is more true than not.

AI accountants or AI structural engineers or AI software developers will be like junior interns and every senior programmer or senior structural engineer or senior doctor or senior accountant will get five AI interns. They may even have some human interns so they can compare the two in their performance.

And they will oversee the work, so we have the safety that we need in some of these systems. But these systems will grow up, and at some point, they don't need the oversight. And each company, each organization will decide for themselves

when they stop the oversight and just let the AI go and do all the work unsupervised, which will happen. I've heard you say that most of the change that AI will bring will come about outside of tech rather than inside of tech. What do you mean by that? Well, take a small business owner. You have a clothing boutique completely out of tech.

Can they do the sophisticated inventory management that Amazon can do? Can they do the marketing skills of a marketing agency they can't afford to hire that Amazon can? Can they close the books as well or manage their cash flow as well? They don't have the expertise. They just love the product they sell and that's their business.

all of these other functions at the level of performance in Amazon or Google or Procter and Gamble can muster will be available to them as an AI that does all these jobs for them in an affordable way because they can't afford to hire a marketing person and pay them $500,000 a year.

like Procter & Gamble can or Apple can, well, they'll get that expertise bundled into all services for small business. We'll help you manage business other than the things you want to do yourself because you love it. So that's a good example of-- - That's a good example. And I think if a lot of people outside of tech

are trying to figure out how does this really impact me? And I think small business owners, this is a great way to look at it. You have so few resources, you have low margins. So if you can supplement that with good technology and there's an easy interfaces and ways to interact with that, I don't have to be Silicon Valley person to use these tools, which AI is

and specifically ChatGPT and some of the other LMs have been really good about the interfaces and making it easy. This could open up your business in ways that weren't possible. And with tooling, does it mean too though that since this is like, it's a whole new shift and it's not just, as you said,

you're not just trying to improve existing workers and jobs and maybe tools by 20, 50, 80%. We're trying to do a sort of exponential shift and change. Does that necessitate everything coming with new tools, every tool?

SaaS application, every enterprise tool? Is it on notice that somebody who comes with a better interface and AI first sort of solution could disrupt it? I think many of the incumbent applications like SAP or a Workday will stay or Salesforce will stay where they are, but they mostly won't be used by humans. Here's a model you should think about. I, as a human user,

want to just get my work done. I don't need to look up what is somebody's salary, for example, by learning how to use Workday or how many days they took leave or pick your favorite example. I'll just say, tell me about this person, how many days did they take off?

not manipulate a menu and know which button to press and which menu item to look for or how do I add a row on the right in Excel. You'll just say add a row on the right in Excel or do me an application to do that. Or I saw a beautiful example just yesterday, the day before yesterday, where they took a large investor and said to the AI, build me a website. First go research this investor.

everything they say, do, and then build a website that's tailored to them, with their company name, their employees, whatever they could find. And it did. End of story. You didn't need programming skills to do that. And that's what will happen. And a small business owner, like a boutique in Carmel that sells clothing, might say, build me a beautiful website.

And I don't want to talk technical. I just want a beautiful website. Here's the clothes I want to show off and I want to change them every single day and I want my website to evolve. And maybe even further, I have these 100 top clients and when they come to my website, I want them to see something specific to them based on what they've seen.

So think about a small boutique in Carmel says, "Hey, I have a hundred important clients that buy, pick a number, $2,000 worth of clothes for me a year. And that's $200,000. And then I have 10,000 other people who buy $200,000 also worth of clothes. When these important people come, build me a website based on their buying history that they see other clothes like that

Extremely customized. So every one of their major users gets a personal website. And this boutique owner, without having ever learned programming or even gone to college. Can deliver that kind of experience without relying on something. That's the way AI will impact us. And frankly, this vision is pretty close to visible, doable today.

It doesn't take a lot. One of my favorite examples of investors is Replit. We also invest in Cognition that builds a whole programmer. But Replit, you can go to their website and say, "Hey, I have a six-year-old son who was asking about the Earth's planetary system." You can tell Replit to build your model and illustrate the whole thing in real time without ever having Dutch-Dutch programming.

That's now possible, people are doing it every day. It's just most people don't know this is possible, but the people who know it's possible are doing it. So as they say, the future is here, it's just not evenly distributed. I don't know who said that, but... I've heard you say it, maybe you said it. You mentioned that these giants, you didn't think they'd be replaced.

Today, many of these tools, they cost hundreds of thousands of dollars, millions of dollars, and you have to hire additional people just to run them. You can't even use them out of the box. And that's already being disrupted by product-led growth kinds of tools and more user-friendly tools. But you think those companies, they are innovating. They will innovate fast enough to maintain their leadership position. Some will and some won't.

You know, you look at old history in any industry where there's massive change, some keep up and some don't among the incumbents. Almost always,

The innovation is driven by a small company. You know, I'm hard pressed in the 40 years of dealing with innovation and innovation only to find one example, one example in 40 years of a large change that was driven by an incumbent or a big company or an expert. Almost always it is some entrepreneur driving this change.

And that's why entrepreneurship is such a great thing, whether it's

You look at Airbnb. Could somebody from Hyatt or Hilton have done that? Look at Uber. Could somebody from a taxi company or Hertz or Avis done that? Or take SpaceX and Rocket Lab, rocket launch companies. Could somebody from Airbus or Boeing or Lockheed have done that? Not a chance. Retailing, it wasn't done by Walmart or Target. It was done by somebody who knew nothing about retailing, Jeff Bezos.

So across the board, I can't think of one large innovation, societal innovation, incremental thing. Intel can grow from a seven nanometer semiconductor process to five because it's fully expected. But these orthogonal out of the box ideas, not possible from incumbents. You've recently tweeted out a picture of

market share between ChatGPT and the other LLMs with ChatGPT being the overwhelming clear leader. Do you think... And growing the most rapidly. Do you think OpenAI's first mover advantage here is too much to overcome or can these other tools, which some of them have very impressive technology, can they make up the gap? Well, depends on what you think they need to make up the gap on. AI is such a large space.

You know, I remember a story from the 80s when I was at Klein and Perkins and I think Lotus and Symantec, if you remember those names from the early PC application days, thought they were competitors because they were in the PC software business. So when you put on a tunnel vision about what the PC business is in 19, pick your favorite year, 1985 or 86,

That feels like you're in the PC software business. In fact, one was in the spreadsheet business, the other was in the utilities tools business, and there were 20 other segments like that. I think AI will do the same thing. Because it's so large, there will be more than one winner. It won't be just OpenAI, but hopefully OpenAI is the biggest winner. Will AI replace...

the jobs that you and I are doing 20 years from now? It will be able to replace those jobs. The first thing to remember is most jobs on this planet today, in America today, let alone the planet, are not jobs really that I'd call respectable. If you are a farm worker and have to bend over in 100 degree heat picking lettuce for eight hours a day,

30, 40 years, or you're an assembly line worker at GM mounting a tire on a car eight hours a day for 30, 40 years, that's servitude. It's not a job. They have to do it because they need to make a living, but it's not something we aspire for in humanity. So I do think the jobs we need to do will be done, but don't want to do will be done by an AI.

And then humans will do jobs they want to do or pursue other things they want to do. So I think it will be truly empowering for human beings after a chaotic period. Let's say one of those jobs was your son was on the trajectory for one of those jobs or your dad was or maybe your grandfather's. What do you tell these people who are on this path? What should they do now to...

prepare for something else or or first i think for young people kids in school pre-college i would say get as diverse in education and as diverse a set of intellectual interests as possible and over time let's decide what you love let me put it this way

Today, I've seen people give advice to young kids in school and say, "Follow your passion." And I always, if I'm there, I say, "That's really bad advice." Because that would be okay. You'd just be skiing all day. Yeah, I'd be skiing. But you know, this young kid, let's say they want to be an artist. The chances are 95% they'll also be doing a restaurant job serving tables in addition.

5% of the artists will make a living in art, 95% will be doing it and struggling with it and having a side job like a restaurant or other things. In the new world, that will become good advice, then you can follow your passion and you'll be able to make very explicit trade-offs. So I'm excited about that because it will truly free up humanity to be human beings as opposed to

just trying to service their mortgage or their car loan or... Just survive. Just survive. Which is more than 50% of the people in the US today that are working. What was more difficult to get into, the Indian Institute of Technology, Carnegie Mellon or Stanford? Or the Indian Institute of Technology by a mile. How come? Because it's a national exam you have to compete in. And I hear the latest numbers are like

It's much lower percentage that get into the IITs in India than get into MIT by a long shot. It's five times easier to get into MIT than into the IITs. Do you still, do you sometimes wake up?

from nightmares of thinking about the exam from when you were a teenager? I don't quite remember that far back. When did you decide that you wanted to build your career in the United States? Well, it was very clear I wanted to start a company after reading stories of Annie Grove starting Intel.

in this country. Andy was a Hungarian immigrant who came to this country, started Intel and I said, "Whoa, isn't that cool?" And tech is cool. So it'd be great to be an entrepreneur and fell in love probably at age 16 with the idea of starting a company. Once I had that and I looked at the environment in India, I said the only place in the 1970s I could start a company was in Silicon Valley.

It just wasn't an accepted thing in any other part of the world. And there was no mecca for entrepreneurs than Silicon Valley. You started

two major companies here in the Valley, Daisy Systems and Sun Microsystems. Do you remember, I've heard a little bit about Sun Microsystems early growth rates and revenue. Do you remember how fast it was growing in the early days? Yeah, in 1980s, early 1980s dollars, I think the very first year, because the hacks we did, I think we sold like seven or eight million in those dollars, which would be like

seven, eight times that today. That's pretty phenomenal for first year sales. The second year was about 38 million and the third year, if I'm remembering right, I could be wrong, was well over 100 million. So it's like getting to a billion in three years in today's terms and that's pretty phenomenal. But we didn't know what we were doing. We were making up because there were four 25-year-olds

I mentioned to a prominent Silicon Valley entrepreneur that I was meeting with you and they shared this story about either when you were building Daisy Systems or when you're building Sun Microsystems that there was this deal with a company called Computer Vision that you were either going to lose or you lost it and you went and chased this deal. Could you just tell us a little bit about that? Do you remember that?

I was CEO and I'd hired a very senior East Coast executive called Owen Brown to be president of Sun. I was relying on him being much more experienced. He played golf, which I didn't. He'd have martinis for lunch with business guests. You have the usual East Coast establishment thingy. I thought he'd close the account.

And I didn't know much about, you know, in the early 80s, if you had an Indian accent and if you were young and people didn't take you seriously and you could barely wear a suit and your cheap suits and wasn't really accepted that a CEO of a large company would do business with you.

So I'd sent him and of course we didn't almost lose the account, we lost the account. We were formally informed the account was lost and that's when I jumped in and said I'm taking this over. No experience in selling or anything but I decided to go solve the problem. Long story of how we did. I won't go into that here but

I think there's a whole Stanford business, Harvard Business School case written on this particular incident. - These are the kinds of things that you did as a builder, as a creator, as an entrepreneur. Do you have these same expectations of the people that you fund and invest in that they do these kinds of things? - So essentially, I was too naive to know what was not possible. And that was a huge advantage for me.

I just didn't take no for an answer. So I do expect entrepreneurs today to be as persistent, as resilient, lots of ups and downs, and just power through them.

There's no mental health support needed. You just power through these things, which is what I did. Frankly, I didn't know anybody else in this country very much. Like I didn't have family or support or anything. Like I had nobody to talk to. I've never had a mentor. And I just powered through things and didn't have the luxury of commiserating with anybody or with myself. It just, you power through it and really great entrepreneurs do that.

Even today. When you got here and you didn't know anybody, was that terrifying, scary? Did you think about going back? No, never. The boats were burned, you couldn't return? Well, I was where I wanted to be doing what I wanted to do, which is start tech companies. And so I was, even in the worst of times, I was always...

Just so focused on my goal. I didn't think about anything else. In fact, there's a presentation I did in 1986 that's still on our website today. It's in PowerPoint or But there was no PowerPoint back then it was done in overhead projectors Yeah, most people don't know what I'm talking about. But these are acetate files you put light through and put your shone it through a lens on the screen and

It's amazing. I have to explain what a projector was back then. It was a completely mechanical device with a light and a lens. That was the only two things you needed. But this presentation is called the entrepreneurial roller coaster. And the tagline is where the highs are high and the lows are low. In every project,

entrepreneur today has to go through the same highs and lows. The lows are really depressing and those who give up will surely fail. And those who don't give up and keep trying have a good chance at succeeding. In fact, it's funny, my son's an entrepreneur and he just sent me a text. I should pull it out, but he saw this quote which said, "You should never stop trying

Because you don't fail till you stop trying but if you stop trying you've been definitively failed He said, you know, this sounds like something dad would say he sent it to my wife I do think that's a pretty key characteristic of entrepreneurs though. There is a time It's a more nuanced answer to know when to give up. Yeah You there's a quote of yours that I've had in

in my hanging on my office wall for almost 15 years and it's that success matters and failure is inconsequential. Yeah. And sometimes I think about it, I've read it so many times and yet when failure comes it's so painful. It's almost like it's so easy to say that and it's so hard to live it and internalize it. It is hard to live

This is the lows of entrepreneurship that I'm talking about on this presentation, which I later transferred to PowerPoint. But I'll make the following contention. You mentioned I started Daisy. You mentioned I started Sun. Do you know what else I started? Tell me. Nobody ever knows what I started. Myself and Scott McNeely, before we started Sun, three months before, also started a company in San Jose called the Data Dump.

Okay. But nobody knows that. Everybody remembers Daisy and son. And that's why I say failure doesn't matter if you're resilient to it, success matters. And that quote came exactly from this experience. Nobody remembers I did the data dump and failed that, you know, life wasn't all rosy. You've managed to stay at sort of the top of your field through multiple cycles, multiple iterations. You're on the Midas list.

You know early in your investor career. You're still at the top of that list today How have you been able to continue to evolve and change yourself? So, you know, I just turned 70 happy birthday That was January, but thank you You know, I feel like I have 25 more years of work to do if I follow Warren Buffett's example He's 94 now. No, so he's my inspiration. I

And my dad was very active and moving around at 95. So hopefully I have the right genes, right genetics. But two things I would say first, you don't retire when you grow old, you grow old when you retire. I've long believed that. And when I go to my business school reunion, I look around and say, all these people are retiring, talking about retiring. But I can understand it. They look really old.

And I'm like, oh, I feel like as energetic as when I was 50. Like, no question about it. And people here will tell you I probably work more hours than almost any of the young people. Yeah. Which is odd because usually. I spoke with someone that works here and they said that you work harder than anyone else here. Yeah. Why? Because it's so much fun.

You know, I get to work on AI in 2018 when nobody else believed it and seeing that vision happen. The same year, 2018, I invested in Commonwealth Fusion. I got to work on a problem that would, two large problems, that would probably save the planet and make it much better. That's a real privilege to work on these hard problems. By the way, invested in public transit the same year.

2018, which is a rich vintage for me, right? Working on problems like fusion for the planet as an energy source, AI that makes life easier for all humans, or public transit that decongests cities. These are such exciting problems to work on. I can't think of what I would do. I mean, golf seems so boring, so boring compared to this.

But you know, because of that, I literally switch my bag because I have so much reading to do every day. I now carry a bigger bag, one of those LL Bean canvas bags to carry my reading instead of my regular bag I carry into work. I'm a glutton for learning.

And I'm learning every day and learning from the best. The best people in fusion who are building fusion technologies come in and explain fusion to me. And I think I'm smart enough to mostly understand. But the same happens in cell therapy, in biology. I love biology too. And we're just investing in a foundation model for developmental biology of how does a stem cell become an islet cell or a cartilage cell.

But from fusion to that and being able to learn about these things and then use them in a way that's useful for society, it's just like such a privilege. Why would I want to retire? What would I do? Play golf for four hours? Well, you could ski more. You could own a sports team. I still do ski. You could own a sports team. You could travel.

Anywhere you want you could do most parts of the world. Why would I want to try? I don't love traveling anymore Well, you because I've been to most parts of the world. I've had most of the experiences I want probably had more aesthetic experiences than most people on the planet I believe you know I've gotten to the bottom of the ocean 10,000 feet under to the volcanic points in the Galapagos and

Wakes are more exciting than most people have done. I've dived with whale sharks. You think about it. I've probably done it. I've heli-skied in Alaska. I've done all these things. And I still enjoy many of these things. But what I really enjoy is more learning. And one of my beautiful day...

go two hours skiing, ski 20,000 vertical feet in two hours and then spend eight hours reading and learning and then have a glass of wine in the evening with the family. That's the perfect day. That's a perfect day for me. So you don't think you don't you don't regret how much you've worked? No never. Yeah. Do you wish you'd worked harder? I've always worked as

as much as physically possible given family and kids. One of the things I did prioritize when my kids were growing up, I had a rule, I would always do 25 dinners with the kids every month, which with my travel schedule was really hard. But I not only had that goal, my level of achievement of that goal, how many dinners I had with the kids were reported to me month end every month.

So just I got other metrics on my performance. I measured that number. It was reported in my my assistant then was authorized to if I wasn't going to make my goal, put me on a red eye and get me back here by dinner the next day. I was going to New York. I'd be on a red eye, do the work and be back here by six, seven o'clock to have dinner with the kids again if I wasn't going to make my target.

So I've been very clear what my priorities are clearly being family first and dogs. I'm a huge dog fan. And then it's learning and startups are great way to learn new stuff. We both have four kids. I'd love some advice from you. My oldest is 15. My youngest is seven.

how can I be the best father possible for four kids? First presence matters. But there's a pretty important age from 12 to 18 with kids. I have a very simple advice to most parents who have kids who are entering that. At age 12 or 13, have your kids make 10% of the important decisions and you make 90% of them.

like should they get their hair dyed purple. That was an actual incident with one of my kids. To them making 90% of the important decisions by age 18 and you help them with 10%. Because at age 19, they're in college, you have very little supervision. They should feel empowered to make those decisions, but they should also learn how to make decisions.

So whenever you can get them to learn instead of edict things, you're going to be better off and they're going to be better off adults and your relationship will be much stronger with them. And I'm really proud of probably the biggest achievement in my life is my relationship with my kids today.

None of them decide to go away from home for college. They all four went to Stanford 10 minutes from home. They are happy to come back all the time. And they're still at home most of the most weeks. I have kids at home. It's beautiful, which is beautiful. So that's a pretty important thing. But very early on, we taught them how to make decisions, even at age five.

They had unlimited access to the number of cold Cokes they wanted to drink, chocolate they wanted to eat, fast food they wanted to eat. But every time they exceeded a certain metric, I just have a conversation with them. You think you're doing too many Cokes or too much chocolate? Having them think is really very important. Another trick I used a lot. They'd come to me and say, "Dad, I have a question."

I said, the answer is yes. I hope you have a sensible question. Imagine how many times they had to rethink the question they were going to ask me. So it sort of taught them how to think about responsible questions to ask or responsible things to do. And that was from age five on.

And it's very important to train them to be this way. I think it made a huge difference. By age 13, none of them were eating fast food. None. At five, McDonald's sounded pretty good, especially because of those toys.

But by age 13, none of my kids even touch fast food now. - Because they've learned the consequences of the bad decisions that you've made. - There's conversations and what is responsible eating. - Your kids don't resent how much you worked or when you were there or weren't there? I mean, 25 days a month is a lot, but do they? - Well, I was there then and I'd say I drove them to school most of the week.

At least three days a week, I'd drive them to school in the morning. So I was there for them and I deprioritized work during that phase, especially those critical years. I've only been married 18 years. You've been married 50. What advice would you have for me? I started dating when we were both 16, 55 years ago. Yeah. 54, 55 years ago. And I think I've been lucky because one party...

especially with the person with the intensity I have has to be much more tolerant and I was very lucky my wife was tolerant. I'll credit her not me. I was totally unreasonable at times. If you look back at these different stages, how do you think people looked at you when you were a founder versus how they look at you now versus how you think they'll look at you or you hope they look at you in the future? Well, I think

I would say, let's start with now. There's parts of my life that are hugely consistent. So I've been with my current wife for 55 years since we started dating. I've been in the same house since 1986. I haven't changed homes. My kids all were born there. They all grew up there. They still have their rooms. They all come back to their rooms. There's like extreme consistency there.

I would say they all say we have a great relationship and I still prioritize them needing my time. So there's that. And then the rest of my life, which startup I work on has so many ups and downs in our portfolio of companies. It's like highly variable. Even the area I'm working on changes every three or four years pretty dramatically.

There's no consistency in one half, but there's consistency where it matters. And I think that's a good role model for kids, but also it's been hugely satisfying. I was 30 when I decided to build a house I wanted to live in. Fortunately, I was well off enough to build a house I wanted to live in my whole age.

And I wrote up a spec of what I wanted in my 30s, my 40s, my 50s, my 60s, my 70s, like an engineering product spec and built a house to that spec. I actually was short-sighted and didn't think about my 80s and 90s or hopefully 100. But it's all worked out really well. Look, none of this is super foresight.

It's a lot of luck, but a couple of things, which is you're internally focused and know what you want and like and stick with it. Most people are driven by what others expect of them. I would urge people, if they want to be happy, to be driven by what internally drives them.

So don't worry about your parents wanting you to have a better title at Cisco or IBM or pick your favorite big company or Google. Do what you really enjoy as long as it's making a living for your family. That was the caveat. But you don't need a lot to support a family. You need basic stuff. You need them to be able to you need to be able to pay for college. Those basics.

I was almost always internally driven. And I was internally driven when I was 25. I wasn't driven by what others expected of me or what kind of car they wanted or what club they wanted me to join. I've never belonged to a club, which is sort of odd for people. No country clubs. Why not?

never enjoyed them. I'd rather go learn something during that time or go skiing, both of which I love hiking. I spend a lot of time hiking because I can think when I'm hiking. So hiking and skiing are my two passions and then family and I'm very clear what I like and I do think it makes a lot of when I was young in the 70s and 80s, I was just maniacal.

you know some of the stories you hear about Elon Musk I didn't understand why people couldn't sleep at work before I had kids and all that you just 100 percent 130 percent all in on your company all the time yeah no wasn't an option failure wasn't an option yeah you just kept going you know you see one of these NFL linebackers and um you know they run into the

Opposing team and their legs still keep pumping. Yeah, even if they're stalled and that's what I Reminded me of that kind of like mentality. Yeah, you never stop. Why did that change? It hasn't changed It probably still hasn't changed. I'm sleeping here though. No, I'm not sleeping here Family changes that age time. I feel like I live a pretty balanced life. Like I'd work a lot I still probably easily work 80 hours a week. I

But I enjoy all of it and I enjoy the work I do at home. But you know, I watch all Sunday, I'll watch NFL, which I enjoy NFL as a sport and do email at the same time. The last thing I just want to ask you is how will you measure your life? Many, many years ago in the early 90s, I lost a dog.

And I realized I was saying to myself, I should have done X or I should have done Y. And in my life now, on my deathbed, I don't want to say I should have done things differently. I end up doing the things I want to do, not what others expect me to do. Thank you so much.

So I think that this belief that you can just build a cool product and it's going to work is very naive. You need to basically wake up every morning in a cold sweat thinking about what is Lockheed trying to do today to take you down. And I think only the businesses that have super high degrees of paranoia are going to work in this market.