- anteayer
Categoría
📺
TVTranscripción
00:00:00We're going to be in an era of abundance that's so large, it's very hard for people to imagine.
00:00:07The simplest way to say it is the need to work will go away.
00:00:12People will work on things because they want to work on things, not because they need to work on things to pay their mortgage.
00:00:19All right. I am super excited to be here with Vinod.
00:00:22Thank you so much for doing this with me today.
00:00:24I really am looking forward to this conversation.
00:00:26Look forward to it.
00:00:26When you first walked in here, before we got started, you were saying, well, you know, the world's crazy and you've invested through many periods of time.
00:00:36Obviously, you know, there have been a lot of ups and downs, but you said the world's particularly crazy now.
00:00:42Can you share why you said that or why you're feeling that way?
00:00:45I've been doing venture capital for 40 years, so I've gone through every large innovation cycle.
00:00:50I've never seen a cycle like this to put an order of magnitude on it.
00:01:01Almost every job is being reinvented.
00:01:07Every material thing is being reinvented differently with AI as a driver.
00:01:14Probably the best way to describe it, if we look 15 years hence, and that's too far to look, we'd have to look back at least 50 years, maybe in the 1960s, to see the delta in change.
00:01:31And so we're going to see this large change in sort of such a short time.
00:01:39It's almost hard to imagine how society adjusts.
00:01:43And if I go even further, almost certainly within the next five years, any economically valuable job humans can do, AI will be able to do 80% of it, with a few exceptions like heart surgery or brain surgery.
00:02:02But by and large, 80% of 80% of all jobs can be done by an AI.
00:02:10And if that's happening in the next five years, the world is going to be a very different place.
00:02:17So I see this sort of like invention happening at a frenetic pace that is so large compared to even the dot-com phase when the internet first came and there was Netscape and TCP IP and all that.
00:02:36It's almost hard to imagine.
00:02:37So no matter where you look, everything's being reinvented in some fundamental way using often the innovations from AI and the capability from AI, but also a few other vectors like stuff happening in biology, stuff happening in fusion and other areas.
00:03:03So everything's up for granted.
00:03:05So I guess that's then leading to a frenzied pace of investment, companies being created, new ideas getting tried all at once.
00:03:16Is your overall read based on, on one hand, it's crazy and frenetic.
00:03:20On the other hand, everything you just said about, yeah, but like maybe a billion jobs are going to be done by an AI in the next couple of years or several years or something like that.
00:03:30Do you net out to being very optimistic about this moment in time for tech or do you think there's going to be a lot of carnage or how do you think?
00:03:38So my view, the next five years to 2030 will look like productivity improvements, the way economists would say productivity is going up, GDP is going up or accelerating on GDP, things like that.
00:03:55And then if I look to the 15 years hence, 2040 and beyond, we're going to be in an era of abundance that's so large, it's very hard for people to imagine.
00:04:08The simplest way to say it is the need to work will go away.
00:04:15People will work on things because they want to work on things, not because they need to work on things to pay their mortgage.
00:04:22Is that going to require us to change anything about like the social contracts to get there?
00:04:27Yes, I'll come back to that.
00:04:29What I want to say is somewhere starting in 2030, the changes are going to be so disruptive that it's hard to imagine how it sorts out and sort out differently by country, by region.
00:04:46Different governments will play different roles in allowing it to happen or not allowing it to happen.
00:04:53And so we'll see this very, very disruptive, almost dystopic type of AI is taking all the jobs kind of thing.
00:05:04But we will have enough productivity, enough goods and services produced to share broadly.
00:05:10Is what you're saying that basically the changes that are already happening will compound to a point where it'll now just be this big disruption?
00:05:18Or are you saying the tech by the year 2030 is then going to be so disruptive that it's going to create all these changes?
00:05:25In other words, are we already on that trajectory or you're just saying the tech's going to get there in a way that will be that disruptive?
00:05:31What I'm saying is tech is disruptive in some areas.
00:05:35And it will be everywhere.
00:05:36And it'll be everywhere.
00:05:38But almost no corporation.
00:05:40By 2040, these are Fortune 500 companies I'm talking about.
00:05:48Should run the way it's running, even remotely.
00:05:51I should say most Fortune 500.
00:05:54Yeah.
00:05:54One of my predictions is the 2030s will see a faster rate of demise of Fortune 500 companies than we've ever seen.
00:06:05It's actually pretty high anyway, but we'll see a rate accelerate dramatically.
00:06:12So let me put it differently.
00:06:15If you're a healthcare company, and I speak to a lot of healthcare company boats, you look at AI and say it's an opportunity and we'll get more administrative efficiency or something like that.
00:06:26But if I said to you, which I've said to a couple of boats, if all medical expertise is free, you have an unlimited number of primary care doctors, oncologists, gastroenterologists, mental health therapists, you name it.
00:06:45How would you redesign the healthcare system?
00:06:48Be crazy.
00:06:49That transition won't happen from existing companies.
00:06:54Almost somebody new will reinvent this.
00:06:58Now, there's the issue of regulation.
00:07:00The American Medical Association doesn't like that.
00:07:03I was going to say, there'll be a lot of people fighting against this who have a lot of power, who will be good at fighting it for at least some amount of time.
00:07:09For some amount of time.
00:07:11Yeah.
00:07:11But fortunately, we have a small part of our healthcare system, very small part, where payments are capitated, what's called a capitated system or Medicare Advantage, where you get paid a fixed amount to take care of a patient, roughly.
00:07:26Yeah.
00:07:27That world will be so much more effective because they won't care about reducing costs, which AI can do.
00:07:36Reducing costs for a hospital system is reducing their revenue.
00:07:41Of course.
00:07:41For a physician, reducing costs is cutting their costs down.
00:07:45Now, I think that the general model, I think, for the next five years, when I talked about productivity improvements, you will see almost every professional start to get interns working for you.
00:08:04So every MD, a practicing physician, will get five fresh MDs as if they are fresh out of Stanford Medical School that will work for them and will increase productivity, maybe provide a higher level of service to their patients, all that.
00:08:23But that then turns into something much more catastrophic over time.
00:08:28So, by and large, the next five years, I think we should assume that all professionals start to get interns working for them who are trained in their profession, whether you're a structural engineer or a salesperson who work for you, good, nice interns behave well.
00:08:50But at some point, these interns grow up.
00:08:53They will have way more expertise than the physician.
00:08:57And then it's hard to roll this clock back.
00:09:00It's going to go fast, too.
00:09:02It will go fast, except in some areas where there's regulatory or other contract.
00:09:07Take the example of the Screen Actors Guild last year.
00:09:12Anybody who signed up for the Screen Actors Guild thing will go out of business.
00:09:17But that change will be slower outside that system.
00:09:23People will be producing movies for $100,000, not $100 million.
00:09:30Yeah, right.
00:09:32The dynamics will be different by sector.
00:09:35If this does play out even directionally, the way you're saying, and now it's 2040, what do you think is left for people to be doing?
00:09:46Like, there must be some things left.
00:09:48Like, are people going to be still all employed or employable via creating their own businesses and pointing AI around?
00:09:59Or, you know, we have a lot of two-person, billion-dollar companies.
00:10:02Does it go that way?
00:10:03Or does it actually just go to, like, a lot of people are just going to consume a lot of leisure?
00:10:08I'm waiting for the first entrepreneur to call me and say, we can build a billion-dollar revenue company with 10 employees.
00:10:16You can definitely do it with 100 employees.
00:10:18We haven't gotten there yet, but we will.
00:10:22That company is already being started.
00:10:24I mean, somebody will leverage it tenfold again.
00:10:27There's a question of, will there be jobs that need people?
00:10:33If you have a billion bipedal robots in the 2040 timeframe, could be five years later, could be five years earlier.
00:10:42But let's say that that was one of my dozen predictions I made when I spoke at TED last year.
00:10:48Almost certainly, they're doing more work than human beings do labor today.
00:10:55Yeah.
00:10:56And they'll be smarter.
00:10:57And they won't sleep.
00:10:59They'll never complain.
00:11:00And you won't have silly notions like a primary care doctor and oncologist.
00:11:05Right.
00:11:05The same robot that, like, cleans your house is also your doctor.
00:11:08Yeah.
00:11:09Yeah.
00:11:09It's this super intelligence idea.
00:11:11I think people will have lots to do, but it won't be because they need to work for a living.
00:11:20Look, the first thing people forget, most of the jobs in this country or on this planet are not really jobs.
00:11:29There's what I call servitude.
00:11:30If you're a farm worker working for eight hours a day picking ladders, and you do that for 40 years, or you're an assembly line worker at GM mounting a tire on a car for eight hours a day for 40 years, those aren't jobs.
00:11:51They are servitude because you have to survive.
00:11:56And what a great thing if we could give those people their lives back and those hours of their lives back.
00:12:01What will people do?
00:12:03Humans will be humans.
00:12:04They'll still compete.
00:12:05We love sports.
00:12:06We love entertainment.
00:12:08We love being great at something.
00:12:13Whether you're even a mediocre artist loves painting for themselves, not because it can be recognized by a New York auction house, art auction house.
00:12:24So people will excel in what they do.
00:12:28You know, exporters, skiing, you name it.
00:12:31I love the idea of people who have a lot more time to care about their children.
00:12:39Actually not be frustrated because children are in conflict with work requirements or career requirements, working a lot more with their elders, parents.
00:12:51So there is going to be a lot to do.
00:12:55What's your confidence level that this is the future?
00:12:58High.
00:12:5980%?
00:13:00Higher.
00:13:00I have very high confidence this world happens.
00:13:06I think greater than 50% that it happens by 2050.
00:13:11The timing is hard to predict.
00:13:13But like you're 100% sure that it happens before like 2200.
00:13:17Oh, no question about it.
00:13:18Yeah.
00:13:19So I'd be shocked if it didn't happen by 2060, where we live in a world of abundance.
00:13:25Nobody has to prove they can buy a more expensive car than somebody else.
00:13:31That part, it may still be there.
00:13:35Humans like competing, including ego.
00:13:38But I do think something fundamental will change.
00:13:42A five-year-old kid going to kindergarten today is being told by their parents, go to school, study hard, get into good college, and then you'll get a good job.
00:13:51That won't be the paradigm.
00:13:53So my kid, my oldest kid's five, and then I have two other younger ones.
00:13:58If you're me, like, what are you thinking right now in terms of like what's important for like little kids right now?
00:14:04Because curiosity, for me, when I'm in a position where I can do anything I want, I could go play golf, which I don't love, or go sailing, or you pick your favorite.
00:14:19I'm mostly driven by curiosity.
00:14:22You know, I want to learn new things.
00:14:25So anytime I can learn something, I'll take a meeting.
00:14:28And I'm fortunate people will teach me, but in the future, you won't need somebody to be willing to teach you.
00:14:36It'll happen.
00:14:37For me, it's that.
00:14:39For others, it'll be producing the best music they can relative to their own bar.
00:14:46And clearly, people will still want to be like Taylor Swift.
00:14:50Will there be like the last person, one of the places in my head was going, was somebody still has to build the AI, build the robots, program everything.
00:15:01Will we get to the last person who needed to do that at some point?
00:15:04And then that's actually even that job goes away?
00:15:07First, high degree of uncertainty with respect to the configuration of what people are doing.
00:15:16Well, I don't believe they'll be working 40 hours a week for meaning in their life.
00:15:24Yeah.
00:15:25Unless it's something they really want to do independent of earning a living.
00:15:30Okay.
00:15:30So that we should be clear.
00:15:32And I just recently gave a talk last month on AI dystopia or utopia.
00:15:38And I, by the way, cover all the dystopian scenarios in it first.
00:15:42Oh, God.
00:15:43Before I get to the utopian scenarios.
00:15:45And so it's on our website and it's on the internet so anybody can find it.
00:15:52What's very clear to me is the dystopian views will be a choice society makes.
00:16:00The utopian view is going to happen anyway.
00:16:04The world of abundance, only a matter of when, not if.
00:16:08And within our lifetimes.
00:16:10It's amazing.
00:16:11And when I say my lifetime, I at least look at 25 years.
00:16:17I'm 70.
00:16:18Easy.
00:16:19More.
00:16:20When you're talking about this dystopia, I think what was implied in what you just said
00:16:24was sort of a self-inflicted dystopia where we just, you know, humans don't decide to
00:16:29organize correctly to take advantage of this amazing abundance.
00:16:32Absolutely.
00:16:33Do you worry at all about like the AI dumerism version of dystopia at all, where like the
00:16:38AI goes awry?
00:16:41Is that something you think about at all?
00:16:42Yeah.
00:16:42So let me answer the question in two parts.
00:16:46Certain societies will choose not to do it.
00:16:50We saw it with the screen actor scale.
00:16:51They essentially sign existing traditional studios up to a paradigm that makes them order of
00:17:00magnitude less effective in content generation.
00:17:03Because we're building an AI videographer in our portfolio.
00:17:09You specify, you don't just generate video.
00:17:12You specify what the lighting angle is and what the, how you pan and zoom the camera in those
00:17:19kinds of things.
00:17:21So the people who resort to traditional means will not do well.
00:17:28People who do do movie production, today we can't do a full movie that way, but that's
00:17:33only a matter of time.
00:17:35Right?
00:17:35Only a matter of time.
00:17:37Yeah.
00:17:37Today, almost all commercials can be done by an AI.
00:17:44Yeah.
00:17:44And the example you gave of, you know, reproducing a $100 million budget thing with $100,000,
00:17:49of course that does, you know, of course that's going to go that way.
00:17:52And we know Coca-Cola got a lot of flack last summer, last Christmas, because they produced
00:17:59some commercials with only with AI.
00:18:01And people were nitpicking, anomalies and stuff.
00:18:05Most people didn't know it was done by an AI.
00:18:08And they did great.
00:18:10And I'm glad they did it.
00:18:13Broke the ice.
00:18:14So there will be a flood of that.
00:18:17So there's that version of dystopia, displacement.
00:18:23Disruption's always fun for the disruptor, very not fun for the disrupted.
00:18:31And so one has to keep that in mind.
00:18:34We have to take care of people who are disrupted in this paradigm.
00:18:39The other end is sentient AI that kills us all.
00:18:45I wouldn't say I don't worry about it.
00:18:48I do worry about it.
00:18:49But I look at it as a complete basket of risks humanity faces.
00:18:57We've gone through an era where an asteroid hit planet Earth and most life was destroyed.
00:19:05So we know that was a risk.
00:19:08And risks still exist.
00:19:09An asteroid could hit the planet.
00:19:11More likely, a much worse epidemic than COVID.
00:19:19Yeah.
00:19:20You know, COVID only killed 7 million people.
00:19:23It's possible it could kill 700 million people.
00:19:26Totally.
00:19:26Right.
00:19:27In fact, I would say that's almost likely.
00:19:32So there's a whole basket of risks.
00:19:38Do I worry about sentient AI killing everybody on the planet?
00:19:42Yes.
00:19:43But more than I worry about President Xi doing nasty things?
00:19:48No.
00:19:48Yeah.
00:19:48So when it comes to AI, I want us to have the Western world to have the best AI so that
00:19:58we are not subject to President Xi or President Putin.
00:20:02That is by far the greatest risk.
00:20:05That makes total sense.
00:20:06And this is what the people who call to slow down AI research miss.
00:20:12Putin isn't slowing down.
00:20:14Not that he's very good at it.
00:20:15But Xi and China definitely can do this.
00:20:21You know, it's a world in which the, and you can't argue with it, they sort of say the
00:20:26end justifies the means.
00:20:28Yeah.
00:20:29We in the Western world care more about the means.
00:20:32Mark Andreessen made this point on the podcast too, which was basically the, so much of culture
00:20:38and production is going to live in how the AI is, you know, programmed.
00:20:43And if, if it's our doctors and our teachers and our lawyers and everything like that, you
00:20:48really want it to be.
00:20:49Well, we've seen it with TikTok.
00:20:51We did.
00:20:52Right.
00:20:52If you're on TikTok, you're more likely to be China inclined, favorably inclined towards
00:20:59China.
00:21:00That's culture.
00:21:01And President Xi had said in one of his speeches, not very long ago, the next battle will be
00:21:11fought on the internet.
00:21:13It will be a battle for people's minds.
00:21:16He was clearly thinking TikTok.
00:21:18So there is, you know, dystopic AI killing humans and all that, or AI in warfare, you know,
00:21:27weapons and robot soldiers and all that.
00:21:31That seems very important too.
00:21:32Yeah.
00:21:33That's very, very important.
00:21:35But it's even more important that we win the race to have social influence.
00:21:41If I'm giving all the entertainment to the rest of the world, if I'm giving all the free
00:21:48doctors, if I'm giving all the few tutors, my philosophy, political philosophy prevails.
00:21:55So I do think by 2040, the biggest risk we might face that I worry about is China using
00:22:04both good AI, cyber AI, warfare AI, but also socially good AI, like free doctors, to everybody
00:22:14on the planet and to embed their political philosophy.
00:22:18Yeah.
00:22:19I think that's real.
00:22:22It's probably the single biggest thing I worry about.
00:22:24But in some ways, it seems like politics, both sort of the way we think about it normally,
00:22:29but also just sort of broadly socially is going to get even more important than it is right
00:22:33now over time.
00:22:35Yeah.
00:22:35Look, we live in a capitalist system, but people forget capitalism is mostly by permission of
00:22:41democracy.
00:22:42Yes.
00:22:43In our system.
00:22:43If democracy sustains, I think we will adjust easier.
00:22:54Now, China has the advantage of using what I call Tiananmen Square techniques to enforce
00:23:02whatever they want.
00:23:03So it's going to be an interesting 15 years.
00:23:08And, you know, it's going to be interesting, but also critically important 15 years to what
00:23:16philosophy starts to dominate the planet.
00:23:18I want to spend most of our time continuing to talk about the future, but I have one historical
00:23:22event I want to ask you about.
00:23:24The tee up to this question is going to sound like, you know, just an opportunity to, you
00:23:29know, talk about being a great investor.
00:23:30But I specifically want to ask you about your OpenAI investment.
00:23:33And the angle I want to ask you about it from is to try to figure out what, if anything, we
00:23:38can learn from how you did that and how you decided to do it.
00:23:41I know that you were sort of like the first, you know, venture investor to do it and you
00:23:47did it in real size and something that probably at the time, a lot of people didn't want to
00:23:53do, maybe thought it was a science project.
00:23:55There was a lot of lack of clarity on the whole thing.
00:23:58I think that like, you know, now it's, you know, it is what it is.
00:24:02But at the time, it was probably an extremely hard bet to make.
00:24:05And so what I was hoping was to try to, as much as possible, play back the tape inside
00:24:10your brain to see what we can learn from that.
00:24:13So it was a very large investment, more than twice the largest investment, initial investment
00:24:19I'd made in four years of venture capital.
00:24:22Wow.
00:24:22So it was a conviction bet.
00:24:26But I want to go back to a different era and talk about this idea of people like to be
00:24:34in herds.
00:24:35You know, the herd mentality was crypto and a couple of others.
00:24:40And by the way, I've never called myself an investor.
00:24:45In 40 years, I always say I'm a venture assistant, health care nurse.
00:24:50It's a much better mental model for me.
00:24:55Not the investing is a side effect.
00:24:57So I can keep doing it.
00:24:59I'll give you another example that relates to this.
00:25:03While people follow herd instinct, I like to think about what are the fundamentals and can
00:25:11they happen?
00:25:13Fundamentals don't always happen, but they can.
00:25:16And in 1996, I invested, I started a company called Juniper.
00:25:21We worked on a business plan in my office for six months with the founder, Pradeep Sindhu.
00:25:28Why?
00:25:29Because the world was saying the internet would be a protocol called ATM, asynchronous
00:25:35transfer mode.
00:25:37So every major company had given up on TCP IP for the public internet and were only going
00:25:45to use ATM because the telcos were specifying ATM.
00:25:49I talked to the senior management of almost every telco and every single one said to me
00:25:58they wouldn't use TCP IP.
00:26:00I talked to the CTO of Cisco who said they have no plans to ever do TCP IP above 155 megabits,
00:26:10which is my home service, it's far below my home service for a public router.
00:26:17But I looked at the data, TCP IP usage had been climbing and I said, this will scale and
00:26:26every expert is wrong in the field.
00:26:30And we should come back to that if you remember this notion of experts.
00:26:34And the data said TCP IP was an exponential and we were near the knee of the curve.
00:26:42So fast forward to today.
00:26:45And by the way, I'd invested in TCP IP and it was a core part of what Sun did in 1982.
00:26:52And this was 14 years later.
00:26:54You see the exponential in TCP IP as a communications medium.
00:26:58And we bet heavily, by the way, we made $7 billion on a $3 million investment when almost no venture
00:27:08investment ever made even a billion.
00:27:11That's crazy.
00:27:12Sounds good.
00:27:132,500 X our money.
00:27:15Yeah, I'd like one of those.
00:27:17Yeah, you only need one, but hopefully OpenAI for us will do better to go back to your question.
00:27:23But fast forward, in the year 2000, I gave an interview, which is in the New York Times.
00:27:28So it's on the record that said, AI will have us redefine what it means to be human, which
00:27:35is this utopia thing we were just talking about.
00:27:38Will people need jobs?
00:27:41That was the year 2000 or 2001.
00:27:452012, I wrote two blogs.
00:27:47One was called, Do We Need Doctors?
00:27:50Under the premise, AI will do all the doctoring expertise that we need.
00:27:57This is 2012.
00:27:582012.
00:27:59It's in TechCrunch.
00:28:00You can look up.
00:28:02And the second was called, Do We Need Teachers?
00:28:05Idea that all tutoring will be done by an AI too.
00:28:08So those were 2012.
00:28:10What gave you confidence at that point that that was going to happen?
00:28:14Because this was before the breakthroughs was transformed, way before all that stuff.
00:28:17So, and this is where I was getting to.
00:28:21What I looked at was the best talent was going into AI.
00:28:26You went to MIT or Stanford or Caltech or wherever.
00:28:31People wanted to work on AI.
00:28:33Toronto, of course, a pretty important place in this history.
00:28:39Best talent was going.
00:28:41In 2016, I gave a talk at the National Bureau of Economic Research.
00:28:47I should probably release those slides.
00:28:49My first slide was the top 20 occupations in the U.S. by a number of people.
00:28:55And I said, I can't see which ones of these can't be displaced.
00:29:00Now, this was to a bunch of economists.
00:29:02That was 2016.
00:29:03I was convinced.
00:29:05But in that talk, I had human performance and current AI performance.
00:29:13And most people were making fun of AI because you'd tell and ask an 83-year-old man if they were pregnant.
00:29:22It made silly mistakes.
00:29:23I looked at the rate of progress and the rate of talent influx.
00:29:30Those two things convinced me the moment for AI was soon.
00:29:35Now, this was 2016.
00:29:37So, by 2018, when Elon backed off funding OpenAI,
00:29:45it committed a lot of money that he didn't meet his commitments for.
00:29:49And when Sam called me, I looked at that and said,
00:29:54I have the right team.
00:29:55I was very worried about China.
00:29:57I thought there was only two other teams.
00:29:59One was inside Google, which was moving very slowly.
00:30:02And another in Baidu, which was stealing Google's people
00:30:06by setting up an office down the air and stealing the expertise.
00:30:10I'm very hawkish on China.
00:30:12And OpenAI needed to exist as a separate vector of AI progress.
00:30:17And it had the right team with Sam running it.
00:30:21When Sam called me, it was almost a no-brainer.
00:30:25Because you were such a prepared mind for this by that time.
00:30:28Yeah.
00:30:28I was convinced the impact would be large.
00:30:34Now, I had no idea when there would be breakthroughs.
00:30:38Obviously, I couldn't predict transformer models,
00:30:41even though the paper was out, was going to be that large an impact.
00:30:46In fact, transformers wasn't what OpenAI was working on.
00:30:52So I wasn't this.
00:30:53This technique works.
00:30:55It was more like talent and the rate of progress.
00:30:59And I actually thought it'd be multiple techniques.
00:31:03I remember talking to Sam as I have a Lego block model.
00:31:07The more different types of Lego blocks you do,
00:31:10the more the commentatorial possibilities explode.
00:31:13I think it was following the data and not the herd.
00:31:17So was there a couple-year period where you were basically very primed
00:31:21that you were just waiting for the right team at some point?
00:31:24Like, it seems like between 2016 and 2018,
00:31:27you probably believed this pitch from just, you just needed the right team.
00:31:31Well, I was convinced of the following.
00:31:35AI would have a huge impact because the rate of progress I was seeing
00:31:39on these benchmarks against human performance,
00:31:42even though they were way down there and almost silly performance.
00:31:46But the rate of change was high and the rate of talent and flux was high.
00:31:51Elon backing off funding OpenAI was fortunate.
00:31:56I can't say I engineered it.
00:31:58I wish I had.
00:31:59Yeah.
00:31:59But it is following the data and not the herd instinct.
00:32:04Yeah.
00:32:05And frankly, 2018, when we made the decision, it closed in early 2019.
00:32:12Everybody thought AI was not a real thing.
00:32:16People thought it was a science project.
00:32:17It was a science project.
00:32:19OpenAI was not using transformer models.
00:32:22It was just my belief.
00:32:24You put the right resources and they'll figure it out.
00:32:28I told my partner, I have no idea whether it happens in three years or 10 years.
00:32:34But the IRR would work out.
00:32:37It was too important to not try was the phrase I used.
00:32:41Some things are too important to not try.
00:32:44And I often try those.
00:32:45At the same time, we invest in Fusion at Commonwealth Fusion.
00:32:49Fusion, again, started it from practically scratch because Bob was a senior fellow at the MIT Fusion Lab.
00:32:59This idea that some things in society are too important to not try.
00:33:04And I'd rather try and fail than fail to try.
00:33:08And most people, most of the time, just fail to try.
00:33:11And so, this combination of factors and knowing AI would have a large impact once we got it was what got me convinced.
00:33:21Okay.
00:33:21So, that's a perfect setup into now looking to the future.
00:33:24That construction got you to a big bet in AI.
00:33:29What are the other areas right now that you feel maybe like, you know, the same way you felt in 2015 about AI?
00:33:37What do you feel that way about now?
00:33:38You mentioned Fusion, but, like, what are a few of the areas that you feel this way about the future now?
00:33:43Well, first, related to AI, robotics will take a little longer, but I think we'll have the chat GPT moment in the next two to three years.
00:33:53What could that look like?
00:33:54A robot that doesn't need to be programmed.
00:33:56It learns.
00:33:57It isn't programmed to do tasks.
00:34:00And it's a generalized, you know, maybe physical robot in our spaces.
00:34:04Well, the world is organized around the humanoid form factor.
00:34:08And so, it's probably a humanoid.
00:34:10Humanoid is the only common enough form factor to get to high enough manufacturing volumes to get the cost down.
00:34:17Almost everybody in the 2030s will have a humanoid robot at home.
00:34:23Probably start with something narrow, like do your cooking for you.
00:34:26It can chop vegetables, cook food, clean dishes, but stays within the kitchen environment.
00:34:33And, you know, if you have a Prius, it costs $300, $400 a month.
00:34:38The robot will cost $300, $400 a month.
00:34:43It won't matter.
00:34:44And, you know, for anybody making more than a certain amount of money, they already have somebody helping.
00:34:51And so, this is help.
00:34:53Yeah.
00:34:53I think that's how it'll start in the home.
00:34:56But factories, farms, farm workers is a massive area that's mostly unaddressed.
00:35:05What part of the humanoid intelligent robot are we most limited on right now?
00:35:09Like, is it about the robotics?
00:35:10Is it like the intelligence?
00:35:12Is it?
00:35:12It's the intelligence.
00:35:14Clearly, we've seen out of China many, many hardware form factors.
00:35:21They're pretty damn amazing.
00:35:22If you see them boxing or running a race, it's pretty cool.
00:35:28But they're not learning robots.
00:35:31You change their environment and they don't do as well.
00:35:34Yeah.
00:35:34I think this intelligence that learns and adapts and doesn't need to be programmed to do a task.
00:35:42Yeah.
00:35:42If you walk a human in here and say, clean up, they'll know what to do.
00:35:49A robot needs to do that.
00:35:51Why is this not going to be like Apple?
00:35:53Like, and I know that, like, you know, I'm not saying I think it is, but like, why do you
00:35:56think that we don't already have the sort of prototypes of this out from the biggest,
00:36:02most well-equipped hardware companies in the world?
00:36:04So I've been doing innovation for 40 years and innovation only.
00:36:10I can't think of very many large examples that where large innovation came from somebody who
00:36:21was large or in the business.
00:36:23Yeah.
00:36:24Can you, can we go through some of them?
00:36:26Yeah.
00:36:28Retailing.
00:36:29You know, when we invested in Amazon, I was a kleiner in 96.
00:36:34Nobody thought you could compete with Walmart.
00:36:37You'd think Walmart would in a way.
00:36:39Yeah.
00:36:40Look at media.
00:36:40Was it NBC or CBS or was it little things like Netflix and Twitter and YouTube and Facebook?
00:36:51And nobody even knew they were in the media business.
00:36:54Uber.
00:36:55Did that come from Hertz or Avis?
00:36:58Car rental or taxi companies?
00:37:01No.
00:37:02Airbnb.
00:37:03Did it come from Hilton or Hyatt?
00:37:05Yeah.
00:37:05They have more rooms than Hilton and Hyatt now.
00:37:08Now, SpaceX or Rocket Lab, instead of Boeing and Lockheed, you'd expect to be able to do
00:37:15space launches.
00:37:16Yeah.
00:37:17Yeah.
00:37:18You know, my own history, I got enamored with the fact that when I was a kleiner, they talked
00:37:25about Genentech, started by an associate at Kleiner when nobody believed biotechnology was
00:37:31a field.
00:37:32No pharma company was doing it.
00:37:34Yeah.
00:37:35All of biotechnology.
00:37:36Yeah.
00:37:36We saw the cloud.
00:37:37We saw self-driving electric cars.
00:37:41Everybody had given up on it.
00:37:44How about cloud, actually?
00:37:44Because that was Amazon.
00:37:46So you kind of give them.
00:37:47Yeah.
00:37:48Cloud is interesting.
00:37:49I think I should say founder-driven companies have permission to try crazy things.
00:37:57Like Apple when Steve was there with the iPhone.
00:37:59Yeah.
00:37:59When Steve was there, they tried the iPhone.
00:38:02And if you go back to the press in 2007, everybody said a phone at $600 or $700 that has no keyboard
00:38:12to dial.
00:38:12Right.
00:38:13Like, what a silly idea.
00:38:15Yeah.
00:38:15Right?
00:38:15So it's founder-led companies that, in a way, self-driving would be 15 years away if Larry
00:38:25and Sergey hadn't said, we're going to do self-driving at Google.
00:38:28Yep.
00:38:28Or Elon had said, well, I'll do electric vehicles.
00:38:32Yeah.
00:38:32Right?
00:38:33So basically, Tesla should be much more likely on humanoids than Apple, for example.
00:38:38It is founder-driven and Elon is paying attention to it.
00:38:41So they're definitely a good candidate.
00:38:43But so are a number of other companies that have a lot of faith in the startup ecosystem
00:38:51of founders with very creative ideas.
00:38:55So that's an area worth delving into in a separate podcast.
00:38:58But that's an example of large innovation.
00:39:04So no matter where I look, so the cloud definitely happened.
00:39:09You know, when I was starting Sun, everybody said, there's these DEC WAX computers and IBM
00:39:16mainframes, and why don't you build a graphics terminal for them?
00:39:20And I said, no, I'd rather just replace those companies.
00:39:24Neither of those companies are in the computer business anymore.
00:39:28And frankly, it took five years to get DEC out of the computer business into mostly a customer
00:39:33support business after we started.
00:39:36Yeah.
00:39:36Which is pretty stunning to me as a 20-some-year-old kid.
00:39:41Totally.
00:39:41So repeating that pattern, right?
00:39:46There's zero chance a company like Siemens or GE could do fusion.
00:39:52Right.
00:39:52Zero chance.
00:39:54What's the mechanic for what?
00:39:55I, of course, agree.
00:39:56But what's the mechanic that's stopping them?
00:39:58Is it risk?
00:39:59Is it talent?
00:39:59Here's the first thing I would say.
00:40:01In these 40 years, in all the examples I've just cited for you, and I could give you more.
00:40:07None of them were done by somebody who knew the area.
00:40:12Experts are terrible at predicting the future.
00:40:17They extrapolate the past.
00:40:19Entrepreneurs invent the future they want.
00:40:24It's a bastardization of an Alan Kay quote.
00:40:27But, you know, so you imagine the world and you try and make it happen.
00:40:32And you're not biased by, what is experience?
00:40:37Experience is a set of biases that prevent you from making mistakes.
00:40:42And that's what experts do.
00:40:44And that's why very rarely do experts innovate in their area.
00:40:49Look, even the COVID vaccine came from two startups, Moderna and BioNTech.
00:40:57You'd think the pharma was very equipped to mobilize.
00:41:02Not a chance.
00:41:04So, you know, it's why the U.S. has done so well.
00:41:09People have permission to try.
00:41:12And if you try and fail, you don't ruin your career.
00:41:17In a big company, you'll ruin your career.
00:41:19To your question, right?
00:41:21You can't afford to fail.
00:41:24So, you start with, I can't fail, which means you can't take a large risk.
00:41:28My view is very simple.
00:41:30Most people reduce risk to increase the probability of success.
00:41:36I do the opposite.
00:41:39Start with, I want high consequences of success.
00:41:41I don't care about the probability of failure.
00:41:44I think that is like one of the best things about tech as an ecosystem.
00:41:46I think not everybody thinks quite like what you're saying.
00:41:50But in general, in tech, you can fail and it's okay.
00:41:54And I think even in other, you know, I worked in France and New York for a hot minute.
00:41:57And even there, it's just like, it's just not how it works.
00:42:00Jack Dorsey is great on paper.
00:42:03But when we invested in Square in 2010, the following was the situation.
00:42:11He had just been fired from Twitter.
00:42:14Twitter was a tiny company.
00:42:15He had just been fired.
00:42:16And he had like four failed startups.
00:42:19His history on Wikipedia.
00:42:21Like he had four failed startups.
00:42:23So, you know, traditional wisdom would say, don't back him.
00:42:29Was your mindset around risk always just increase the consequence of success?
00:42:33Or did that evolve over time?
00:42:35It was very much that from, you know, my first startup was go on, take on the computer business.
00:42:42Everybody big is dumb.
00:42:43Which is a little bit of an arrogant viewpoint.
00:42:47But if we hadn't had that kind of hubris, we probably wouldn't have attempted what we attempted.
00:42:56So I think hubris is an important ingredient of great entrepreneurship.
00:43:03Yes.
00:43:03We'll build an AGI, right?
00:43:06Like that's hubris.
00:43:07It's arrogance.
00:43:09Only I can do it.
00:43:10And so these are entrepreneurial characteristics that are very important.
00:43:15But also thinking from first principles, which is why experts do poorly in their areas.
00:43:24Because they have too many biases of what works and what doesn't or how to do things.
00:43:29While somebody new in the area reinvents the area.
00:43:33They figure it out from first principles in the new environment.
00:43:37And that's an essential characteristic.
00:43:41So, you know, one thing I always get asked is what do you look for in entrepreneurs?
00:43:46You know, I seldom look for deep expertise in their area.
00:43:51I do look for people who are thinking from first principles and learning rapidly.
00:43:56Even in YC batch, if I'm talking to a YC partner about some startup, the question that's most important to me is how much have they changed their plan over the last three months?
00:44:08Yeah.
00:44:08Rapid evolution, rapid thinking, rapid learning is what matters.
00:44:14So this went into all over the map.
00:44:17No, it was good.
00:44:18This is I love this, Eddie.
00:44:20But I want to keep getting sort of the other areas you're very excited about.
00:44:24That was robotics, which I agree.
00:44:27Humongous.
00:44:28If, you know, some ways that's like that's like the final thing to figure out.
00:44:32You mentioned fusion as a big thing.
00:44:34I'm very bullish about energy.
00:44:36Well, sort of many formats of energy.
00:44:38So I think there's two large formats.
00:44:41Fusion is one.
00:44:42And there's multiple approaches to fusion.
00:44:47But I'm also very excited about super hot geothermal.
00:44:53So, again, this ties to this risk idea.
00:44:58In geothermal, completely uneconomic business.
00:45:03You buy power because it's green, not because it's cheap.
00:45:08And people like cheap stuff.
00:45:10Turns out almost all geothermal in the U.S. is at 200 degrees or 250 degrees max.
00:45:18If you increase the temperature to 450 degrees, you get 10 times to 6 to 10 times the power,
00:45:25depending upon conditions, from the same well.
00:45:28That's surprising.
00:45:29For two reasons.
00:45:30One is Carnot efficiency goes up, but the heat content also goes up.
00:45:34So the energy content of the steam goes up and the efficiency percentage, so you get a multiplier effect.
00:45:41Don't need to get into technical details.
00:45:44But now it's cheaper than natural gas if you can get to 450 degrees.
00:45:49Nobody has to worry about super hot geothermal is green.
00:45:54It's just cheap.
00:45:56Wow.
00:45:56Right?
00:45:57And so why don't people attempt it?
00:46:00Because people have decided you can't drill at 450 degrees because drill bits collapse.
00:46:07Got it.
00:46:07So I sort of approached it differently.
00:46:10I said, what would it take to drill at 450 degrees?
00:46:16If I can solve that problem, everything else takes care of it.
00:46:20And we know there's 450 degrees.
00:46:22The most parts of the earth, it's only a matter of how deep you go.
00:46:26And so I set off on that problem.
00:46:29I think we are making great progress.
00:46:30So I think it's as powerful as fusion in supplying most of Western United States.
00:46:37Yeah.
00:46:38You can get that.
00:46:40Our location in Oregon, one location produced seven gigawatts of power.
00:46:47And I hope this year I can prove we can do all seven gigawatts competitive with natural gas.
00:46:55Wow.
00:46:55So without worrying, it works in the Trump era where he doesn't care, believe in climate change.
00:47:01It still works.
00:47:02And I love that kind of project.
00:47:04I'm also very bullish.
00:47:05Commonwealth fusion will do well.
00:47:07I don't know enough about Sam's project in Helion, but I'm glad diverse techniques are being allowed.
00:47:14But both of these are areas where you're like, the science is going to work.
00:47:17The impact will be enormous.
00:47:18Yeah.
00:47:19Yeah.
00:47:19It's worth doing.
00:47:20Solar also.
00:47:21Is that on your radar?
00:47:22We are doing solar.
00:47:23Well, we probably have in the Western world that China has some great solar technologies, the next generation of solar called perovskite solar cells.
00:47:34We're working on that.
00:47:36But there, the impact isn't as large.
00:47:39Yeah.
00:47:40Right.
00:47:40Right.
00:47:41And that power is already so economic that there's other uses for that power.
00:47:48So we are doing that.
00:47:49But, but look, I'm relatively optimistic.
00:47:55I see our path clear to getting to zero emissions by 2050.
00:48:01I think most of the technologies do that will be proven in the early 2030s.
00:48:09Is there any sort of a countervailing force with the fact that we're going to be consuming a lot more power via AI?
00:48:16People always ask me that.
00:48:18But I think if we get fusion right, I think very early 2030s, we'll get our first power plants.
00:48:26And that power curve.
00:48:27That power curve just catches up.
00:48:30Yeah.
00:48:30So there'll be this period between now and 2035 when these things aren't scalable yet.
00:48:36Yeah.
00:48:36But beyond that, I don't think.
00:48:37It should catch up.
00:48:38We should, powers will be cheap and plentiful and zero emissions and zero resources.
00:48:46And by the way, if you're creative enough, you can do this in every area.
00:48:52We're producing cement today.
00:48:55We are, we not only don't say like environmentalists do, let's go shut down cement plants every time.
00:49:05There's one in Cupertino that the environmentalists have been trying to shut down.
00:49:10My approach is very simple.
00:49:12They are emitting carbon dioxide.
00:49:13Let's capture the carbon dioxide and turn it into carbonates.
00:49:16We get more cement out of the same amount of limestone mine at a lower cost per ton.
00:49:26Now, what's there not to like?
00:49:28Do you have to care about climate to have that work?
00:49:31No.
00:49:32And we are in production on this now.
00:49:36So I'm sure cement is solved.
00:49:39Steel is the same way.
00:49:40I'm pretty certain we'll be able to produce steel with zero emissions.
00:49:47Or low emissions initially at a lower cost or similar cost to today's steel.
00:49:58I'm pretty damn excited.
00:50:00I think this is mostly solved.
00:50:04So, you know, climate is a very solvable problem.
00:50:07Yeah.
00:50:08Even without AI and AI will only turbocharge climate.
00:50:13So that's energy, generally speaking.
00:50:16Energy, resources, steel, cement.
00:50:18Yep.
00:50:19Robotics we got.
00:50:20Robotics we got.
00:50:21What else?
00:50:22Well, I'm super excited.
00:50:24So I was mentioning this talk I did.
00:50:27It's called Plausible Tomorrows.
00:50:29So you can Google that too.
00:50:31I have a dozen predictions that look really implausible unless somebody sits down with me and I walk them through each and they say, I see no reason this won't happen.
00:50:44I think by 2050, we can replace most cars in most cities.
00:50:49Self-driving or just don't have cars?
00:50:51Public transit that meets the following requirements.
00:50:55It's always cheaper than cars and cheaper than today's public transit.
00:50:59Cheaper than today's public transit without losses for the transit agency.
00:51:04That's always faster than public transit because it's like Uber.
00:51:11It's hailed, not scheduled.
00:51:14It's all personal.
00:51:16Well, it's two and four person vehicles, not, and it fits in a bicycle lane.
00:51:22Well, not only, by the way, public transit, Fusion, OpenAI, I invested in the same timeframe, roughly, around 2018.
00:51:32So I'm very proud of that vintage year.
00:51:35That's a good year.
00:51:35But people said, public transit startup, you must be totally crazy.
00:51:41And I said, Google's doing Waymo is the wrong way to do transit.
00:51:46Because the vehicles are too big or what?
00:51:48No, it just increased congestion.
00:51:51Okay.
00:51:51So what should it be?
00:51:52So, and I think Waymo will have a great business for delivering people from transit routes to the last mile.
00:52:00There's only one question in a city when it comes to traffic.
00:52:06Can you increase throughput 10x for the same street width without redoing the street width?
00:52:12So we designed a public transit system of self-driving vehicles in bicycle lane width.
00:52:19So it's easy to insert in a city, has 10x the capacity of a light rail system, or a car system.
00:52:29So more capacity than light rail system in a bicycle lane width.
00:52:33And it's on demand because the cars just, you hail them like an Uber.
00:52:38So, turns out, every single project we have bid so far, we bid a lot.
00:52:46Four have been decided.
00:52:48We want every single one for a public transit system from a startup.
00:52:54And the startup wasn't invited to bid on any of these projects because nobody knew about us.
00:52:59Yeah.
00:53:00Nobody invites a startup.
00:53:01San Jose invited 32 bidders to build a transit system from the airport to the Google campus and then to the Apple campus in San Jose.
00:53:14We bid over the transom because most of these bids are open.
00:53:18And we want it outright.
00:53:19Think about it.
00:53:20But is this, so what is the, what's the vehicle?
00:53:23It's a little pod.
00:53:24Think Disney World.
00:53:25It's like a little narrow car.
00:53:27Yeah.
00:53:28Yeah.
00:53:28It makes sense.
00:53:29It's in a bicycle lane.
00:53:30Yeah.
00:53:30By the way, it can fit above a bicycle lane too.
00:53:33Yeah.
00:53:34If you can go twice as many as cars on a street or something like that.
00:53:37So, it turns out it has way more capacity than cars in a street.
00:53:41Or even if everybody took buses, this would have more capacity in passengers per hour for a given street width.
00:53:50Are they also coordinating with each other so the cars are closer together?
00:53:53Is that part of it?
00:53:54Absolutely.
00:53:55You can coordinate.
00:53:56Yeah.
00:53:56Yes.
00:53:56Yeah.
00:53:56If you build a raised infrastructure, it's so simple, you can do prefab.
00:54:02Yeah.
00:54:02You don't have to do the Boston tunnel dig for 20 years.
00:54:06Right.
00:54:07Wow.
00:54:07You just do a prefab, put it in place.
00:54:09Yeah.
00:54:09So, it is thinking from first principles.
00:54:13Yeah.
00:54:14That matters.
00:54:16And it leads to, and being ambitious.
00:54:19Okay.
00:54:19And how about medicine?
00:54:20Because that's like 20% or more of GDP.
00:54:22That's a huge one.
00:54:23Yeah.
00:54:23Obviously, AI is a big part of this, but you spend a lot of time in medicine, separate from AI, with AI.
00:54:27So, off, which is a big part of our GDP.
00:54:33A quarter of that spend is medical expertise, is think doctors.
00:54:39Mm-hmm.
00:54:40Okay.
00:54:41Almost certainly that goes to zero.
00:54:44Can go to zero.
00:54:45So, what are we doing?
00:54:47We're building AI primary care doctors, AI physical, mental health therapists, AI physical therapists, AI oncologists, AI, you name it.
00:54:56Actually, maybe just to get specific on what you mean about that going to zero, if we have, for every 100 doctors per capita we have today, in like 2050, do you think we have like zero or 10 or, like, what do you think this is?
00:55:11I don't know the answer.
00:55:14But there are some.
00:55:15You will always have some to learn from.
00:55:19For the AI to learn from.
00:55:20For the AI to learn from, though the AI can boot itself to be much better.
00:55:25Mm-hmm.
00:55:26And I'll give you today's data of a multi-center study at Stanford.
00:55:32They took the best academic institutions and judged complex AI diagnosis.
00:55:41Physicians only.
00:55:43And these are academic learning centers with the best physicians.
00:55:4873% accuracy.
00:55:50So, 27% of the patients were getting the wrong diagnosis.
00:55:53It's like a lot.
00:55:54It's like one in, yeah, it's a lot.
00:55:56It's wonderful.
00:55:57Yeah, yeah, yeah.
00:55:57It's true.
00:55:58Think about it.
00:55:59And this complex means it was a serious thing.
00:56:02It wasn't just the flu.
00:56:03AI alone, 88%.
00:56:05Accurate.
00:56:06Accurate.
00:56:07But then they gave the AI to the doctors.
00:56:10The doctors improved from 73% to 76%.
00:56:14The AI got degraded from 88% to 76%.
00:56:17That's funny.
00:56:19And this is a serious multi-center study.
00:56:21I love medicine.
00:56:23What will happen is everybody on the planet, for less than a dollar a month, will get a free
00:56:30primary care, a free physician, initially called primary care, but they'll do everything.
00:56:38My son's company, QRI, is called multi-specialty primary care, because the AI knows enough
00:56:44gastroenterology to provide that.
00:56:46Now, the hard problem is the American Medical Association won't let the AI write a prescription.
00:56:52Yeah.
00:56:52And so, we have human physicians who approve the diagnosis and the prescription or testing
00:56:59recommendation or whatever.
00:57:01Yeah.
00:57:01I was going to say, I mean, if the humanoid at home plays out too, then it's probably
00:57:07also doing some amount.
00:57:08I mean, it's not going to do surgery, but it might be doing some amount of giving you
00:57:11a physical or giving you an injection or whatever else you need.
00:57:14Or watch your gait and say, hey.
00:57:16Yeah.
00:57:17Something's up.
00:57:18Yeah.
00:57:18Something's up.
00:57:19Yeah.
00:57:19So, that's a quarter of medicine.
00:57:22What about the other three quarters?
00:57:25A quarter is pharmaceuticals.
00:57:29That, we are seeing an explosion in AI in drug design.
00:57:33Yes.
00:57:33Both small molecule and...
00:57:37Are you pretty bullish on that?
00:57:38I'm very bullish on that.
00:57:40Do you think it will speed up the rate of drug discovery slash make it cheaper?
00:57:44Or do you think it's more about it will help us discover undiscoverable things?
00:57:48Probably a little bit of both.
00:57:50The hard part in that is the regulatory cycle.
00:57:54We have a separate effort to say, could you trust test anything against lots of human genetics
00:58:01and lots of human organs?
00:58:03So, we have an effort to do 10,000 tests, and it could be 1,000 different genetics, so people
00:58:12and 10 organs each, in one week, in one untouched robot.
00:58:20So, it just does the test, screens the drug, and looks for effects and growth and all that.
00:58:25So, maybe something like that would replace human trials.
00:58:31Most likely, at some point soon, it will reduce the size and length of these trials.
00:58:39So, I think drug discovery will see not only better molecules.
00:58:44The bigger problem in drug discovery is not to do what's called lead compounds for a disease,
00:58:50for a drug, but to increase the probability of success.
00:58:55The probability, once you have a lead drug candied to actual FDA approval, is so low.
00:59:03Like, out of 100, less than 5 will make it.
00:59:08So, it's increasing that probability because you're starting with a better candied
00:59:13and you're managing the process better.
00:59:16So, that's a quarter.
00:59:18I do think there is diagnostics imaging.
00:59:23And by the way, we had the most coolest self-driving for MRI machines, so you don't need MRI machines.
00:59:31Now, you need, today, technicians to run MRI machines.
00:59:35But if you have a self-driver, you don't need that.
00:59:38Cardiac MRI is so specialized because the heart is moving
00:59:42that you can't staff centers with cardiac MRI technicians.
00:59:47We turn an hour and 15-minute exam into 20 minutes.
00:59:52Single button press.
00:59:53You can even do it remotely.
00:59:56Because, so we did the same for ultrasound.
01:00:00Self-driving for ultrasound machines.
01:00:02An ultrasound machine might cost $25,000,
01:00:05but a technician to do a cardiac MRI will cost you $150,000.
01:00:09So, why care about the machine?
01:00:11It's the person.
01:00:12If you have self-driving, we had this company that GE actually acquired from us.
01:00:18So, all I'm saying is diagnostics and imaging is a quarter of the spend.
01:00:23It'll get cheaper.
01:00:24I'm shocked it's a quarter of the spend.
01:00:26I didn't realize that.
01:00:27Testing of all sorts.
01:00:28And then a quarter is in-hospital care.
01:00:33These are very rough numbers, but in-hospital care.
01:00:37And that will improve too.
01:00:39So, the net of all of this is like a multi-trillion dollar retrofit.
01:00:44Yes.
01:00:45Yeah.
01:00:45But, if expertise is free, and you can give somebody 10 times the amount of care,
01:00:56the average U.S. citizen gets one primary care visit a year, roughly.
01:01:04In Australia, it's four or five times that.
01:01:07So, we are poor anyway.
01:01:09If you provided way more care up front, you'd reduce the disease burden downstream.
01:01:17Fewer people would have heart surgeries, which are super expensive, or emergency room visits
01:01:23because you took care of them at the beginnings of the detection of their disease.
01:01:29So, every expensive disease, whether it's cardiac, Alzheimer's, Parkinson's,
01:01:35we are well on our way to discovering them 10 years before the symptoms manifest.
01:01:41So, I've argued a piece I did called 20% Doctor Included.
01:01:47In 2016, I said symptom-based medicine will disappear.
01:01:52You don't detect disease from symptoms.
01:01:55You detect disease based on what's going on in the body.
01:01:58So, that's medicine.
01:02:01Yeah.
01:02:01I'm pretty excited about it.
01:02:03With the level of precision that's possible, I think in the next five years or 10 years,
01:02:09biological intervention will approach computer programming in being able to precisely program
01:02:16things, design molecules.
01:02:19So, we are starting to, we have a couple of startups looking to do drugs for one patient
01:02:24at a time.
01:02:25So, the whole drug only applies to one patient for one genetic anomaly.
01:02:30And that's because that includes like the patient's DNA is factored in or something?
01:02:36Yeah.
01:02:36Their disease, their specific mutation can be addressed for them.
01:02:42Wow.
01:02:42There's also broader categories like a single shot treatment for sickle cell disease, right?
01:02:53Massive problem.
01:02:54Yeah.
01:02:54So, if you can do a genetic treatment, one shot, I think that'll happen in the next 10
01:02:59years.
01:03:00As I'm listening to you talk about all of these different areas and like you think on
01:03:04such a grand societal scale, but you probably also have a bunch of sort of maybe seemingly
01:03:12boring, but like good ideas kind of come through Coastal Ventures all the time too.
01:03:16Yes.
01:03:16Do you still enjoy making those kinds of investments as well?
01:03:21Is that more where you have like a partnership and there's other people on your team who are
01:03:23more interested?
01:03:24Well, one, there are other people in the partnership who enjoy just successful startups.
01:03:30Yeah.
01:03:30I do too.
01:03:31Yeah.
01:03:31Look, I enjoy winning.
01:03:33You know, when I do a GitLab, that's a win.
01:03:37When we do Cognition, it's a win.
01:03:39When we do Replet, it's a win.
01:03:41I mean, the growth rates for some of these companies are stunning.
01:03:45Shocking, yeah.
01:03:46So, yeah, it's still fun to be disruptive.
01:03:50Yeah.
01:03:50It's just sort of a different bucket of enjoyment.
01:03:52It's a different bucket.
01:03:54Yeah.
01:03:54I would say in the more societally impactful technologies, I will go through actually starting
01:04:04the company.
01:04:05Like Fusion.
01:04:07Bob was a senior fellow, Commonwealth Fusion.
01:04:10Bob was a senior fellow at the MIT Plasma Fusion Lab when I met him.
01:04:15And he had ideas.
01:04:17And I got him sort of like instigated, so to say.
01:04:24He's done all the work.
01:04:25And then he instigated the field of Fusion.
01:04:27So, whether he succeeds or not, there's so many different startups now.
01:04:33One will succeed.
01:04:35So, Commonwealth Fusion or Helion don't have to succeed for Fusion to succeed.
01:04:40So, I like instigating things.
01:04:43The same thing we did.
01:04:44It was very profitable to make 2,500 times your money.
01:04:48Yeah.
01:04:48And having the world be TCP IP instead of ATM.
01:04:54And it's interesting to look back at the press from 1996.
01:04:58Just assumed.
01:04:59Yeah.
01:05:00It was going to be, the internet was going to be ATM.
01:05:03So, I'm actually, you know, I always tell Pradeep, the founder of that.
01:05:07Yeah.
01:05:07If we hadn't done that, the internet wouldn't be TCP IP, which is horrific to think about.
01:05:14It'd be in the hands of AT&T engineering the protocols.
01:05:18Yeah.
01:05:19Which they had done.
01:05:20And this is why Lucent went out of business.
01:05:22Practically, they're still around.
01:05:25Alcatel, Newbridge, Nortel.
01:05:27All the telecom players sort of gradually disappeared because they didn't win place or show in the race for the internet.
01:05:37And the same will happen in other areas.
01:05:40You used the word instigating when you could have maybe used a word like incubating.
01:05:46I also heard earlier in the conversation, I think you, you know, you talked about like not being an investor.
01:05:50You know, I think also like the way that you work with founders and I've heard you talk about not being founder friendly or that there's a different way to think about being founder friendly.
01:06:00What I'm curious about is there's a lot of sort of new ways to describe what VCs do and how they describe themselves and mindsets around it.
01:06:11Do you think that in general, the role that venture investors see themselves playing, do you think we're trending in a direction where it's getting better all the time?
01:06:20Or do you think that in any ways the industry has lost its footing on some of these concepts?
01:06:25Look, there's all kinds of VCs with all kinds of approaches.
01:06:30Some are marketing labels.
01:06:33Founder friendly is a real disservice to founders.
01:06:38It's like if you said to your kids, if you told your five-year-old.
01:06:44Do whatever you want.
01:06:45Do whatever you want.
01:06:47I will always say yes.
01:06:48Yeah.
01:06:48That's, you don't pay the price.
01:06:51You want to go swimming without me?
01:06:52Yeah.
01:06:53Like whatever.
01:06:53Yeah.
01:06:54Or eat as much candy or eat as much Coke.
01:06:58And by the way, my kids, when they were growing up at age five, they could eat as much candy and as much Coke as they wanted, as much junk food.
01:07:07So we always had it laying around.
01:07:10They could, they didn't have to ask us, but we taught them how to measure themselves on how much Coke they drink, whether it's good or not.
01:07:21So I would say we taught them control, not gave them yes, no permission.
01:07:27Yeah.
01:07:28But the most important thing is that you set those parameters for a kid.
01:07:31Yeah.
01:07:31Yeah.
01:07:32Right.
01:07:32And in literally at age 12, I thought of they can make 10% of the important decisions.
01:07:41By age 18, they have to make 90.
01:07:43Yeah.
01:07:44And my job is to teach them how to make the leash longer.
01:07:48Right.
01:07:48But the same is true of entrepreneurs.
01:07:51So a brilliant engineer from DeepMind starts a company.
01:07:56What do they know about finance?
01:07:58What do they know about marketing?
01:07:59What do they know about even hiring or managing?
01:08:04Our job is to teach them, just like I taught my kids.
01:08:09By 18, my kids were pretty independent.
01:08:11I didn't worry about what they did in college.
01:08:13All four went to college five miles from our house at Stanford.
01:08:19Yeah.
01:08:19None of them had this, I want to get away because they felt they had enough leash and enough permission.
01:08:26They didn't need to, quote unquote, get away.
01:08:28Can I put up like a steel man argument of the other side to hear your reaction to it?
01:08:33Basically, it would be something like the sort of the strongest entrepreneurs, which is where all the returns are, don't really need much help or much management.
01:08:42And so better to just invest in the people who didn't need help in the first place.
01:08:46I disagree.
01:08:47That is very much the founder's fund argument.
01:08:50Yeah.
01:08:50They do very well with that.
01:08:52Great founders do well, whether you like it or not.
01:08:57Yeah.
01:08:57Whatever you do.
01:08:59Where I will disagree is they can do even better if they have a debating partner.
01:09:07So my view is very simple.
01:09:10I almost never work at a board.
01:09:13So I don't even go to boards now.
01:09:15You won't show up to a board.
01:09:16Do you mean?
01:09:17I don't get on boards.
01:09:19Yeah.
01:09:19I see.
01:09:20Yeah.
01:09:20And Square is a great example.
01:09:23I got off the board the week before the IPO.
01:09:28And everybody is like, hey, this is sexy IPO.
01:09:30Why are you getting off the board?
01:09:31I don't like the burden of the governance, the governance and all that.
01:09:37My responsibility shifted from having Jack do better to governance for the public markets.
01:09:46It's not very fun.
01:09:49It sounds glamorous, but it's not.
01:09:52You know, Jack and I would do regular monthly dinners.
01:09:56And I'd always challenge him and always push him.
01:10:00And he loved that.
01:10:01So he took the time to do these dinners.
01:10:04You talk to Max Lefgen at a firm.
01:10:06It's a public company.
01:10:08We long ago distributed stock.
01:10:10I still do quarterly meetings and challenge him for an hour every quarter.
01:10:14Right.
01:10:15Even the strongest founders like Jack and Max and need this.
01:10:21It's like even a pro sports player has a coach.
01:10:23Yes.
01:10:23You want people to challenge you and debate you, but not make decisions.
01:10:33I find it silly that boards want to take a vote.
01:10:36I won't even take a vote on the acquisition.
01:10:39I'll just say I'll support whatever management wants to do, even if I disagree strongly.
01:10:46But I do want to have the opportunity to debate it with the management team.
01:10:51So Jigdeep Singh is a founder, one of our robotics companies.
01:10:55I've done six companies with him.
01:10:57There was a company, I'm forgetting the name, called Light Something Optical Company.
01:11:03The team wanted to sell.
01:11:06I only asked to give me two hours with the whole team to make.
01:11:11And I did over the weekend the presentation I gave to them on why they should not sell.
01:11:16But before that presentation, as long as I had the two hours, I signed the papers and said,
01:11:23you have my vote anyway.
01:11:24You can choose to use it or tear it up.
01:11:27So I will always give the final say to the management team, but I do want to debate it.
01:11:36And I think every strong founder loves feedback that challenges them.
01:11:42Yes.
01:11:43And it's the only thing really valuable to a strong founder.
01:11:47Great founders love that.
01:11:50Not so great founders are so uncertain of themselves, they don't like that.
01:11:55And there, I prefer to teach them how to be a great founder.
01:12:00Yeah.
01:12:00But almost every strong founder loves feedback.
01:12:04I had Keith on this podcast.
01:12:07And what you're describing is a big part of, I think, what drew him back to Kosa.
01:12:13Yeah.
01:12:13By the way, on that topic, I'm curious.
01:12:15I read that as one of the sort of like generationally great sort of recruiting or re-recruiting moves that has happened.
01:12:25And I think he's an amazing investor, obviously.
01:12:28Did you know for a while that you wanted that to happen?
01:12:31We're a different kind of firm.
01:12:33And I said earlier, I've not called myself an investor ever.
01:12:37I always say I'm a venture assistant.
01:12:41If you look at our website, since we started the firm in 2004, the first tagline is venture assistants.
01:12:49By the way, another phrase that has stayed on our website since 2004 is, I prefer brutal honesty to hypocritical politeness.
01:12:59In vain most VCs give polite feedback to an entrepreneur, they're doing a disservice.
01:13:05They're basically saying you're doing great.
01:13:07Don't examine anything.
01:13:09You know, makes you feel good as an entrepreneur, but doesn't help you build a better company.
01:13:14And the only things that make you feel better are things that change your thinking or challenge your thinking.
01:13:20So it's an important characteristic for founders to keep in mind.
01:13:24What are they looking for?
01:13:25The best help or just people who will go away?
01:13:29So it's a style that's different.
01:13:33We don't think of investments.
01:13:35We've never calculated, never means not in the last 10 years.
01:13:39I can't think of a single instance where we calculated an IRR.
01:13:44Not one.
01:13:46An investment firm that never calculates IRRs because we want to build something significant.
01:13:53And then the economics will work out.
01:13:55So that's our approach.
01:13:56But we do want to increase the probability of success of the founder.
01:14:03And we want to increase the magnitude of the potential success also by thinking about, should you do X or Y?
01:14:11Founders underestimate that.
01:14:13So let's say a founder is selling 10% of a company, 10 million at 100 million post.
01:14:24The founders forget they're keeping 90%.
01:14:28And what happens to the price per share of that 90%, whether they give up 10% or 12%, keep 90% or 88%, trivial compared to the 500% swing you can get by taking one path or a different path.
01:14:45That's where the opportunity for maximization is.
01:14:50That's where it's not about investing, but getting the best help building a company.
01:14:54And that's what I sort of spend all my time doing.
01:14:58Just yesterday, I was talking to this founder.
01:15:02It's a very cool company, self-driving for these heavy machines, bulldozers and loaders.
01:15:10That's cool.
01:15:10Right?
01:15:11And we had an interesting session.
01:15:13And I reframed his business from automating big Caterpillar machines to, I'll just rent you the driver.
01:15:23You have trouble staffing drivers.
01:15:26Most of these are in mines and remote places.
01:15:29Equipment, a multimillion dollar piece of equipment sits idle because you don't have an operator.
01:15:37I said, just do rental by the hour and guarantee you will always have a driver available.
01:15:43Always have a driver.
01:15:45This reframing, Antrin was pretty excited about.
01:15:49It's the same thing, but it's not.
01:15:50It's a much easier sell.
01:15:54It's not saying change your equipment.
01:15:57It's saying, hey, you can't keep your equipment running all the time because you don't have operators.
01:16:03I'll run it for you.
01:16:05You're paying operators.
01:16:07I'll do it better, faster, more hours.
01:16:10You get more of your equipment.
01:16:11It's a simple framing change.
01:16:13But you could have a similar technology discussion, too.
01:16:17I know you're going to have to go in a minute.
01:16:19So I just want to ask you, there's one, I mean, there's a million more questions I want to ask, but there's one I really wanted to ask, which is you've been not just like doing venture, but you've been doing it like really successfully.
01:16:31And in a super relevant way for a long time, through a lot of cycles, before we started, you said, you know, you're turning 70 or you turn 70 this year, you feel 25.
01:16:41This is kind of like the dream.
01:16:42And do you know what you attribute all of this to?
01:16:46Yeah, very importantly, I gave a talk at Stanford in 2015, Stanford Business School 2015.
01:16:54I said, I'm internally driven.
01:16:57I don't care what others think of me.
01:17:00So it feels obnoxious sometimes because I'll do my thing independent of what others think.
01:17:05I'm doing this because it's fun.
01:17:07You know, I get somebody teaching me about mining and constraints in construction or self-driving an MRI machine or a big loader, the same thing.
01:17:21Clever ideas like public transit or can we make fusion happen?
01:17:26So I'm in it because it's fun for me.
01:17:29And most of the people at Khosla love what they're doing and would do the same thing if they didn't get paid.
01:17:39And so we do it for a different reason than making the most money.
01:17:45Because we focus on larger impact, the returns really do well or take care of itself.
01:17:53You know, it's sort of like what's motivating you.
01:17:56In fact, this was a conversation I had Sam when we invest in Open Air.
01:18:01He knows I care about the impact.
01:18:04Something happened, making something happen.
01:18:08Fusion is too important to not try it.
01:18:10I want the impact of fusion.
01:18:12There's nothing I could do in making money that would impact me more other than saying I paid this tiny part in getting fusion to start and hopefully happen.
01:18:23Even if Commonwealth fusion doesn't do it, almost all the companies have been enabled and investing in the field has been enabled because people looked at Commonwealth fusion.
01:18:34So this idea of instigating change is very motivational.
01:18:40It's way more impactful and also fun and way more learning than almost anything else.
01:18:47I say I'm like totally, I have one addiction, which is to learning.
01:18:51I can't learn enough new things, whether it's about biology or an AI algorithm or hypersonic flight.
01:19:00We are doing Mach 5 flight too.
01:19:03And we haven't talked about national defense.
01:19:05We are spending a lot of time there.
01:19:06I would love to.
01:19:07But all these areas are just fun.
01:19:11Yes.
01:19:11And that's why I don't mind.
01:19:14I still work 80 hours a week and I hope 25 years from now I'm working 80 hours a week, health permitting, touch wood.
01:19:22But that's, I think it's a very different motivation of why we do things.
01:19:27I tell our LPs, I'm probably the only one foolish enough to tell my LPs, if I can have more impact and less return, I'll pick more impact every single time.
01:19:36As long as they get a minimum return that they're comfortable with, so they can.
01:19:42Well, you seem to be getting both.
01:19:43Vinod, thank you so much.
01:19:44This was amazing.
01:19:45And I am super appreciative of your time and all that you've done for the industry.
01:19:49Great.
01:19:50Thank you.
01:19:51It's a lot of fun.
01:19:52Thank you.
01:19:52Thank you.
Recomendada
1:58
|
Próximamente
1:58
8:59
1:54:12
1:21
1:08
1:59:45
1:55
0:33
27:55
1:54
7:31
2:44
9:20
2:24
44:17