Bill Gates on How to Tax the Rich, AI, Misinformation, & the Election | On With Kara Swisher

0:00Rinse takes your laundry and hand delivers it to your door. expertly cleaned and folded so you could take the time once spent folding and sorting and waiting to finally pursue a whole new version of you like tea time you or this tea time you or even this tea time you said you hear about Dave or even tea time tea time tea time you so update on Dave it's up to you we'll take the laundry rinse it's time to be great

0:33When did making plans get this complicated? It's time to streamline with WhatsApp. The secure messaging app that brings the whole group together. Use polls to settle dinner plans. Send event invites and pin messages so no one forgets mom's 60th. And never miss a meme or milestone. All protected with end-to-end encryption. It's time for WhatsApp. Message privately with everyone. Learn more at whatsapp.com. Running a business comes with a lot of what-ifs. But luckily, there's a simple answer to them. Shopify. It's the commerce platform behind millions of businesses, including Thrive Cosmetics and Momofuku.

1:11And it'll help you with everything you need. From website design and marketing to boosting sales and expanding operations, Shopify can get the job done and make your dream a reality. Turn those what-ifs into... Sign up for your$1 per month trial at shopify.com slash special offer.

1:32Hi, everyone. This is Pivot from New York Magazine and the Vox Media Podcast Network. I'm Kara Swisher. We're off today for the holiday, but we have an episode of On with Kara Swisher for you. Oh, yay. Yay. Who's on this time? The head of DEI from the National Forestry Service. No, no, no. That's just that's what I say for Pivot. This is my interview with Bill Gates. You might have heard of him. Oh, my God. In front of a live audience. I got to be honest. That's impressive. That's impressive. Thank you. I try. It's on!

2:32It's called What's Next? The Future with Bill Gates, and it's about some of the biggest issues we face at the moment. AI, disinformation, global health, the wealth gap, and how technology could help or not. It's a funny and substantive series by Morgan Neville, who did docs like Won't You Be My Neighbor about Mr. Rogers, Roadrunner about Anthony Bourdain, and 20 Feet from Stardom, which garnered him an Oscar. It's a good match since Neville got Gates to actually use his own experiences as case studies in some of the episodes. It's surprisingly intimate, actually, and I actually really liked it.

3:06I didn't think it would be as good as it was, and it really is, especially for people who don't know a lot about these topics. It was also great to be back on stage with Bill, whom I've interviewed many times, including an historic interview with the late Steve Jobs. While we've had plenty of disagreements over the three decades I've covered Microsoft and also Bill Gates, I took the opportunity to dig a bit deeper and ask him about his interests and investments into many of these areas, including nuclear power, how he would handle attacks on the rich, and the outcome he's hoping for in the upcoming presidential election.

3:38Plot spoiler, he implied, referencing climate change deniers, that he's voting for Vice President Kamala Harris, although he certainly didn't make an endorsement. I hope you enjoy it. And by the way, we love that you're listening. And it's even better if you hit subscribe to follow the show to get even more exclusive conversations like this in your feed every Monday and Thursday. So let's get to it live from the Paris Theatre.

4:12Hi, everybody. How you doing? I used to come to this theatre as a kid. It's fantastic to be here. Owned by Netflix. Wow, that's something. Anyway, thank you for coming. It's a little lightness here, inequality, AI, climate change, disease. What else are we talking about? But at least we're not Eric Adams, are we? Okay, I had to do that. Oh, come on. Anyway, I love New York. Anyway, I'm very excited actually to talk to Bill. I've interviewed him dozens of times over the many years. We first met when he came pouring out of a cab in very sweaty, summery, Washington, D.C. I thought he had arrived in the limbo, but it was some other guy, and he just got out of a DC cab and did an amazing interview with the Washington Post editorial board and came himself and everything else.

5:05So that's where we started. I've covered Microsoft for years and talked to him a number of times. We've agreed, not that much, but disagreed a lot. But also I'm really, I think this phase of his life is really interesting, and he really is an information sponge curiosity, and he loves to take apart problems. And I think what he's done around disease and other things is really interesting. The last interview we did was about climate change, which is a book that he wrote. And it's actually coming true today, a lot of the stuff he was talking about. So without further ado, Bill Gates.

5:45Oh,

5:49we meet again. So, there's so much to talk about. I think what we're going to do is we're going to throw to some clips and then talk about them. We're literally going to have to rush through this stuff because we've only got about 40 minutes to get through all these major topics. And you could do hours on each of these things. But let's start with AI. Let's go to the clip to start with. And you're talking to James Cameron. And a me is a, hey, innovation can solve everything type person. Right. Says, oh, thank goodness. Now I have the AI on my team. Yeah. I'm probably more of a dystopian. I write science fiction.

6:21I wrote The Terminator. Where do you and I find common ground around optimism, I think, is the key here. I would like the message to be balanced between this longer-term concern of infinite capability with the basic needs to have your health taken care of, to learn, to accelerate climate innovation. You know, is that too nuanced a message to say that AI has these benefits while we have to guard against these other things? I don't think it's too nuanced at all. I think it's exactly the right degree of nuance that we need. I mean, you're a humanist, right? As long as that humanist principle is first and foremost, as opposed to the drive to dominate market share, the drive to power.

7:07Or if we can make AI the force for good that it has the potential to be, great. But how do we introduce caution? Regulation is part of it, but I think it's also our own ethos and our own value system. So that's a perfect person, the creator of The Terminator, which I think has informed a lot of people about AI, the idea that it's here to kill us, essentially. And, you know, the person I talked about AI with was Elon Musk, actually, who was quite worried about it at the time. He thought it was here to kill us. Later, he evolved it into treating us like house cats. Then he said we were like anthills, that the highway will come through and just cover us, et cetera.

7:47Always a happy day with Elon Musk. But when I was watching this, I watched it again on the train up here. It's advanced quicker than anyone, including yourself, had thought, right, that it's moved quicker. and you talk about this, and it's a very good explainer in the film. But I want you to sort of unpack why you're still very sunny about it, I would say. And I think you lean towards the possibility. So talk a little bit about that. Well, the idea that computers would eventually become intelligent and do human-like things, you know, as soon as you learn how to program, you're thinking, wow, what can't software do and why not?

8:28And so prosaic things like visual recognition, speech recognition seemed out of reach. I was age 15 when I saw this Stanford SRI, Shaky the Robot video, and I thought, wow, we are getting close. Look at that. Shaky almost can do these things. So it has been mostly disappointing when the neural nets came in. Then the prosaic stuff, the speech and video got strong. But even then, the idea of representing knowledge, being able to read and write, still seemed out of reach. And I believe that we'd have to explicitly understand how to represent knowledge. I didn't think that pure reinforcement would cause this great knowledge representation to emerge.

9:17And even GPT-3, you know, it's like, it's still statistical. It's not really a world model. And so it's going to be deeply flawed. But by GPT-4, that was wrong. And so, you know, right up there with my first demo of graphics interface with Charles Simone, I'd say, you know, that demo of chat GPT-4 was the most amazing demo I've ever seen. I think a lot of people attribute the huge amount of information we've been putting into the internet right now has the leaping point. What do you, because I was at a dinner with a very smart group of technologists, and they didn't know. So talk about not knowing, because software is about knowing.

10:00When you put something in, you take something out. But what do you attribute the leap to recently? Well, there's a subtlety to how the world gets represented that this is captured. If you told me to write a piece of software that you could say, you know, rewrite this like Shakespeare would, or like Trump would or like Harari would. This thing, I would have no idea how to write a piece of software that is as fluent as this is. And so that's an emergent set of capabilities that shows that it is, in a deep, deep sense, representing what it is to write like Shakespeare. Right. And how our algorithms generate that, you know, is subtle enough, and we have a lot of mathematicians working on, you know, trying to figure out those representations, but it does make it harder for us to know, okay, where is it correct?

10:58Where is it incorrect? Right, but it's not Shakespeare. It can write like Shakespeare. When does it move to whatever Shakespeare may be? That's the idea. It's representing us. Well, if your goal is utilitarian to say, compare humans in terms of medical diagnosis to this ability to look at all the medical records and see outcomes, to look at all the latest medical literature, to look at your sensor data over your lifetime, clearly the computer is very close and will exceed human diagnosis capability within the next few years. It will be superior at a very profound and important task. And, you know, that's just provable that it will be better.

11:48If you're talking about art, okay, then, you know, we don't have an agreed utility function. You know, why is some art more popular than others? Okay, that, it can't enter that realm because there's no data to center it on a common metric. What then is it lacking? I mean, you're talking about the utility. You often talk about healthcare. Education is another area. Drug discovery, you know, gene research, et cetera, et cetera. These are all utilities. But a lot of the worries are about things that aren't utilities, that it will start to make decisions for us. Where's the worry point that you have?

12:29Because often most of the technologies talk about the positive parts about it. And I understand why you would talk about the negatives in the beginning, given our past history recently. Well, there's three things you can worry about. One is that people with bad intent will use AIs for cybercrime, bioterrorism, nation-state wars. And so in that case, you think, okay, let's make sure the good guys have an AI that can play defense against those things. And that makes you want to move ahead and not fall behind. The second thing you could worry about is that the rate of change where technical support jobs, telesales jobs, You know, in the same way that that medical diagnosis will be superior with the right training set and a few more turns to the crank on how we drive reliability, that there's great progress on that.

13:33It will be superior at a telesales or telesupport type job, which, you know, those are big parts of the economy. And so even though you can say, okay, that frees those people up so we can have class size of five and every, you know, handicapped kid can have a full-time aide and elderly people, you know, can be engaged in social activity. We have unmet needs for labor. If we free up all this labor, we can shorten the work week. But the rate of change is scary. And so those are the two I worry about. The third one that comes up a lot is the loss of control. My view is if you manage to get through the first two, that actually that's not the hardest of the three.

14:18And it's actually kind of weird to go to that one because it's pretty far from the future. We know that bad people will use AI in a bad way. The AI itself, you know, I think. It doesn't care. I don't think that's the one to obsess, particularly given the rate that this is happening. This is not a generational change. Is this within a 10-year type change? Sure. One of the things that's, when you just say loss of control, I think some of us are loss of control to giant companies that are going to dominate this area. OpenAI is raising money at$150 billion valuation. There's some sort of kerfuffle going on among all the executives, but it's turning into a for-profit corporation.

15:02It started as a nonprofit to help humanity. Now it's going to help humans, a few humans, a few select humans. Microsoft now has vaunted ahead with its investment in open AI and everything else. If there's a few corporations running this, I would worry about that. Would you? Well, in general. Although you're a big shareholder, Microsoft. We have this thing called competition. And the main thing you want is that it's so competitive that the improvements go to benefit the user. That is, the quality of medical diagnosis improves the health system. The availability of that personal tutor that I was out in Newark seeing a few months ago, that that becomes super cheap for kids in the inner city, not just wealthy kids.

15:55And I've never seen such intense competition. You know, it's the same way as the internet in 2000, 2001. The failure rate will be unbelievable because they're sort of a, oh, it's an AI thing, give it a high valuation. So consumers will be the primary beneficiaries. Now, if you ever get to the point where one company has knocked the other companies out, then fine, the competition authorities can come in. But we're in no, you know, we're not even close to that. You know, Google is still extremely capable in this space, OpenAI plus Microsoft. But, you know, Elon's got XAI. You've got Anthropic. Backed by Amazon.

16:40At least 15 companies that aren't that far from the state of the art who are lowering prices. You know, they're investing way above the revenue level. And so this is the formula for the benefits to flow to society. Okay, we're going to rush this, but what's very quickly your biggest worry, a real worry, and your biggest benefit right now? Well, the biggest benefit is that I think this technology, and this is my primary engagement with it, we can get it to the inner city. We can get it to Africa within the same time frame that it gets to the wealthy. Unlike any other technology in history, this one I think we can get it out on an equity basis.

17:28And your worry? My worry is just that this is the most unbounded thing. This is not just a tractor, you know, obsoleting farmers where you say, okay, it turns out there are many other unmet needs there. This, you know, which we thought would first come for blue-collar work through robots. You know, the surprise was that with this language facility, it actually is coming first for white-collar jobs. But the blue-collar thing is happening just a little bit later, certainly within the next three, four, five years. And so it's so disruptive at a time where government, who you expect to take the excess productivity and spread it around in a fair way, reduce the work week, our general trust in the capacity of government to do complex things is pretty low.

18:26Yeah, so it was a good run, guys. So congratulations. All right, let's roll the next clip, which is an area I really spent a lot of time. Do you find yourself, even at this age, using your phone and staying on social media more than you want to? Oh my gosh, yeah. TikTok is so addictive. I'm on it like all the time. And have you ever run into crazy misinformation about me? Crazy misinformation about you all the time. I've even had friends cut me off because of these vaccine rumors, but I'm a public health student at Stanford, and I think that there is just so much nuance on how do you communicate, like, accurate public health information or scientific data.

19:04I don't know. I need to learn more because I naively still believe that digital communication can be a force to bring us together, to have reasonable debate. I think one thing, like, you don't really understand about online is, like, it's not really, like, logic and fact that went out. Like, people want an escape. They want to, like, laugh. They want an engaging video. They want to, you know, be taken away from boring reality. And so, like, the most popular video of you online is you literally trying to do the dab. Bill, can you do the dab real quick? Damn, Bill! Or you jumping over the chair.

19:39Is it true that you can leap over a chair from a standing position? It depends on the size of the chair. I'll cheat a little bit. Yes! Those are the most popular because people want to escape from things, so I don't think fact and reason always went out online. But the thing about, you know, I make lots of money from vaccines, it's even hard for me to figure out where that comes from. It's not like a political organization. It's just madness. And who promotes that? I think it's fear. I mean, everyone was stuck at home during a pandemic. We're all scared for our lives. No one really knows what to trust or what to believe.

20:12So that's what our society does. Okay. So can you do the chair right now? No. A smaller one. Yeah. A little stool. stool. I can do a stool. Your daughter's very wise, but most of the stuff about you isn't funny, actually. I spend a lot of time telling people you're not Satan, that you don't put chips in people's brains. I spend an inordinate amount of time doing that, which is an unusual position for me to be in. But let's talk a little bit about information. You talk about Anthony Fauci, who has had death threats. I just interviewed him recently. One of the reasons government is and the trouble is in, is the polarization, which is fueled by online stuff.

20:55And there's no question about it. I know they're trying to come up with studies that it's not true, but a new study just came out showing it is absolutely true. The polarization has been further impacted, especially around, we use the word misinformation, disinformation, but it's really propaganda, isn't it? That's really what's happening. 1.2 billion views were of Elon doing misinformation on the platform he bought, bought, which he enjoys to spew misinformation on. It's actually the biggest purveyor of misinformation right now. Can you talk a little bit about this and the worries you have? Because it does directly impact you yourself, which I'm not worried about you, but everybody here in this room.

21:36Well, I didn't anticipate that if you want to belong to a group, there almost become certain beliefs, you know, the election was stolen. You know, vaccines have negative effects. And recently Haitians eating dogs. Fauci is a self-interested individual, you know, who somehow gets royalties from vaccines. I mean, you know, totally false, provably wrong. I want to track the position of people. I don't know why I do, but it's amazing. Because you're Satan. But go ahead. But even Satan doesn't need to know where people are. It's fair. What does he do with it?

22:24So that's, it's just strange. And, you know, we always try to deal with it with humor. And as you say, I have nothing overall to complain about. But the idea that your sort of group beliefs are reinforced online, And so you have developed a reality that when people say, no, the election was not stolen, you're like, no, that would be almost denying that I'm part of this group. And so I'm going to behave that way. It's incredibly scary because it's definitely putting us into separate camps. And it's hard to see. People say, oh, we didn't regulate social media properly. Well, do we know now? Are we regulating it today?

23:13Is there a clear understanding of it? And then, of course, AI that we just discussed, if anything, supercharges the ability to create credible misinformation. What would you do now to stop this? Because I think people are in their own separate— Microsoft was an early investor in Facebook when it was a$15 billion valuation. Nice one. That was a good one. I doubted that. I'm sorry. I was wrong. But when you think about what should be done now, because if people are in their own information bubbles in a way that's profound, and then AI, for example, can supercharge it, as you say, what would be the solutions to avoiding that?

23:51Well, most things where, you know, the country has been off track, or even the world has been off track, you start to see the harms. And then, you know, people self-correct, you know, parents play a role, educators play a role and some degree of banning the extreme behavior plays a role but you know first we have to come to a view that okay this really is a problem the United States may not be the first place to get this under the control we have the first amendment and you know our divisiveness is is particularly high I wish I could say okay such and such a country has done this extremely well.

24:36You know, the country that actually controls craziness on the internet is China, but they do it in a way that destroys - They do like to track everybody where they're going. Yeah, and they don't let crazy things out there. They do better, but at the loss of democratic freedoms. So what do you do in a situation when, say, you have someone like Elon who's suddenly decided to let misinformation flow toxic waste all over the place. And in some cases, people who try harder, like a Mark Zuckerberg, who you were a mentor to him, he and I had a big argument about Holocaust deniers a couple years ago. And I kept saying, this is going to lead to anti-Semitism down the line.

25:18And he said, well, we're not going to regulate them at all. We'll see what happens. And you can see what happens when you anticipate. So is there anything government can do? Or are we at the mercy of these companies to decide? You know, because years later, Mark did clamp down on Holocaust in ours, but it took him to do it, which is problematic, I think. Well, you know, the question is, do you create some level of liability for these companies? And many forms of that would essentially mean social media companies would go out of business. But, you know, we, a little bit, I think we ought to edge in the direction of forcing them to take some more responsibility.

26:00It is very tricky because if somebody says, you know, vaccines in general kill people, that's wrong. And it caused older people who needed COVID vaccines not to take them. If people say, hey, vaccines sometimes have side effects, that's true. You know, we need to talk to people about the net benefit and how we avoid those side effects and things. So the exact dividing line where, you know, all of a sudden, you know, some AI wakes up and shuts it down or labels it, I don't even feel like we're trying to find that happy medium. Who is responsible? Is it the parents? Like California is banning different things.

26:40Different states are doing. Where is the line? Because it can't be Elon Musk and Mark Zuckerberg. It just can't be. Well, nobody's really there. It hasn't been much regulation. You know, take even young people getting online. That's not an enforced thing. So I'm a little surprised how hands-off we've been on these things. With kids. I mean, we over-regulate them offline and we under-regulate them online. It's really quick. I thought Anxious Generation and many observations along those lines are very profound and we should take action on. So if you could do one thing, what would it be? I know what I would do, but what would you do?

27:16Misinformation actually is the one topic, and I say it in the series, that I say, hey, young people, you grew up with this thing. You understand the phenomena better. We basically pass this as a big problem to you. In AI, I have thoughts. Global health, I have thoughts. But misinformation, I'm just stunned that there aren't more clear, constructive things relating to policy and technology. Yeah, so you're just like, good luck. Good luck, young people. You know, actually, the way my kids dealt with it is they took all social media off their phones, which I thought worked rather well. We'll be back in a minute.