Hello and welcome to the NVIDIA AI Podcast.
I'm your host, Noah Kravitz. Today's guest is making a return visit to the podcast.
Having joined us roughly two years ago, back in November 2018 to talk about his then-new book, AI Superpowers.
Today, Kai-Fu Lee is back, and he has a new book, AI 2041, which is a very unique collaboration that uses short works of fiction to describe what our world might look like 20 years from now in 2041. thanks to AI.
Kai-Fu Lee is the CEO of Cinevation Ventures and a New York Times bestselling author of the aforementioned AI Superpowers.
Li was formerly the president of Google China, and he's now also co-chair of the AI Council at the World Economic Forum.
And I'm delighted to welcome him back to the NVIDIA AI podcast.
Kai-Fu Lee, thank you so much for taking the time to join us.
Well, thank you. Great to be here, Noah.
So I've had a chance to read some of the book.
It's divided into 10 stories with explainers and an introduction at the beginning.
I've read, I think I'm on my fourth one now, so I've got a pretty good sense of it.
And it's really, as I said, a really interesting, pretty cool approach to to the topic, but why don't you tell the audience about it?
What is AI 2041 about and how did you Get the idea, or how did the collaboration come to be with your co-author Chen Chufan, also known as Stanley Chan, who's a prolific writer of fiction in China?
Yeah, I just think AI is such an important topic that more people should be exposed to it. still see and hear a lot of people who feel intimidated that this is rocket science, don't ever get it.
But because it impacts their lives and their children's lives so much, I wanted to see if we could create a new, innovative system piece of work that is not only accessible, but also engaging and perhaps even entertaining to more people.
And by presenting stories to them, by the end of the 10th story, they will have taken the first lesson in AI.
So that was the idea. And of course, I'm not a fiction writer.
So I reached out to my friend and former coworker at Google Stanley asked if he would do that.
I was a little hesitant because many science fiction writers would not want to be constrained.
Because in this book, I didn't want to just go all out and use all of his imagination to talk about technologies that are and aren't feasible.
I wanted to make it a realistic, achievable 20-year-old vision.
And he said, okay. So we went ahead. and we co-designed a special way to make a book which was that I had about 20, 25 technology areas I wanted to cover.
Things like deep learning all the way to quantum computing.
And then we pretty much order them from simple or basic to more advanced.
And also we wanted to apply them to 10 different industries like healthcare and entertainment and education, to indicate how AI changes and disrupts all industries.
And then lastly, he had this great idea that it should happen in 10 countries. so that people get and understand that this impacts everybody, not just US and China.
So we then mapped out the 10 stories, what the storylines would be, which technologies they would have, what country, what industry, And then he went off and wrote 10 stories.
After each story, I wrote a commentary that described what are the technologies used in the stories, How do they work?
What are the possible upsides and downsides?
And how might we in the future deal with the downsides?
So that's how the book was written. It's a very cool approach.
And as soon as I started, I guess I started with the little press introduction that was sent over before I dove into the book itself.
But Before I was done with the introduction, I had a little list going in my head of people who I wanted to tell about it.
And they were people who, as you said, are fascinated, but perhaps intimidated by AI, perhaps maybe a little... to their predisposition was maybe colored a bit too much by the robot overlords are coming, they're going to take my job, that whole line of thinking.
And so what I found right off the bat great about the book is that, first off, the stories are very accessible.
They're not talking down to people, but you don't have to be a rocket scientist, as you say, to understand them.
But then the explainers right after the story were really nice, again, very straightforward, but not dumbed down way of explaining These are the technologies at hand.
And as you said, these are all of the different ways.
There's great diversity amongst the characters and the locales and the industries. all the different ways that these things are happening and will continue to happen. actually quite excited for some of these folks I thought of to read the book and get back to me.
And I'm curious to find out. what they've learned and how that knowledge might change their perceptions.
I'm wondering, and as you were going through the creative process, if you had folks like that who maybe weren't so up on the technologies, give you feedback from that perspective.
Yeah, yeah. I sent them out to a number of friends who read them.
And several of the proofreaders for the book were also neophytes in technology.
And they gave a lot of excellent feedback about how much they had learned about technology beyond that, the degree to which they thought they could.
So I'm glad it can serve the purpose And you mentioned that the book is sequenced so that the later stories, would you describe them as more challenging or just kind of getting into more advanced technologies?
Yeah, the first seven stories are really digging into technologies and then the last of which is quantum computing. and autonomous weapons, which I thought were some of the more advanced uses or challenges that would be presented to us.
The last three stories were more alternate ending kind of stories.
They do have some... technologies, but some are more dystopian, some are more utopian.
I talk about the economic implications, what happens when AI disrupts and changes manufacturing, automates it, reduces the cost, and energy becomes a manufacturing problem because We just need to make very low cost solar panels and new batteries.
And with AI technologies making that easier and better, what happens to various economic theories as we reach a future of plenitude. also explore topics like if AI can brainwash us, like some people say, by using social networks to show us content that makes us more extremist or something like that.
But why can't we use that? for our own good.
Can there be a future AI that will make us happy is another topic explored.
I think the last three were more economic and social implications, although they also do have some technologies embedded.
Being a father of two school-aged children and a former teacher, and I'm married to a lifelong educator, I found Twin Sparrows to be...
I wouldn't say more or less interesting than the other stories, but just it struck a chord with me and it got into notions of... of education, obviously, but personalization and these ideas of AIs that develop over time and really get to know their user, if you will, and can kind of help shape a life in that way.
I don't want to put words into your mouth, but I was struck by the sense, and having read AI Superpowers as well, that for someone who's seen as much of you as you have and knows what's going on with the technology to the extent that you are, you seem like you're an optimist more than a pessimist about this stuff.
Is that accurate? Yeah, I think that's accurate.
And and the reason is that I know there are a lot of issues with technologies.
I also worry about many of the issues that are discussed in the book, issues with fairness, transparency, personal data, also issues with autonomous weapons, I'm concerned about all of these.
Sure. But at the same time, when I look back at all the technological breakthroughs, such as electricity, or the internet.
When they were first launched, there were also very big concerns.
The internet brought about viruses to our PCs, which weren't so bad before.
Electricity could cause people to be electrocuted, but ultimately people found solutions, and usually technological solutions, like the circuit breaker or antivirus software.
So I'm an optimist because I'm a believer that technological problems will always happen at the advanced stage of a technology breakthrough.
And over time, it will be overcome, has been overcome and largely by using technology solutions.
So that's one reason I'm optimistic just looking at history.
The other is that I feel there's so much negativism out there and while the basis are justified, but I think people are becoming too cynical or even dystopian So I feel like there needs to be a counter voice that shows the constructive side of things that, yes, there are issues with fairness, but when they're dealt with, it could bring us greater happiness.
So I wanted to present the other point of view so that the viewpoints are more balanced out there.
So you used the word plentitude before in describing kind of moving towards an age of plentitude that AI may well help bring about.
And there's a story in the book called The Job Savior.
This is one of the ones I have not read yet, so I don't want to speak out of turn about it.
But it deals with this idea of what's going to happen to human jobs as AI is able to automate apply efficiencies and make some of these human tasks redundant.
I've always been fascinated by this notion and might speak more to my own relationship to work than anything else.
But this notion that eventually sophisticated technologies can do the grunt work for us or can do so much of the grunt work for us that humans will be able to take on the more interesting creative jobs or perhaps not work at all as we think of it and have more time for other pursuits. without asking you to obviously spoil anything in the book.
But can you speak a little bit about that notion and I don't know, maybe dispel the pessimism in me that thinks like, oh, well, that could never happen.
It'll just be a widening of inequality. or the people who control the AIs will benefit, but the rest of us won't, I guess I'm wondering if you could speak a little bit more about this idea of plentitude that AI might help bring about.
Okay, so first on the jobs issue, I'm one of the believers that AI will take away jobs.
It will also create jobs, but probably not create them fast enough at the same speed that it will be decimating them initially.
Because is by definition designed to do what we currently do because it's intelligent and therefore it is by design going to take away our jobs.
And especially the routine ones, mostly the routine ones from white collar routine jobs to blue collar routine jobs.
And that's a sizable So the story, Job Savior, was designed to point that out to people so that people are aware that if one has a purely routine or largely routine job, it will be displaced by AI.
And moving forward, what it will cause is that the companies that use AI to displace people or develop AI to sell to companies to displace people, these kinds of companies will become dramatically more valuable.
And at the same time, jobs would be lost in the process, thereby leading to wealth inequalities, so that this is one of the worrisome chapters in the book.
And I outlined a possible solution. I don't think it's the definitive answer, but people do need to think about answers.
People talk about universal basic income.
I think that kind of a measure is needed to reduce the degree of wealth inequality But more importantly, people need to be aware what jobs are going to stick around, which jobs will be displaced, and which jobs might be new. we can actually make a pretty accurate projection even right now for the next five years or even longer.
We know auto mechanic jobs are going to be changing and the old types are going to be diminishing.
Plumbers are going to be okay for a while, but eventually they'll be gone. but in the long term, because our plumbing is different house to house.
And healthcare services, we're going to be in a shortage of that because people will live longer with a larger population that is more well-off and people don't want robots to take care of them.
So we have a reasonable map of which jobs are going to get displaced.
And I think it would be irresponsible if we don't let people know and redesign vocational training. and make sure that people who are being displaced are getting the training into a job for which they're interested and qualified and that doesn't get destroyed again.
So the story in Job Savior is about the company, which is in a new category of companies called the job reallocation.
They're brought in when there are layoffs. and they're paid by the government to help people find a new job.
And then it proceeds to find interesting ways to give people effectively practical training because routine jobs becomes rare and rare as AI will take them over. and most jobs will require more training.
So it really does a, finds a very clever way to get people into practical training.
On the planet to an idea, The next step after jobs are automated is that the energy costs will also go down the cost of distributed, storable solar energy has come down by about 85% in the last 10 years.
And if that continues for another 20, the cost of energy will be a tiny fraction of what it is.
And the cost of making goods whether it is our electronics or clothing or even food will all come down because they can all be manufactured.
And as they're manufactured, there are three types of costs, materials, energy and labor in increasing order of magnitude.
And labor will come down dramatically. Energy will come down significantly. and materials will also come down.
And that means for the first time, in 20 years or so, we will be able to, we as the human race, will be able to afford to essentially wipe out poverty and hunger.
Whether we do so or not with the national boundaries and selfishness and greed is another question.
So the last story, Dreaming of Plenitude, talks about a country, Australia in this case, that has realized that it definitely has the capability of wiping out poverty and hunger.
But at the same time, the economic models need to change and people's motivations need to change.
So it tries different policies. And then we get to watch certain people who are doing jobs that wouldn't be considered jobs today, but because they don't introduce as much economic value, but they do have a lot of social value and that they don't necessarily earn money, but they earn some other token that give them a sense of accomplishment and pursuit and even self-actualization.
So that's that last story. As I'm listening to you talk about this and thinking about your—well, this call is being conducted— across oceans and time zones and national boundaries and such, I'm wondering kind of two things.
One is I'm thinking about the relationship between industry and technology companies and governments and how it's seems like in certain areas of the world and the United States where I live comes to mind that Maybe the capabilities and the knowledge of the companies when it comes to these technologies is outpacing the knowledge and the capability of government institutions to... regulate or harness or whatever the best word is.
And so I'm wondering that sort of in conjunction with the response, to the book and the different stories, how they might be similar or different in different parts of the world.
And I guess that's kind of a word salad of getting to asking, in your experience working across all the different organizations and different countries and areas of the world, Do you see markedly different takes on the importance of technology and the relationship between technology and government in particular? in kind of making these decisions and raising these things to people and educating people about the future of their jobs and the future of everything.
Do you see market differences in different parts of the world?
And did that play in at all to your process and thinking about setting these stories in different places?
Actually, the governments are relatively similar in the way they view the positives and negatives of AI.
The timing may be different. I think saw Europe first became alarmed about the power of the internet companies and how they needed to be directed. followed by US, followed by China, but they're all on board with the same need. to avoid too much power in the hands of a small number of companies that might have near monopolistic power.
About job displacements, similarly, the countries are also relatively similar because I think governments tend to not want to address issues until they are proven.
So right now, job displacement is still not yet showing up, are still not showing up. yet in the unemployment numbers.
And I think COVID also will further mask that because To the extent there are more displacements happening, which I think there are, people might say that they're caused by COVID rather than technology.
So I think most companies are not yet ready to jump in with the job displacement issue, which I think is unfortunate because there are some things that governments can do even without much cost.
So that's that aspect. the other aspect is i think most governments do think about ai as a arms race and think that's certainly countries want to be better and more advanced in technologies but for the most part ai should not be any more of an arms race than a, you know, genetic sequencing or other types of important technology.
And I think countries ought to cooperate more.
But I think right now that's not the trend that the countries want to be better than other countries, more advanced, and they view it as a national priority.
I suppose that's good from the perspective of more attention and more funding into AI. but it's not creating a global collaboration that I had hoped that we would continue to see.
My guest today is Kai-Fu Lee. Kai-Fu has a new book out with his co-author Chen Chufan.
It's called AI 2041, and it is a look at our near-term future, what the world might look like in 20 years. thanks to AI and related technologies.
And as we've been discussing, it's a series of 10 short stories that, Kai-Fu, if I have this right, you kind of worked out a technology roadmap, so to speak, that you presented to Stanley that he wrote from his drafts.
I'm wondering, was the book or was part of the book written during COVID?
Yes, it was all written. It was all written, okay.
Yeah, we get more free time. So we were able to-
We actually did it all over video conferencing because we both live in China, but I'm in Beijing and he was in Shanghai.
Okay. We use the extra time that COVID gave us.
I'll use video conferencing, the COVID tools, if you will, to pretty much work through the book.
Do you think the experience of COVID, I mean, either of the particulars of how you had to work, and I'm thinking kind of bigger picture, did that change your vision at all or even influence, you know, sort of some of the, I don't want to call them predictions, but some of the visions that you landed on for the stories?
Yes, it did certainly have an impact. In the stories, you will see some very interesting COVID inspired ideas. including healthcare advances, new drug discovery, and longevity. you'll see COVID induced isolation, sense of and psychological impact on people. that if the virus keeps not going away, will people become more isolated and what are the implications there?
You'll see ideas about work from home taken to an extreme, and you'll see ideas like When you work from home, can you be accompanied by all kinds of robots that take care of the chores around the household and also maybe keep you company?
So all of these ideas would not be in place had it not been COVID.
So I've heard this, and I think attributed to you, whether in this book or in a different piece of yours, but I've also heard it other places that One of the things that, when we look back on it, COVID will have done is accelerated a lot of processes, and specifically talking about technology. and working from home and some of the technological automations that...
I do a lot of writing for my living. And when you were talking before about job displacement, I was thinking about how the The autocomplete features in something like Google Docs have really advanced over the past couple of years They can't quite write my pieces for me, but they complete sentences in ways that I'm a little surprised by how prescient they sometimes seem.
Have you seen impacts from COVID related to AI and whether automation specific or not, that perhaps have surprised you that you think will really jump out when we get to 2041 and look back or just kind of things are moving a little faster than you might have thought pre-COVID?
Yeah, there are three types of accelerations that I can see. just the general focus on how healthcare technologies are becoming digital and intersect.
Yes. and more efforts to put AI into protein folding, drug discovery, collecting all kinds of digital content from mRNA to genetic sequencing and how AI can be used there because we're all much more focused on healthcare nowadays.
The second is work from home. And that essentially digitizes the workflow and the work stream.
And that allows AI to be plugged in to either enhance or displace people.
And I think we're going to see that as technologies like Zoom, DocuSign, and others causes companies to naturally have essentially a digital copy of the workflow upon which AI can operate and help them save and make money.
The third and the last area is in automation.
In order to have better social distancing, and protect workers.
Many factories in China used robotics as a way to do that. uh so that would be human robots working together in assembly lines and factories And we're seeing that making significant advances in surprising areas.
For example, by first doing a COVID robot, a COVID nucleic acid test robot a company has developed a robot that can be a almost generic platform for laboratory use.
So that automation becomes a surprising advance in automation.
And then that spills over to more industries in starting with factories, moving on to commercial and moving on to residential.
So in China, especially in China, we are seeing many restaurants adopting robot waiters that are not These are trays on wheels that navigate to your table so you can take the dish from it.
These are not fancy. They're Danny's like restaurants.
It's happening in China. And then at home, there are not just the iRobot-like robots, but there are many more robots doing things like delivery.
For my apartment building, all of my packages and the takeout food are delivered not by a human, but by a robot to my door. because it's easy to navigate in an apartment building.
So I think what's happened in COVID is For the US, I think work from home, workflow, digitization will cause US to be really strong and leading in enterprise AI.
And then in China, automation from factory to commercial to homes will cause China to have a the largest number and fastest growth in the area of robotics and automation.
And in both US and China, we will see healthcare plus AI make rapid advances.
If I might put you on the spot for a second, if there's one area, however narrow or broad, that you really hope outpaces your expectations for advancement in the next 20 years as relates to AI, what area would that be?
Would that be healthcare or something else? it would be a very close tie between healthcare and climate.
And these are also areas that we can hope that the AI giants, US and China, would be able to work together because it's for the common good of the human race.
Do you think that's possible? I think for climate it certainly is because that's become widely recognized. as something every country needs to contribute to.
And the US and China are both making this a high priority.
For healthcare, it's a little trickier because there are issues with personal data and the view that there is a competition there.
But there are also some efforts of collaboration.
So we'll see. And so the book is out and you're doing a book tour of some sort.
Is it virtual, in person, a mix of both?
All virtual. All virtual. Book tours in...
The early 2020s, I guess, are all virtual these days for a number of reasons.
What's next for you? You're still running Sinovation?
Yes. So time to go back to the- Back to the grind?
Yeah. Well, technology investments are more important than ever.
And so we're seeing many exciting tech projects.
We have even a larger percentage of tech coverage now.
So it's quite exciting. And is there another book project brewing perhaps for another two years down the road or not just yet?
You know, after each book, I would say no.
And then a year later. You'll get the itch all of a sudden and it'll start to take shape.
That's right. I need to have an idea next.
Excellent. Last thing, I guess. Somebody who's listening who maybe was on – well, wherever they are on the spectrum of being a neophyte or steeped in this and they –
They saw your name come up on the podcast feed and had to listen.
Any advice that you would pass on to somebody who's thinking, okay, 2041, 20 years from now, most likely I'll still be around.
I have time to change careers or change my habits or maybe start learning a coding language on the side.
Any advice you might give to somebody thinking, what can I do to prepare for whatever AI has in store for us 20 years from now?
Yeah, I think AI will impact every profession, but it doesn't mean everyone needs to go and learn to program AI.
AI will manifest itself in tools. There will be tools that help writers and reporters, as you discussed. helps you write better.
And there will be tools to help accountants, lawyers, and scientists do a better job.
So to most professionals, the AI will manifest itself as tools that help each of us become more productive and amplify our capabilities, much as the PC and the Internet have.
So don't view it as something that's threatening.
View it as something that may be empowering, especially for the professional workers.
Excellent. I like landing on an optimistic note.
Kai-Fu Lee, the book is called AI 2041. It's available. everywhere.
You are also all over the internet as a man with your stand. and your track record would be expected to be.
Is there a website for the new book or where is the best place for people who'd like to learn more about the book and your other work?
To go online. Come to AI2041.com or just go to Amazon.
There are lots of reviews and content there.
Or follow me on Twitter. I'm Kaifu Lee. Perfect.
Kaifu Lee on Twitter. Well, Kaifu Lee, thank you so much. for taking the time to come back.
And whether there's a new book or not, we would love to have you back again two years from now, 20 years from now.
Whenever the time is right, come back and talk to us again on the podcast.
Okay. Thank you, Noah. Thank you. Thank you.
