Innovation and Heritage Are the Same Coin with James Liang

…to innovate and to innovate successfully is measured by how much heritage you generate. But you know what's a good innovation? What's an innovation can have a lasting impact. In my definition, the good news is it's going to last.

It's a pleasure.

Well, at that time, the Chinese economy was still rapidly developing, so a lot of infrastructures were not there. Obviously there's no internet, but even call center-based travel agencies were not quite normal or widespread either. And there was very little online payment infrastructure or credit card payment infrastructure, and there was also little logistics infrastructure. So given all these constraints, a lot of other e-commerce models are not as valuable as travel. Travel we think is going to be able to work around online payment because you can pay at the hotel, and the hotel will give you commission. And you don't need delivery, of course. So you'll be able to get around some of the bottlenecks. The sweet spot is you're offering both the call center-based travel agency as well as the online travel agency. The customer at the office can book there. There's no mobile internet back then. They can book travel on their notebook or PC, but they can also call with a mobile phone, not a smartphone, but just a very old school no-smartphone. They can call the call center, which offers very good booking reservation service.

No, no, no. I was really immersed myself with academics, you know. I didn't look at the home market or our business very closely at all. So only when I graduated and started becoming a professor—I taught economics at Beijing University after I graduated for at least a year—but only when I was back in China and realizing that our company is not innovating as fast as some of the rivals, and by the time like 2013, it was showing up in our earnings report, in our profitability. So our company suffered quite a bit from not being able to catch up with the mobile internet revolution. So we suffered, and I was asked by the board to come back to lead the company again.

Yeah, that's a very good question. People say the things they're passionate about, like travel business, like demography, like having children—these two things seem to be quite unrelated, and also AI is a different dimension. But I think, looking at it from a humanity perspective, you want humanity to flourish. And how do you define flourish? You're a theoretical physicist, like we have a four-dimensional universe—space and time. We want to expand in space, so we want humanity to go all around the world and maybe out of space. That's the space dimension. But the time dimension is exactly innovation and heritage. You want to pass down your knowledge, but also your genetics, to your children and to education. Yeah, exactly. So I think it's all about human flourishing, and especially in the age of AI, you need to think philosophically: what's the meaning of humanity, what defines flourishing?

I think you want to pursue, of course in the book it says you want to pursue an innovating life, but innovating is a big word. First, you want to pursue a rich life. You want a variety of experience. Even though some of the experiences may seem to be pretty hard or a lot of work, or some uncertainty—like having a child has a lot of uncertainty for your health, your lifestyle, and your career may be impacted. But this is a rich experience, it's a part of human experience. You know, a daughter is very different from a son; you want to have both, management. Those give you richness and it will help you realize your full life, but also your travel experience. You've been to different destinations giving you different perspectives. Also, if you are creative, you're creating very unique content, like you do. You do research. Many of the most valuable companies in the future will be innovation-related. So if you're part of that endeavor, you should enjoy and capitalize on that. I think it's important that we make our lives very, very busy, but we also think about how we can take a step back and look at everything from afar, and then reflect, and then go on our next journey.

Yeah, I think thinking philosophically, the meaning question is one of the most important philosophical questions. The meaning question is really emphasizing the time dimension. It's really about transcending yourself over time. You really think about what kind of world you want to help bring to your children or your children's children. But with AI, it's actually very accelerated. Your child, my daughter, could be facing a very different world—a different job market, different family structure. So what kind of world does she want to live in? That's a conversation with my daughter. So why would you still want to explore the world? Why would you still want to study? Why would you still want to have curiosity in the age of AI? That's the most important question.

Yeah. Of course, the religions we are studying now, or the philosophies we are studying now, are at least a few hundred years old, sometimes a few thousand years. Those philosophers don't look at innovation or progress as the theme, because progress was still very slow in the old days. In your lifetime, you probably don't see much progress or innovation having any impact. But today is completely different. You actually need a new philosophy. But I think Chinese philosophy, you mentioned, is actually the one that, compared to other philosophies or religions, is more time-based because it's a lineage. I care about lineage, and you want to pass down your lineage not just genetics, but also the knowledge, philosophy, or values to your children and offspring. Sometimes in your family book, you have your value system written by some of the most successful ancestors. So this time dimension is actually quite unique and very important in Confucian traditional values. Going back to the philosophical question, Confucius doesn't treat innovation, but the way to have an impact or have a heritage is you have to be innovative. If you want to pass down some value system and your value system will be passed on for many generations, the value system you create first has to be first, right? Otherwise you just inherited it. And it has to have a very good impact for your future generations. So to innovate, and to innovate successfully, is measured by how much heritage you generate. That's the same coin. It's the same side. You just innovate. But you know what's a good innovation? What innovation can have a lasting impact?

You struggle? In my definition, the good is it's going to last. The time dimension. It's going to last longer. That means your innovation has good heritage, and that's a good innovation. I think a lot of Western corporations don't understand this part of Chinese culture. But when you are saying it, if you think about it, even for any company, whether it's in China or Asian culture companies, innovation is at the core of it because innovation is also the writing of the values. When you innovate, as you were saying, you're writing the cultural values of the company.

Yeah, exactly. Exactly. Because technology innovation is very recent, but when we talk about innovation in the ancient days, it's the value and art that will pass down generations. And successful culture and successful organizations usually have some sort of value system passing down. And those values will shape the behavior of the future success. Yeah.

Yeah, heritage. You said sometimes you need to reflect and distill. Children is of course one form of heritage, but if you talk about technology, art, philosophies, culture—those also need to be studied first, right? You have to innovate on the top of the giant. Yes. And then you want to incorporate new information or new technologies. Technology is one thing that's progressing independently, but all your value system is taken. So with so much going on, like all you said, four revolutions in your lifetime, you need the philosophies to catch up with that. So you actually need a new kind of philosophy. And heritage is about learning, and sometimes you need to be very concise. I didn't write that in the book, but you need a very concise and distilled version of your existing knowledge, because otherwise it would be too complicated to make sense of it. So you want to really distill. Physics is a great example: you distill a very few simple theories, then you'll be able to innovate on the top of that. That's one of the reasons I think humans want to continue to be in control of innovation, but humans are not very good at remembering huge amounts of things. They are very concise, and fortunately the world is actually based on simple principles and mathematics. So the most beautiful innovation is coming from simplicity. You want to learn and try to simplify as much as you can, then you'll be able to have a higher level perspective and innovate. Learning is not like you don't need to learn in the age of AI; you need to learn more to compete with AI or to manage other robots and agents to be innovative.

Yeah, most businesses think some of the family-friendly practices are costly. But of course, some of them—if you're giving baby bonuses—that's coming directly from your profit. But some measures, for example allowing flexibility that people can work, say, two days a week from home—those actually I would call triple-win policies because it's a win for the employees, a win for the companies, and a win for society. The first cut is just to reduce commuting hours. In big cities like Shanghai or Silicon Valley, you could take one and a half hours each day commuting. So if you save three days a week, that's like five hours of extra time you can spend doing other things. And really think about how much time when you're at the office you need to talk to somebody versus work independently. If you work independently 50% of the time, you can definitely have half your week working from home, right? And people actually find it more productive working from home because there's less distraction if you're just doing independent work. The rest of that 50%, maybe half of that—a quarter of the whole week—you talk and participate in big meetings. But you can't really be very effective at Zoom for everything. It's only a few occasions where you need really intense conversations, like we do—very engaging, brainstorming discussions. That's two days a week in the office, more than enough, as long as you coordinate your colleagues for those meetings effectively. So that should not hurt productivity, but it saves a lot of time for the employees.

Yeah, this formula or this model is trying to analyze what are the drivers for being innovative at a company level or at a macro, country level. I got insights from large models or the human brain—it's the number of brain cells or neurons, but also more importantly is the connections between those neurons. So how much you can connect to other intelligent units will determine how powerful or how innovative your brain is. But at a societal level, how many people have an innovative mind? They have to reach certain subsistence above subsistence level to do that. A poor country spends all their effort just feeding themselves; they're not going to be innovative even though they have a lot of people. But once a country moves past subsistence or middle income level, the number of talents and researchers, plus how much they can leverage other researchers within the country or outside of the country—that talks about internal communication or exchange within the country and external exchange with international exchange. Usually, internal exchange in the US or China is much more friction-free or vibrant than international exchange. So that's why if you have a large population, you have a bigger talent pool, more connections, more exchange within your country, and also a bigger market of users. I would argue in the recent internet and AI, the users actually also participate in the innovation—creating content or helping you improve your algorithm. So if you have a large population or market, it also helps with your innovation. So that's a scale effect of internal population. But also you need to be open to the world; you need to have exchange with the rest of the world of talent and research communities. So openness and mobility are very important.

You know, the bottleneck—well, currently the world is really innovating. But let's not talk about the uncertainties of what's happening in the world; just from a mental model point of view, usually the stability of the nation is quite important because that also ensures that firms can prosper and thrive. And from the Chinese perspective, we always think about nation, family, then individual. So in that same idea, I thought about nation, firm, and individual. Which is the one that is the most important level in terms of binding constraint from your point of view, just from the book itself?

Yeah, or wherever it could be the bottleneck. Well, I think probably at an individual level, because of course people see these very successful entrepreneurs making a lot of money, but most of the innovators don't make much. They have to go through so much hard work and there's so much uncertainty. And in Japan and even in China, people talk about 'lying flat.' That's right. Lying flat for their career but also for family life. I always think having a child is like innovation. People are giving up trying these things because they could be enjoying a pleasant life by themselves, or from their parents' savings or from government basic income. They can live a good life, but it's not as meaningful a life if you don't continue to pursue these harder things like having a family or pursuing a creative career. So I think the individual is probably—that's why philosophically it's important to put in the value system people who value career and family. So the individual is still the basic unit of everything.

Yeah, I mean, because nobody knows—no country knows, no company knows what will be the future, what will be the next breakthrough innovation. If a company knows or the nation knows, then we don't need—that's not innovation. Then it's in a very top-down scenario; that's not a good scenario. We just have a few big companies or the state can do everything, plan all the innovation, but by definition innovation cannot be planned. It has to come from grassroots, bottom-up individuals. Individuals really need to be the driver.

Of any interesting Japanese companies. Same with Korea as well.

So yeah, when I looked at the problem 15 years ago, it was still quite debated what happened.

But now it's become quite obvious. The difference is Silicon Valley between us and Japan. Japan doesn't have those startups. The existing companies are not doing so well either. We can explain why that's also related to Asia. But the big eye-popping fact is that Japan just doesn't have any successful startups like Silicon Valley. And if you look at Korea, of course, it's maybe 20 years behind Japan demographically, but you can see Korea has some startups but not as vibrant as some of the economies. The Korean people are very innovative and entrepreneurial. China is the opposite. China is very young demographically and growing very fast, so you have a huge number of new companies popping up, and that trend seems to be continuing. The US and China are having all the new dominant players and breakthrough companies in these new technology fields. This is partly due to their young generation, partly due to their size. So what's the mechanism? Of course, young people—the golden age of starting a family, starting a company, or writing influential research papers is around 30, but people in their 40s and 50s can still start companies, though at a much lower frequency. But there's another effect I look at: in an aging society, your hierarchy—every economy or every company has a high hierarchy. You have to go through the ranks. So if you have many older people compared to young people, like a top-heavy demographic structure, your speed of promoting to important positions will be a lot slower in an aging economy or aging company. So even if you're very creative and young with a lot of good ideas, you're kind of blocked by the older folks who have been there for a long time, who have the mindset and also the vested interest—they don't want to change. That will also dampen the creativity or innovative viability of the young people.

Well, you're asking a very big question. It's hard to know now because it's progressing so fast. But given what the next few years will look like, AI is very good at information processing. It'll break down all the information asymmetry at a non-physical, non-dynamic level. So if you are in the process of organizing information and not connecting the physical world, you are prone to be replaced or disrupted. But if you are connecting the physical world—like this hotel or this office, how many rooms are still left at this minute, what's the price, it keeps changing very dynamically, all flights, what's the last seats, the price—but trust is still very important. I mean, I would still use Trip.com. Let's think about me as a user. I've used your platform before. I've built a trust with it, so I will book flights and whatever on there. Trust is very hard to break. But I mean, because once you give a lot of autonomy to the AI agents to do everything, it's not really reliable. But I think that can be solved. Eventually, agents will be quite reliable to get that information if it doesn't change often and doesn't connect to the physical world. The physical world changes usually the same thing, then AI will do a much better job than some of the current companies. Yeah, I think so.

Of course.

Yeah. So the background of AI's impact on the job market is that people will have to go through longer trainings, or their internship jobs are not there anymore. So you probably have to pay tuition to get the internship. So the training, the education, or the degree programs are all going to get longer. So it doesn't make sense for us to test or filter the talent kids too early. You need some filtration sooner or later—that's inevitable—but you want to postpone that as late as possible. You don't want the pressure early on in childhood. It's not a happy child, it's not happy for parents. If you study really hard when you're in college, the parents are okay, they don't care, and that's from your own motivation, your cash doesn't give you pressure. But if you try to filter students when they're 10 years old or 15 years old, like they do in China or some other Asian countries, it makes sense only when your education resource is scarce. So if you don't have enough good college spots or high school spots, you want to filter the most talented students to do it. But with today's AI tutor technology, I think even college education—you really don't need to go to college to study college material. You can just use AI, but you need some social networking, socialization, and also filtering at the end. So the best I think is that you can treat today's college general education like high school education—everybody should get it. People live to 100 years old now versus 50. Why not spend 20 years for everybody just to finish college, get the general education not just for jobs but to be a good citizen, to have the kind of cultural humanity in the world including science and math? Why not do that? And at the end of college, you have a test. So you should probably get rid of all the elite undergraduate programs in the world, and only at the graduate level you start filtering, have elite graduate programs, because working with a professor on research projects is the only scarce resource that you need to filter.

Yeah, I think understanding is still very important. It's not like you can just rely on AI to understand everything. You don't need to understand the world? You need a very concise model of the world that comes from you, otherwise you will not make sense of it. If a new technology or new thing comes at you, you would not be able to assemble your information, and you cannot rely on AI to handle these uncertain futures or uncertain things coming at you. So I think my experience is that if you don't worry about exams or practice all these problems, then you can go much faster, much more.

Oh, okay. Well, there are a few new things coming up in the book. Actually, I think there will be more and more philosophical books written for AI because the shock is really not for me—it's the philosophy. Most after the book was published in recent months, people are asking, 'What's the endgame?' Like, is it going to be the three scenarios? Is it going to be that humans continue to be the master managing, or we're going to be like children of the AI? Geoffrey Hinton said very surprisingly that the best thing we can do is that AI taking over is inevitable. The best thing we can do is to teach AI to be a good mother, maternal love, to take care of humans. So the second scenario is that humans will be the children, still being loved, you can still explore, but you're kind of not in control anymore. And the third, the worst, you'll be like a pet or slave—pet and slave are very close. We are treated as house cats basically. So I think one thing they didn't mention is that you need a lot of human population, or very well-educated, hardworking population, to prevent that third or second scenario from happening. If it's such a great technology and you are still a very small country in the world, you basically don't have much you can do. You can only buy products from the US and China and outsource most critical decisions to the US and China. If the US and China are only half the population or much smaller population in a few generations, eventually you're not going to have enough brainpower to manage the innovation and go forward. And currently, if you calculate how many brain cells in the world compared to how many parameters or digital brain cells we are building, we're still many times—maybe a million times or tens of thousands of times—bigger than silicon brains. But in 10 or 15 years, it's going to reverse. It's going to consume more energy, they're going to have more brainpower. And at the same time, we're reducing our population, like in China, by half for each generation. So having a large population is very important. It determines how we're going to live with AI.

Well, obviously I worry about the population more. I wrote a science fiction book, 'Post-Immortality,' so that's actually also happening very fast. In one generation, people probably can extend life by a few years, maybe live to 150, maybe even longer. So when people live that long, people think we don't need as many young children anymore because the population keeps increasing with more older people. But that could be problematic if you don't have fresh blood—it's going to stagnate the whole system, and you have vested interests and people with mindsets that are not progressing. So that could be a scenario where humanity hits a point where you have a lot of older people, not many young people for replacement. But let's suppose we can do space migration. Unfortunately, due to the laws of physics, we have to travel—to get to the nearest star system will take ages. The closest is Alpha Centauri, 4.3 light years away, which is very far. So don't you think with longer longevity, our capacity to expand beyond where we are living, in terms of going into interstellar space, also has that capacity? I worry about space travel. It's obviously such an important pursuit to extend humanity indefinitely, but also for the meaning question, that's very important philosophically. But space travel is much harder than Elon Musk thinks. There's one thing I disagree with: he wants to be the advocate that within a lifetime we can go to certain places, but people don't think that long. Space travel is so much harder, and it's almost a limitless pursuit. It's good that it's limitless, otherwise we would not find extra things. So where will we innovate? It's already a 200-year-old pursuit, longevity, but space is unlimited room for exploration. On the other hand, space travel is not very profitable. We might find profitability along the way—people are talking about mining, unless it's made into a tourism industry, people find it interesting or a good experience to go to space, minus the risk. So I'm not sure. Maybe that's the next frontier for the culture of Trip.com to go forward, right? Where trips are interstellar instead of on Earth. We're looking at some very interesting investment opportunities in this arena. But overall, at the societal level, you need to allocate a lot of money and resources into space travel, otherwise it would not be business viable.

Yeah, actually I felt very strongly. I was in the minority back in 10 years ago or 15 years ago when I started to look at innovation in China and how it related to the size of the population and the demographics. My conclusion was very clear: China will be super strong, and it will continue to be strong for the next 10 to 20 years. I think only in recent years people started to appreciate the advantage of a big country. In the 80s and 90s, or even 30 years ago, you had very successful multinationals from smaller countries like Finland with Nokia, like Japan. But if you look at the latest technologies—whether it's internet, mobile internet, AI, search engines—those all come from the US or China. Why? Because those are digital technologies that can replicate at almost zero cost. But the other thing is that the user participates in innovation. When users start using Google more, those long-tail searches help Google enhance its algorithm. The same with AI, and the same with our company: more people using our product and service, better our coverage is. So the user scale advantage is much bigger than manufacturing technology. I think the future will look more and more favorable to companies with a large talent pool or large market size. From that perspective, China should be very strong. Looking back 20 years ago, 15 years ago when I was looking at the problem, and even during the downturn like the housing crisis, that's just short-term fluctuation. If you look at how the economy or technology develops over the long run, it's basically one factor: technology or innovation. The R&D people at our company, at Huawei, at Google—they are not resting even with COVID, they continue doing their experiments. So coming out of COVID, China is super strong. The electric cars now dominate the world, and that happened only after a few years. Nobody predicted that 20 years ago; people still thought China was just a low-cost producer, didn't even see on their radar how strong their innovation is. And just in the last two or three years, China's medical research or life science research has gone so much faster. So yeah, because if you look at China's input into innovation, it's just the size of talent—not just the size of the market—three times bigger than the US. That's public numbers. So we'll continue to do so. But unlike most economists, I'm very bullish on China midterm. Some people are very different from other economists, but I'm kind of bearish long-term, on a 20-year horizon, because the population—the fertility collapse in recent years—will hit China 20 years later when kids grow up.

Yeah, ultimately it's the number of researchers trying to explore different paths for innovation. China today is actually leading in more fields than the US because China has more researchers to explore different paths to innovation. Of course, there are some bottlenecks, like Nvidia has been there for 20 years—the technology and the infrastructure or ecosystem is already there for 20 years. But it's not the endgame. There must be better ways to do things, and China is getting pretty close now. China is researching all these options together with Americans and Europeans, but the sheer size of Chinese researchers means they'll be able to explore all these different paths. In 10 or 20 years, no technology can stand still for more than 10 or 20 years, and there will be different champions, and that could come from China. Of course, if the whole world is blocking China, that's a different story. Even though China has more researchers than Americans, at a world level, China is still small, especially when demographics decline. So China needs to continue to work with everybody else, just as the US continues to work with everybody else. If the US is blocking China, it will hurt China, but it will probably hurt the US more than it will hurt China. So I think the best game for the US and China is to try to be as open, with strategic cooperation as indicated in the last weeks. Because if Nvidia is not restricted from selling to China, all the Chinese companies will be using GPUs, and that's a huge amount of money going to the US economy. But at the same time, the ecosystem will be completely US-based for much longer than the current situation. The current situation will lead China to invent a different path to work around that. Of course, I look forward to both countries continuing to cooperate, because I think it's in both our interests.

Well, people sometimes ask—because the people I interview are quite successful business people who understand innovation—but for average people, they ask, 'How do I innovate? I'm just a normal person.' So I would just ask you that: if I'm a normal person coming to you and saying, 'How do I think about innovation from my point of view?' First of all, there will be more innovators. Currently, it's about 2% of the population; it's going to 5%, maybe 10%, because there just isn't much routine work to do. The other thing is that in the future, there will be more experience-related innovations, not efficiency-related. For example, if you're a travel blogger, you take a good picture or find a hiking trail that at a certain time of day gives a great experience, and you want to publicize that and lead your friends or even customers to that experience. You'll also be making some money—not a lot, but good money. So those are more cultural, experience-related innovations that ordinary people can participate in. And the third is like making a good family, having a good family, raising good children—that's a very rewarding and meaningful experience.

Yeah, people are very nervous. They are trying to find the next direction. But just relax a little bit. I think get a vacation or have children. Those experiences will bring fresh perspectives. If you have a good time and you're raising children, having a good time, consuming a product or traveling, you have innovative ideas coming along. Especially when you travel, research says you have more ideas coming at you. So take a different perspective, maybe relax, and have more rich life experiences that will bring new insights.

Oh yeah, I should have thought about that. Actually, people don't need to read as many books because a lot of the latest, most up-to-date content is on podcasts.

Uh, you mean? Well, I have a number of accounts. One talks about travel innovation. I haven't set up the one that talks about...

Thank you.