【深度访谈系列】AI时代狂想曲——我们可能是最后一代需要动脑的人

你对世界的未来有自己的猜测吗？我最近发现，低维度的人喜欢思考自身过去，中维度的人喜欢思考自身未来，而高维度的人在思考世界的未来。这位老哥给年轻人提供了一个高维度的参考答案，他的名字叫拉斯·特维德，而这些都是他写的书。在这个系列的上期，我们戴上了一顶他的思维帽子，叫“世界从未改变”，但拉斯还有另一顶帽子，叫“世界即将颠覆”。他对 AI、对健康有很多几乎极端的想法啊，让我听完大为震撼。20年后的世界会怎么对待年轻人呢？让我们分4块地来看看：经济结构、健康、地缘政治和投资机会。Welcome to the new world。本就如此。经济的形态决定了 AI 世界的赢家和输家。拉斯在《超级趋势》这本书里描绘了一张2050年的地图。站在山顶上，你会同时看到3个世界。This is one thing where my thinking also evolved when I was writing the book because at first I was thinking about how will the world economy be, and then I got to the point is which was there will be different economies, and and so I I divide them into the super clusters and the slow zones and the lost pockets. So the super clusters are this like Shanghai and Shenzhen, it's like Singapore, parts of Switzerland, the city of London, so on. It's frantic, you know, lots of innovation, people, you know, their adrenaline is flowing. But other people are not happy with that at all. That's not what they want. They want to live on a little farm on the countryside and see the flowers grow and talk with a cow called Flower and some sheep and so on. And it's very important that that we recognize that people are different and that in this frantic AI-driven and robot-driven economy, lots of people don't want to be in involved in that. They maybe they don't even want to see robots much, but they do want to live longer. They want better healthcare. They want cheaper stuff and so on. So they would like to have the benefits from this frantic technology evolution. But they have to recognize they're not going to be the ones who get rich from it because they're not making it, but they will be the ones who get access to the final products. And this, this is the slow zones. Yeah, gotcha. Yeah. So now we have, you know, the frantic places, the the super hubs, you know, where everything is happening. You have the slow zones, and then you have the lost pockets. So the lost pockets are the ones that just are completely outside of it, and it's it's a part of the economy where the ladder has been pulled up, so to speak. So if you take a poor African country, for instance, the ladder, the social ladder or the economic ladder for that is okay. So they can, you can put a factory, somebody from China or Europe or America can put a factory there because the salaries are low. But then when they work in a salary, they learn some skills and then after a while they deploy the skills in their own companies and so on. Essentially cheap labor in manufacturing. Yeah, it's cheap labor. Yeah. And the issue is that the cheapest labor will end up being robots, so you will have these places that are left behind. So, so I I think it's really the easiest way to describe a realistic outcome is that we will have these three kinds of economy. And when I say that we'll have a passion economy, what I really mean is that we'll have two economies in par in parallel. We'll have the hyper intelligent economy and the passion economy. And many of the people who are really into the passion economy will live in a slow zone. I see. Yeah, yeah. And the people who are really into innovation will live in these frantic clusters. But there are also some people who will work in the frantic clusters. But then they will take their vertical takeoff and landing aircraft and fly to the slow zones and wake up on the countryside while they're still working on quantum computing or whatever. In this discussion, I think it's very important also to be fully aware of the capabilities that machines are going to have. So sometimes, quite often, you hear people saying that we we are still needed for judgment. Machines are not creative. Yeah. Have you heard that before? Yeah, I don't think that's correct. So first of all, I think machines in virtually any field eventually can make better judgments than humans. And second, I think that machines are already incredibly creative and they will just be become orders of magnitude more creative. So it's not about that. It's not about that they eventually are superior to us in any particular way. It's just that we we can build culture and passion and purpose and find lots and lots of things to do there. Think about wolves and dogs. Today there are not many wolves in the world, but the the wolves got got turned into dogs by humans through selective breeding, and there are actually far more dogs today than there ever were wolves. So yeah, so the sum of it is I think that that humans will thrive, but the transition to it will of course be turbulent. If I were to just make sure I capture the analogy fully, we are the dogs in the scenario that you just described, right? So we will be the dogs of the future, or are we raising AI as dogs? We, yeah, so it is that we are the dogs of the future. Yeah, people should get that wrong. But dogs dogs are adapted to humans, and and humans will adapt to machines, and that means that because machines do the hard parts, we can do the soft parts. So a golden retriever is much more gentle than a wolf, because it it can afford to be more gentle. It actually gets rewarded from being gentle. 那么把镜头拉远，国家和国家之间，谁又能赢下 AI 这一局呢？拉斯的答案给得很直接：短期看美国，长期压中国。 We obviously know that AI is a heavily geopolitical topic, and there is a lot of competition between US and China, right? We talked a bit about incentives, but what factors do you think generally determines which country wins out the AI race? And maybe can I ask you to to rate China on some of the factors that you've mentioned? Let's say on a scale of one to ten. Okay, so summarize. So America is leading on frontier models, but Chinese models are instead of being a year behind, they're more like weeks or months behind. I would say very, very close, and will surely, in different ways, overtake the Americans very soon. And here are my reasons. First of all, the biggest bottleneck right now in AI is chips, and that's one of the reasons that the US has advantage. China right now produces about forty percent or forty-two percent of its AI chips. China has clearly a strategy to become completely self-sufficient in AI chips. Prediction is that it will be seventy percent of the AI chips by twenty thirty, and then you can just run the the trend forward and say by twenty thirty five it's probably a hundred percent. So it will it will have a complete competitor to ASML, the machines that make you know the foundation of all chip manufacturing. It will be able to make Huan Day and and another who will be able to make leading edge chips will have high speed memory chips. It will have everything by then. But there are two more things. So first of all, the AI plus strategy in China of spreading AI, getting AI out in society, and promoting that with open, open model, open weight or open source model. Yeah, yeah. That the combination of really wanting people to adopt small language models and agency AI and having an education system that from your six years old prioritise AI and where technical educations are more prioritised than in almost any other nation. I just think Switzerland has a lot of tech education also. But so you have you will have open models. You will have lots of people who have the education to take those models. And the third advantage is power. Yes. Yeah. So over the last few years, China has invested more in grid, electric power system, than all the rest combined. I haven't asked an AI why China started building out electricity system before AI became a thing. Maybe we should do that afterwards. So my suspicion is it was because of electric cars, and not because of AI. But whether it was because also of AI, I don't know. But in retrospect, it was a fantastic move. It was really, it was really smart to do that. And if you look at predictions for the US, even though the, I mean Donald Trump's administration is doing everything they can to provide more energy, the predictions are still that by 20, 20, 30, I think it is, they will have a massive shortage of energy. So now I mentioned the education system, the open models, and the electricity.But this number four, and the number four is really, really big, and that's physical AI. So physical AI is, of course, also what you call embodied AI. It's AI inside physical machines. Areas, serious prediction you can read from major research institutions or well reputed investment banks and so on. Predict that by 2050 there will be billions of thinking machines, many of which will be more intelligent than humans. Even though the activities in physical AI, especially in America, China is so good at making machines. I saw these insane numbers that compared the square meters of production factories in China, US, and Europe. Yeah, unfortunately I cannot remember the name, but but the number I think China is twenty times as big as America, and it's actually, excuse me, but I'm a little bit off here. But I think maybe five times Europe. So Europe actually has a lot more manufacturing than America. And America? Yeah. So America has has has has become super weak on manufacturing, and they know it, and they are trying to fix it. But these things are not easy to fix. For sure. Yeah, but China has that, and so I mean, the Chinese robots are already amazing, and and and they're far far better than they were two years ago. So China is ready for that, and China also has something that could be a blessing in disguise, as you say, and that is rapidly, you know, the prediction of a rapidly falling workforce. The reason I say that is can be a blessing in disguise is that if you know that you will be short of labor in the future, that gives you a very strong incentive to automate and to deploy robots and and disembodied AI as fast as you can. Not all nations would understand that or would be able to do it, but China is doing it. But we know that when you feel that there's an there's a really strong incentive or a really big threat, you do things faster. You you you invest more in it, you improve it faster and so on. Being aware that the workforce will be a problem will make China automate much faster than it would otherwise do. But the automation takes its own life. The robots can build robots, the AI can code robots, AI. It's like you're sinking the ground for something that will become incredibly powerful by itself after that. And that means that the standard of living in China, because of this strange dynamic, might become much higher than it would if you didn't have a declining workforce. So yeah, so it it can actually be a blessing to have the declining workforce. Economic 版图重绘的同时，健康版图也在悄然改变。拉斯这次端上了一热词啊，叫长寿逃逸速度（longevity escape velocity）。当科技每年给你续上的命比你老去的还要多的时候，人类就第一次跑赢了时间。Do you mind explaining what longevity escape velocity means and what your thoughts on that? Like given you've mentioned health improvements. Yeah, so when you talk about how long time people live, you you talk about the chronological age. I mean, how many years old you are. But then scientists also have a term called your biological age. So there are different tests you can take, and then you say that you are actually biologically older or younger than your chronological age. This this this concept that was launched some some years ago, which is that with the use of science, are able to to slow down the biological aging enough. Then for every year we live, there will be so much scientific breakthrough that you can add a year to your expected lifetime. That means that even as you chronologically get older, you're not getting closer to death. Death keeps evading you because of new scientific breakthroughs. If we then dig into what it actually consists of, it consists of a number of different breakthroughs. And so let's start with the things that are more mundane. You could say that is preventing disease or curing disease. Yeah. So different things that removes what would normally kill you. But in anti-aging, you're you're talking about a completely different thing. You're talking about understanding the essence of what makes you age, and and then going in and preventing that thing from happening, so that you actually do not age anymore. What we can say with very high degree of certainty is that within next ten to twenty years there will be so many medical breakthroughs that people will have the ability to slow down their aging and extend their expected lifetime substantially. Here you have to distinguish between two different success measures. So lifespan is one. You know, let's say you can live to 150 years old, and then many people say, yeah, but I don't want to, because I don't want to sit in a wheelchair for 60 years. It's not good quality life, yeah. And that brings the second term in, which is health span. And if you are able to implement anti-aging technologies, then your health span will be very long. Also, so it is about getting as many healthy years that you as you can. So maybe if you have a 150 years lifespan, you may have 130 years health span, and that means when you are 130 years old, you can play tennis and and party with your friends and so on. So I think it's very likely that AI will bring us this, and this is something I write about in the book that I I I don't hear people talking about. I actually I personally haven't heard anybody talking about, and that is that the biggest benefit we can get for AI might be time, and the time comes in two forms. The the one is the one we just talked about. Yeah. You have a lot longer time to live your life, but the second thing is that during the normal week and the normal day, because machines will do so many things that are just essential for survival. You will have time to pursue your passions and your purpose, much more time than otherwise. And lots of people, I think, have no time, like zero time, to pursue passion and purpose. But it will give lots of people a lot more time for these things. 好，趋势聊到这儿，最实在的2个问题来了，钱该往哪放？普通人又到底该怎么办呢？ Bringing back to investments, can you give us some sneak peeks on what your portfolio looks like right now, or what your investing strategy looks like? Yeah, so I am very positive on the AI business, so that's a part of my portfolio is very directly related to that. First of all, AI is a market where demand is much bigger than supply, which is a good thing if you are doing the supply. And I think it will be supply constrained at least past 2030, possibly much longer. Right now, supply is mainly hit back by lack of chips. In the longer term, it will also be lack of energy. The products are getting so much better so fast that they will disseminate into society. Demand will continue to grow exponentially. So, the money in the value chain is made in on the supply side, which as a whole is very profitable right now. And it consists of the hyperscalers and you know the data data factory suppliers and model operators on one side, and then the chip manufacturers on the other. So, yeah, so we are very invested in that. We had this nice dip, so we bought some more of that. Then another thing that I'm really, really bullish on in the longer term is industrial metals. So we talked about this long-term 20 years cycle in industrial metals. So the last ten years, the miners have actually reduced their production, which is completely unsustainable. It's completely out of whack with the prediction for for future demand. So we have future demand driven by renewables. You know, these are very, you know, solar panels and windmills, electric cars. They're all very heavy machines. Electric cars are very heavy because there's a lot of metal in them, but also by reammonds, especially in Europe, also by the acceleration in global wealth that AI will create. And it takes on average around 16 years to open a new copper mine, for instance. So what what they have done is they're squeezing the last bits out of many of the mines, but not really opening new mines. Interesting. So it's a consensus prediction is that we will have a shortage of a number of industrial metals over the coming few years, and that will take at least ten years to get out of that, and possibly fifteen. And last time we had one of these cycles was from 2000 to 2008, and there you saw mining companies go up typically four to six hundred percent. And so they have gone up some, but I think there's a lot more in it. So this is long time and long term, and you have to be diversified. Of course, yeah. There will be extreme shortages of uranium as it looks, extreme shortages of copper, shortages of lead, of aluminium, but also a number of other metals. So I think it's really good to have. Fascinating stuff. No, thank you. That that makes a ton of sense. For my final question, I wanted to bring it back into something maybe a bit more personal. So, last for context, a lot of our audience are like relatively young people in China, right? Who are trying to navigate their lives, and I'm curious, like basically, how would how should a young person prepare for the the changes that we we have coming forward? But I'll give two two pieces of advice. The first one is to get to know yourself. You can take a lot of self tests on the internet, and so they actually typically on a very high standard, or at least some of them are. So take a number of self tests, try to find out who you really are, because from that you can find out what is your passion, what what would you really be happy doing. And the second one is to make AI your friend, not your foe, and that means learn to work with AI at some at some level. I mean, the most sophisticated levels surely will be that you're able to make agentic AI and create new services, even build companies with AI. And and the lesser level is you just use it to make whatever you're doing more efficient by using large language models and maybe some other applications. So these two, I think, none of that is particularly difficult, but I think they can make a big difference. But there's also something in the attitude that you know, don't assume that there's no jobs for humans in the future. There's most likely lots of jobs, and they will you will earn a higher income, but they will be more interesting and they will be closer to what most humans actually want to do. So I don't think humans mentally are made for working at a factory doing the same thing all the time. Well, maybe a few are, but we are we are we are mostly biological optimized for some other work, and it will become easier for us to do things that we really thrive with. So to go into the future super scared and pessimistic is not a good idea. The world will most likely be far better than you think. That is very reassuring. Even if you think it will be far better, it will still be far better than you think. That that's always a nice highlight to to end on. Okay. 好，那么这期视频到这里就结束了。感谢看到这里的狗蛋，感谢点赞支持的狗蛋，也感谢拉斯参与到本次的访谈之中。希望大家都可以做个复杂世界的清醒人。我们下期再见。