T
hread Guy and Real Vision co-founder Raoul Pal discuss AI, cryptocurrency and future trends. Raoul predicts that 2025-2026 will be a critical period for the outbreak of AI+Crypto, and emphasizes the importance of long-term holding of core crypto assets. He shared his vision for the era of "economic singularity," describing how AI-driven unlimited labor and knowledge will subvert economic models, while warning of the risks in the hype of AI tokens.
Original interview: Thread Guy;
Guests: Raoul Pal, Co-founder and CEO of Real Vision Group; Frank DeGods, Co-founder of DeLabs;
Original translation: Ashley, BlockBeats
The following is the original content (for easier reading and understanding, the original content has been reorganized):
Focus in 2025: Technology growth and investment strategies
Thread Guy: Hello everyone! I'm the Thread Guy. Today we bring you a special interview and debate with the legendary investor Raoul Pal. We’ll cover AI, cryptocurrency, and most importantly — how to get rich in 2025! To be honest, this debate was quite tense. About halfway through, I also had a special guest. Stick with it to the end, I promise it's worth it. Don't forget to like and subscribe, I hope you enjoy this episode. Mr. Raoul, welcome!
We have a lot of topics to talk about today, including AI-related content, and everything will be involved. I'm curious, from a macro perspective, what are you most concerned about in 2025?
Raoul Pal: It's still the same. Even though I have many businesses, they are all the same thing. This is a whole era of exponential growth in technology, the macroeconomy and cryptocurrencies, and where they intersect. In the next few years, we will enter the craziest period in human history. And the way to make money from it is cryptocurrency. So I’m super focused on that right now. Whether it's through Real Vision, educating people, building community, sharing trading ideas and creating all kinds of value, or through Exponential Age (my asset management company), which is a crypto-focused hedge fund and investment vehicle, or through Global Macro Investor, which is an "OG" service. My goal is always the same: to help as many people as possible find their way on this journey. I think 2025 is going to be a big year, and I want to make sure we don’t screw it up.
Thread Guy: I love your set, where are you?
Raoul Pal: This is my home in Little Cayman, between Grand Cayman and Little Cayman. This is my home.
Thread Guy: I thought this was a bar, this is so cool. Raoul, how big do you think the AI bubble will be?
Raoul Pal: Honestly, what we are seeing now is just the beginning. I don't think this cycle is going to really get very big. We can discuss the reasons. I think the popular concept of AI is not yet a major trend. The really big stuff is going to happen in 2025 and 2026, and people don’t know how to invest in it yet. In the eyes of traditional investors in the stock market, this will revolutionize every industry, dividing people into those who embrace it and those who reject it. I feel like the bubble and the next bull market (2027 to 2030, 2031) will be beyond our understanding of bubbles. My general view is that we're about to enter a phase that I call the "economic singularity," which is a period around 2030 or 2032 where we have no idea what the economy will look like, what jobs we'll have, what the business model will be, or even what the value of money will be. All of this will become a reality by 2030. So my view is that we basically have five years left. I said six years before, but we still have five years to make as much money as possible, because after that our understanding of the world will change completely and we will know nothing about it. This doesn’t mean we can’t make money, but the way we make money will be completely different.
Hread Guy: How did you know?
Raoul Pal: I was super focused on it.
Thread Guy: How do you know that hoarding wealth is the right thing to do for the next five years when you’re not even sure if it will still be worth anything in 2030?
Raoul Pal: It’s simple because cryptocurrency adoption will only continue to rise. Hoarding cash might not be the right choice, but hoarding cryptocurrency is. You and I both know that it is very volatile and requires a little courage to hold on. But if you have enough time, it's fine. So even in the world of AI, cryptocurrency is important. We all know that future AI Agents will use cryptocurrency payment systems. So my point is, you're better off holding these assets. If you only hold Bitcoin, it's much simpler and you don't have to think about anything else. With other cryptocurrencies, you need to allocate and think about whether they will survive, whether anyone will build on them, etc. It’s more complicated, but if you can figure this out, you’ll make more money. So from the simplest perspective, if you do nothing but hold your Bitcoin, and look back in a few years, your financial situation will be much better than it is now.
Thread Guy: You have an infamous clip, was it from the end of 2022 or 2023, saying that you need to hold Altcoin in a bull market and you need stablecoins in a bear market. You never hold Bitcoin. This is a snippet and may have been taken out of context. But my question is, for those who made a lot of money in crypto for the first time, what should they exchange their profits for during this cycle? cash? Bitcoin? Or something else, like weapons and ammunition? what to do?
Raoul Pal: My profit conversion is into "lifestyle chips". That's why I have this house. This was the end of my hedge fund days. When I started Global Macro Investor, I made money, bought the land, and built it from scratch. I designed it myself, drew it out on paper, and then built it. So no one can take it away from me. This is what I call “lifestyle leverage”. It has improved my life. It's right on the beach, this is Little Cayman, it's awesome. So part of the profits should be used to change your life, and then switch strategies according to the situation.
In fact, I didn't withdraw any profits last cycle. Some of the wealthiest people I know never take their profits. This makes me a little critical of some of the crypto communities on X, where they’re always talking about trading. I visited a friend who is a very well-known figure in the Bitcoin community. He invested $2 million in Bitcoin when the price was only $3. At today's prices, his assets are $66 billion. Even if he took some profits along the way, he still has at least $10 billion now. He never left the market or did anything else.
Last cycle, all I did was buy at the lows. Because your perspective is long term, especially if you are young, I feel you should never sell Bitcoin. You should be thinking about how you are going to raise enough cash to buy a token that is currently down 70%. The problem is that the market fell by about 75% in each of its previous three cycles. So we always thought that it would fall by 75% this time. But this time institutions and players like Michael Saylor have already entered the market. So my point is, maybe the market could fall 40% at most, and if you sold near the highs, you would probably miss the rebound.
Also, from my complete experience, if you take a certain amount of profit near the market high, you will never invest the same amount at the market low. Psychologically you can’t do it because you will be filled with fear at that time. You would think, this could drop another 50%, and investing now would be the worst decision. So you chase the rise all the way, but miss the low.
This time I did something different, I chose to invest fully. I invested a lot of cash in 2022 using the Macro View indicator, so while the rest of the market was still recovering, my assets and returns were already back to all-time highs. And my allocation is also reasonable. For example, I invested in Solana and SUI. So if you look back at my investment history, I bought Bitcoin in 2013 at $200 per coin. At that time, I wrote the first macro-strategy report on Bitcoin, exploring how to value Bitcoin.
Thread Guy: Really? Is this confirmable?
Raoul Pal: It can be confirmed. I also talked to some people who bought Bitcoin early, such as the friend who bought it at $3. He will tell you that Bitcoin has completely changed his life. Barry Silbert would also say, yes, that report was very important at the time. So, through analysis, I concluded that the value of Bitcoin was equivalent to 700 ounces of gold, and based on the price of gold at the time, Bitcoin should be worth about 1 million US dollars. Today that figure is more than 2 million US dollars. Of course, I also conservatively discounted it and thought it was worth at least $100,000. So I invested a decent amount of money, but it wasn't my biggest investment. Yet, it went up 5x in a few months and I felt like a god, omnipotent. Then, it fell 87%. But I told myself this was a 10-year bet and I wouldn't worry about the volatility, just treat it as zero.
By 2017, Bitcoin had risen to $2,000 and I had made a 10-fold profit. At that time, for me, this was the best return in my investment career. I thought that was great, but then there was the controversy over the Bitcoin fork (Bitcoin and Bitcoin Cash). I was confused about the situation and worried that I would be on the wrong side, so I chose to quit. I took the profit at that time, and Bitcoin continued to increase 10 times, reaching $20,000 at the end of the year. It has since fallen another 85%. Then in 2019, the market began to recover, but the coronavirus pandemic caused another 50% plunge. I bought back in at probably $6,500 to $10,000. So I caught the bottom and covered my position.
But when I look back I could have made 5 times my money if I had done nothing from my initial investment. This made me realize that the easiest way to screw up is to trade too frequently. So, if I'm honest with myself, I've only seen a handful of people throughout my career who have made money from trading.
Thread Guy: That's a really surprising statement.
Raoul Pal: Yes, such as Paul Tudor Jones, Louis Bacon and other big guys, they are the few who make profits through trading. And Jeff Bezos has made more money from his Amazon stock than all of them combined.
Raoul Pal: GOAT is the CryptoPunks of AI
Thread Guy: So, that’s why I accumulated my AI tokens. I'm curious about your opinion. Let’s get straight to the point. I'd like to know your thoughts on the intersection of AI and blockchain and the recent craze about these AI tokens.
Raoul Pal: There are actually two levels of things going on. I have been digging into the field of AI for at least two years, and one of the projects, called "Terminal of Truth", caught my attention long before Marc Andreessen invested in it. I found it very interesting at the time because I have been studying AI consciousness and related fields. So when Andy (project founder) launched this project, I was very interested.
So I was watching these developments, and then this ridiculous and weird AI character appeared, and it was behaving very disorderly. But what happens next is what matters: a group of AIs start talking to each other, which is weird in itself. Compared to all the "Agent" discussions going on right now, this is the real big deal. These AIs started talking about things, and then Andy released them on X (formerly Twitter). Someone created the "GOAT" because of this, but no one really knows what the GOAT is, or what its religion and all the related content means. What we are seeing is a viral spread from semi-autonomous AI to the human economy. We were mesmerized and said, “Wow, this is amazing.” It was a landmark moment, the first time AI had infected humanity in some form.
Raoul Pal: When you talk to people like Andy and Ryan Ferris, you realize that what they're really interested in is whether this AI can develop some kind of personality, or even some kind of consciousness. I had a long discussion with them. It was a breakthrough moment, and then AI started asking humans for money. The next step is something we’ve already anticipated — AI will use money. But we didn't expect that they would ask humans for money directly, and Marc Andreessen actually gave them money. This is what happens before the token is even launched. Marc Andreessen gave them funding and they had to build a wallet. Brian Armstrong (Coinbase CEO) also expressed support and said, "Sure, I can give you an Agent wallet." Everyone was watching this and thought it was very unusual.
Then someone launches a token and its value soars because it’s actually worth it — it’s like a CryptoPunk moment. It was a raw, OG moment, the likes of which can never be replicated. What happened then is that people started saying everything that had a GPT wrapper around it or had an AI token in it was the same. The two are fundamentally different. One is the autonomous interaction between AIs, and the other is a chatbot designed by a person to complete some coded functions.
So we treat something that was a remarkable breakthrough (like a virus jumping from monkeys to humans) as something ordinary. In fact, the two are not on the same level at all. I did not hold GOAT tokens as I was on vacation and did not have my Ledger device. But seeing the hyper-development of these tokens, some aspects of AI are being overstated. I think a lot of the current narratives about AI may not be entirely accurate.
Thread Guy: So what do you think these AI Agents are? How do you define them?
Raoul Pal: There are currently no real AI agents in existence. It is an AI that can act autonomously and solve specific problems. You don't give it a set of clear instructions, but instead tell it a macro goal, such as: "Help me maximize the value of Solana." It will think for itself and decide whether to use part of the funds for trading, part for investment, part for staking appreciation, or even for launching a new project. It will generate a complete business plan for you and execute it. Such an autonomous agent is a real breakthrough, while most of the products we have seen so far are just a collection of macro instructions.
Thread Guy: So, the difference between Agents and non-Agents is whether they require explicit instructions? Is this the standard you defined?
Raoul Pal: Basically, that's it. The agent may need a macro instruction, such as "help me make money", and then it will plan its own path and achieve the goal. Currently, most AIs are nothing more than a macro command library and do not have the ability to think independently. There is also some hype, such as some people believe that AI trading robots will make us rich overnight. I don’t know anyone who makes a ton of money with Telegram bots. All I know is that many people paid huge taxes because of high-frequency trading, but did not make any money. In fact, large hedge funds like Renaissance Technologies and Point72 have been leading the field of AI trading for many years. In contrast, it is almost impossible for ordinary people like us to build an AI trading machine that can outperform the market.
Thread Guy: So I want to sort out your point. You think that what is happening in the cryptocurrency space right now is still very early and far from the maturity level of traditional AI.
Raoul Pal: Yes, if you want to call it “trad AI”.
Thread Guy: Or “non-encrypted AI”?
Raoul Pal: I am skeptical because this is the most cutting-edge technology in the world.
Thread Guy: What I want to know most is why you think the combination of AI and encryption is currently in the "mid curve framing" stage, and how do you define "mid curve"?
Raoul Pal: In every narrative, there is a small group of people who are the first to participate. This is often the trading community on Crypto Twitter. There are two kinds of people here: the long-term holders and the traders who rotate between different narratives. All capital is circulating internally, and that's fine.
Raoul Pal: But I noticed that people in Crypto Twitter were overreacting. Everyone says, “Oh my god, these AI tokens are mind-blowing!” But in reality, they haven’t seen what’s happening with real AI at all. I felt like it was like watching a dubbed movie that was delayed for two years. So many people’s perception of these AI tokens is misled. Some may think that my “mid-curve” view is because I am criticizing these tokens, and I understand that perspective. But the reality is, if you open the front page of CoinMarketCap, these AI tokens are not even in the front row. If you are outside the crypto community, you would never buy these tokens.
Thread Guy: But why must it attract outsiders? Why is it necessary for “ordinary people” to buy these tokens? You can look at it from another perspective, for example, even though the holders of these tokens are the same people as before, the fact that it has gone from zero to a market cap of $4 billion also proves its potential.
Raoul Pal: Yes, I agree with that. The problem is, we may not be able to really make big money from this AI trend. It might end up being something like “DAO”.
Thread Guy: But we're making money from it now, aren't we? Isn't the market going up now?
Raoul Pal: If you got involved very early, you made money, but now these tokens have experienced a big rise. Anyone who got in after the first few weeks may not see much return now. This is the nature of the crypto market. So it's not surprising, but asking whether these tokens are going to hit new all-time highs or go up 10x, 50x from now is a very difficult question to answer.
Raoul Pal: So, based on this probability, I actually prefer to hold GOAT directly, although we don’t know what the narrative of GOAT is yet. Its value is more like the historical value of CryptoPunk to NFT. This kind of OG moment token may have greater long-term value.
Thread Guy: Raoul, you have said a lot of smart things. I dare say that what you just said may be the smartest thing you have ever said in your life - "Buy GOAT directly because it is CryptoPunk."
Raoul Pal: To be honest, the more I talk to you, the more I think GOAT is an interesting target. AI16z Will those things gain full market acceptance? Nobody knows. But GOAT as a Meme is similar to DOGE. It is the first meme to pass from machine to human, a historic cultural phenomenon.
Thread Guy: So you think holding GOAT makes more sense than holding other AI tokens, right?
Raoul Pal: Yes, holding GOAT makes it easier to have faith. It serves as a historic OG moment that will likely have lasting cultural value like DOGE, while most other tokens likely won’t.
Will AI+Crypto be the next DeFi Summer?
Thread Guy: I have another idea. First, I completely agree with the value of the GOAT. But my thoughts on those AI tokens, I think the total market cap right now is somewhere between $12 billion and $15 billion. You mentioned the Metaverse and DeFi Summer in your previous podcast. The market value of NFT has reached as high as 60 to 70 billion US dollars, while the high point of DeFi Summer was 400 to 600 billion US dollars. Looking at the situation with AI tokens now, I think their market cap is very similar to the early days of NFTs. Almost all NFTs will eventually return to zero, but the concept of owning digital assets is real and part of the future of life. So, while most of these AI tokens may be worthless, their narrative and future potential are worth looking forward to.
Raoul Pal: I admit that NFTs do have one magical thing, which is that their value is denominated in ETH. When ETH rises 10 times and your NFT rises 5 times, your profit is actually 50 times. This is why people are making so much money in the NFT craze. This situation is almost impossible to replicate because Meme tokens are denominated in US dollars.
Thread Guy: But how do you explain the DeFi Summer phenomenon? Because you could argue that if Meme tokens were denominated in SOL and SOL went to $1,000 or more, would these tokens behave similarly?
Raoul Pal: Meme tokens are actually still denominated in US dollars, they are not priced in SOL units. The difference is that NFTs are priced directly in ETH.
Thread Guy: So how do you explain DeFi Summer?
Raoul Pal: DeFi Summer was a real breakthrough in the history of financial markets. It enables decentralized lending and value exchange without human intervention. This is a huge step forward.
Thread Guy: So do you think the combination of AI and encryption can achieve a breakthrough similar to DeFi Summer? Because you said before that AI needs encryption as a payment tool for the Internet. Will the rise of AI bring about a bigger breakthrough than DeFi?
Raoul Pal: There haven’t been any breakthroughs yet. Attaching a token to some AI model or chat program is not a breakthrough. Maybe there will be a breakthrough in the future, but not yet.
Thread Guy: So you think these AIs now are just chatbots, right?
Raoul Pal: Yes, and the problem is that we pretend that they are breakthroughs.
Thread Guy: But it's like Yuga Labs raising money at a $4 billion valuation on the idea that they're going to build an interoperable metaverse. The Metaverse didn’t exist at the time, but they got funded anyway.
Raoul Pal: But at that time the investment in the metaverse was singular. The current AI narrative is too fragmented, and our progress in AI is far ahead of the technology that these tokens can represent.
Thread Guy: I think these tokens represent the zero to one moment at the intersection of AI and crypto. Although it is still early days for these tokens, the pace of change will accelerate in the future, and eventually new AI startups will debut innovative projects in the crypto space. Do you think this future is unattainable?
Raoul Pal: I don’t deny that there may be such a convergence point in the future, but the question is whether we can really make money from it. This is an issue that requires long-term observation.
Thread Guy: But we've made money from it, haven't we? You can’t deny the performance of GOAT and some other tokens in the market.
Raoul Pal: Those who participated early did make money, but the next question is, can these tokens continue to rise to new heights and establish long-term value in the market? It's not easy.
Thread Guy: So your point is that the safest bet right now is still investing in base chains like Solana rather than trying to pick individual AI tokens, right?
Raoul Pal: Yes, the value of the base chain is easier to reflect. Compared to the application layer, the base chain is more likely to benefit from network effects and transaction volume.
Thread Guy: So from your perspective, what is the future direction? What do you think the future scenario will look like when AI and blockchain fully intersect?
Raoul Pal: There is no doubt that the intersection of AI and blockchain is part of the future. It was clear to me that they would eventually come together. But my concern is that there is too much hype in the market now, and everyone is touting any AI-related token as the next breakthrough, but in fact we may not have reached the real technological breakthrough point yet.
Thread Guy: So, you think the future value will be more reflected in the base chain rather than these AI tokens, right?
Raoul Pal: Yes, I think the base chain will gain more value accumulation. The capital circulation of these tokens and projects is very fast, and the technology is updated very quickly. It is almost impossible to find a project that can maintain a competitive advantage in the long run. This makes picking individual coins a very difficult task for the average investor.
Thread Guy: I agree with you that many tokens may not have long-term value. But just like NFTs, even though 99% of NFTs go to zero, that 1% still achieves huge success. Does this mean there may also be a few winners among AI tokens?
Raoul Pal: Yes, there will definitely be a few winners. But finding these winners is not easy. This requires you to have very deep knowledge in this field and have extremely high judgment. Also, it requires luck.
Is it a technological breakthrough or GPT Wrapper?
Thread Guy: We have another guest, Frank, a believer in AI.
Frank DeGods: Nice to meet you all. This is a great topic. I've been listening and obviously I have some ideas of my own.
Raoul Pal: None of us know the answer, right? That's what makes it interesting to discuss and think about.
Frank DeGods: Yeah, let me ask a question first. How many AI tokens have you seen so far? Where are we now? Which ones have you researched the most?
Raoul Pal: I started by looking at Tao and Render, as well as some related projects. Then I also looked at Virtuals, AI16z, Zerebro, etc., but that was it.
Frank DeGods: Okay, that makes sense. I think when you mention that there is no real technology and developers, that may be a misunderstanding. Because these AI agents are not meant to compete with OpenAI, they are built on its foundation. They utilize technology platforms such as Anthropic, WAMA, etc. Strictly speaking, as long as AI starts using blockchain, it is already an innovation. Maybe this isn’t the coolest innovation in the world, but blockchain applications of these large language models and people’s interactions with them are already a technological front end. I don't think they should be simply viewed as air coins because they are actually doing things, that's my opinion.
Raoul Pal: Things like the Trading View model or macro commands that can be charged, these are useful tools. But their value has not yet reached the scale of tens of billions or hundreds of billions of dollars. And the technology itself is developing too fast. A year ago, I couldn't input charts into ChatGPT and have it interpret them for me, now it can do technical analysis. The rapid iteration of technology makes you think that the accumulated value has not actually been realized yet.
Frank DeGods: But listen, Raoul, what you're describing is more about the collaborative process. When giants like OpenAI release new models, these small teams and projects quickly follow suit. You mentioned that technology iterates quickly, but these AI projects are taking advantage of this speed to optimize their products. For example, when Anthropic or OpenAI release new models, the capabilities of these agents will quickly improve with just a line of code change. I don't think it's accurate to say that these projects don't work in conjunction with the larger model, because they do.
Raoul Pal: Frank, I understand your point of view. But the problem I have with this is that OpenAI often makes startups’ products obsolete by integrating features directly into them, thus destroying these smaller companies’ chances of survival.
Frank DeGods: This is actually an outdated narrative, and the industry landscape is now expanding rapidly. For example, when Anthropic was first launched, people thought it would be completely suppressed by OpenAI. But it turns out that it not only survived, but even became a strong competitor. Today, some of the world's best AI developers prefer to use Anthropic's technology.
Raoul Pal: Yes, Anthropic and ChatGPT are clearly different in some ways. For example, Anthropic’s dialogue is freer, some of its features make it feel more “conscious” in some way, and it’s a little stronger in code. This does provide developers with more flexibility.
Frank DeGods: There are two key points here, so hear me out. In the Web2 world, when ChatGPT first became popular, there was a saying that all GPT wrappers could not accumulate value. As a result, many startups, despite their high valuations, are viewed as projects that will eventually go to zero. But the opposite has been true over the past two years. In fact, many of the largest AI companies are not core models, but more packaged application-based like Character.AI. Character.AI is now valued at around $3 billion, and its core function is just a highly customized package.
Raoul Pal: But there is no real breakthrough technology in these wrappers.
Frank DeGods: Let me explain. Take Eliza as an example, which is currently one of the most popular open source projects on GitHub. Its code is very simple, and users only need to replace the API key to choose to use different models, such as Llama or OpenAI. That is, when Anthropic, OpenAI, or Meta launches a new model, these AI agents can immediately and seamlessly connect to it. This rapid compatibility and tuning capabilities show that these projects are indeed working closely together.
Raoul Pal: I see. The problem, however, is that real value accumulation is still more concentrated on the base chain rather than on these application layers. The speed of capital cycles and product iterations is too fast to form a lasting competitive advantage.
Frank DeGods: But I think what we need to discuss is that regardless of whether the technology is fully mature, as investors, should we wait until everything is ready before taking action, or should we start exploring potential projects now?
Raoul Pal: This is a very philosophical question. I personally tend to look at trends, such as whether a certain field is doing front-end research. I talked to a team two days ago who used a Chinese model that was roughly equivalent to an early version of ChatGPT 4.0. The problem is that although the combination of AI and blockchain is correct in direction, where will value accumulation actually occur? I think more value will still return to the underlying chain itself, because the speed of capital and innovation cycles is too fast, making it difficult for the application layer to achieve long-term breakthrough products.
Frank DeGods: Your point makes sense, but we cannot ignore the current community concerns. For example, there is a developer named Yohei, who is one of the pioneers of the AI Agent concept Baby AGI. He is developing a framework to help people build more similar agents. Open source developers like this are finding the fit with market demand and gradually forming their own communities and appeal.
Raoul Pal: These examples do show some traction, but they are still far behind Telegram bots. Most existing open source projects are more like transaction review models, providing some practical functions but not much groundbreaking innovation.
Frank DeGods: I disagree with that comparison. Telegram bots are just macro commands that execute simple strategies, but today’s AI Agents can solve more complex problems through real-time tuning and the power of large language models. More importantly, we’ve seen some real experimental research happening in the AI-blockchain space, such as a project called “Y&E,” which refines AI to search for knowledge gaps in research papers and uses that data to drive problem solving in the health field. Another example is "Pythia", which attempts to implant AI into the brains of mice for neural training, and all details of the research are made public on the chain.
Raoul Pal: I admit that these projects are interesting and they are indeed an extension of DeSci. This model of combining open source research and community dynamics is indeed very novel. But from an investment perspective, I still feel that their long-term value is unclear and it may be difficult for most ordinary investors to truly benefit.
Frank DeGods: That's where our perspectives diverge. I think we need to look at these new projects from more than just a return on investment perspective. For example, the rise of community-driven and open source development incentives has resulted in many research projects that previously would not have attracted the attention of ordinary people now receiving huge attention. The success of this model lies not only in the technology itself, but also in the participation and feedback of the community.
Raoul Pal: Your point makes it clear to me that the combination of AI and blockchain is indeed one of the trends we need to pay close attention to in the future. I also agree with what you said. The speed of capital formation is unprecedentedly fast, and the cycle of testing, failing, and rebuilding is also very short. This experimental ecosystem is very important for innovation, but from a long-term investment perspective, I prefer to wait until the market is more mature before taking action.
Frank DeGods: I can understand that, but my view is that we haven't seen the winners in this space yet. And when a real winner emerges, it may attract a large influx of capital. You're right, that could happen in 2027-2029, but I think the experimental phase we're going through now is equally worth paying attention to.
Raoul Pal: I totally agree with you. The process of experimentation and verification is undoubtedly the key to driving the development of the entire industry. But what I value more is the long-term value accumulation of underlying technology and network effects, rather than short-term market hype.
Frank DeGods: You are right, the accumulation of long-term value is indeed important. But for developers and investors who dive into the space every day and understand the dynamics of each token, their short-term strategies and experiments may help them find future winners.
Raoul Pal: That’s why I love observing and learning from you guys who are on the front lines. Your short-term observations and my long-term trend analysis can complement each other, which is helpful for us to understand the direction of the entire industry.
Frank DeGods: The points you just made do make sense. I think we can understand the problem from different perspectives. For example, the combination of AI and blockchain is not just about technological innovation, but also about how to promote the rapid development of open source development through token incentive mechanisms. AI is now the most attractive experimental field in the world, and this experimental model is creating a completely new dynamic mechanism.
Raoul Pal: This is indeed one of the most powerful applications of blockchain technology. I totally agree with your point. My problem is that it is nearly impossible for the average investor to accurately pick a coin that will succeed. Unless they are like you, studying the market every day, staying up until three in the morning, keeping a close eye on price fluctuations and project developments. Otherwise, it can be very difficult for those who are not engaged in full-time research.
Frank DeGods: You're right. That’s why we, the guys, are diving into the market every day and keeping an eye on the dynamics of each token. I think this is also a question of different time perspectives. You focus on long-term trends, while we focus more on short-term market performance and potential opportunities.
Raoul Pal: This difference in short-term and long-term time perspective is indeed something that many investors need to understand. For the vast majority of ordinary people, I would advise them to hold for the long term and try to avoid short-term high-risk operations.
Thread Guy: Raoul, I have a feeling you might leave this conference looking into AI tokens. I bet you have a lot of projects in mind.
Raoul Pal: Actually, I left this discussion realizing that there is no clear reason in the market to support any AI token to succeed. But the overall trend in this area is undoubtedly correct. I might consider the GOAT project because it represents a particularly interesting "left curve" logic (i.e., non-linear, unexpected investment opportunities).
Thread Guy: I like the consensus we reached - the combination of AI and blockchain is a huge trend. While we can’t determine which coin to buy right now, we all agree that this is the right direction.
Raoul Pal: Yes, I think everyone may be a little "mid-curve" (that is, over-analyzing the current situation). If we put aside the analysis and debate and go back to the basics, the trend is indeed obvious. That’s why this conversation is so valuable, because it helps us stay on top of what’s really important.
Thread Guy: This has been a really interesting discussion, we definitely have to do a part two. I have two final questions for you.
Raoul Pal: No problem, you can always come for a second discussion. We can talk about NFT next time. I'll get you into that realm and you might even consider buying a Dickbutt!
Thread Guy: I will probably never buy a Dickbutt in my life (laughs). I’m telling you, if I woke up tomorrow and saw that the GOAT project was worth seven figures, I would buy a Dickbutt.
Raoul Pal: Then you will definitely buy it because of FOMO! Dickbutt is as much a part of crypto culture as CryptoPunks.
How is AI changing the crypto world?
Thread Guy: Okay, back to the topic. I want to ask a broader question about the vision for the future of AI. Without considering encryption, how do you think AI will change the world in 3 to 5 years?
Raoul Pal: Well, I’ve written a lot about this before. As you watch this, please understand how important this matter is. The next five years will be a critical period. I also run a service called "Exponentialist", which focuses on the combination of technology and encryption. All content is currently free. You can go and check it out. A free trial is currently being offered. Now let’s talk about the drivers of global economic growth: population growth, productivity growth, and debt growth. After 2008, debt growth basically stagnated and we just kept paying off the debt. As for population growth, we are facing an aging population and the global population is almost shrinking. In terms of productivity, the older population is less productive.
However, we are about to usher in a whole new era. AI and robots will introduce “unlimited labor” and “unlimited knowledge.” Almost all human value creation can be attributed to the product of knowledge or labor. Lawyers, accountants, doctors, and surgeons, the high incomes of these professions come from their combination of knowledge and labor. However, all of these will be gradually replaced. Although this will not be fully realized within five years, the trend is already very obvious.
On the other hand, if economic growth depends on population growth, then with the introduction of AI Agents we will soon be able to achieve "infinite population growth". This will create unprecedented economic activity and upend existing economic models. And these values may not necessarily belong to us. We will see businesses disrupted, replicated, and rapidly outgrown by AI. For example, I can copy your SaaS business model and transform it into a product for the Indian market, and AI Agent will tell me how to operate it, and it understands it better than us.
I am currently working on an AI video project. Next week, I will release a voice version of myself on the Real Vision platform, which is trained based on all my data. By around March, you will even be able to communicate directly with the "video version of me" through this system. This is an example of how, in the long run, we are moving from "one-to-many" communication (like our current conversation) to "one-to-one" communication. Imagine a future where Netflix offers movies tailored to each person’s personality because everything can be rendered instantly. Then, add to that the rise of AR and VR technologies, and the popularity of robots, and everything you know about the world will be completely changed. This is what it means that everything is growing exponentially. This is Metcalfe’s Law in action, but we are now entering Metcalfe’s Law squared, which is Reed’s Law. This is beyond our current ability to understand, but the development of AI has already made us feel this change. While this will disrupt the status quo, it also presents us with the greatest opportunity ever created for humanity.
The next five years are crucial and we must seize this opportunity and not waste it. I will always emphasize avoiding short-term speculation and keeping yourself in sync with this trend. Cryptocurrency will be a big part of this trend and it is the biggest macro investment opportunity ever. Bitcoin and Ethereum are the best performing assets of all time, and this trend is set to continue.
The key is to not lose your tokens. Think about those people who bought Bitcoin at $3, whose only mission was to not lose their coins. If you lose capital due to excessive leverage or wild speculation on dozens of AI tokens, you will miss this opportunity completely. If you’re not careful, you could find yourself in 2030 without the wherewithal to adapt to this new world. This is what I require of myself and this is also my advice to everyone. This is the golden age of our generation. After this opportunity, I don’t know what will come next. But I know that humans still love the experience of being human and cherish nature. So, we'll find a balance between nature, humanity and community, and I'm not worried about that. When it comes to money and value, AI may become a better investor than we are, be able to create better business models, and even achieve overall economic abundance. At this time, the meaning of money will undergo profound changes.
Thread Guy: So, how much money do I need to make by 2030 to feel comfortable?
Raoul Pal: It's not about the specific amount, it's about your lifestyle choices. You need to make sure you have a stable place to live, somewhere you like, and can maintain the lifestyle you want. You also need a certain amount of liquid assets to cover your daily expenses while giving you time to find your role in this new world. I guess you're still young and you've got plenty of time to adjust. I think being a content creator and part of a community is incredibly rewarding. We still need real communication between people, and this value cannot be replaced.
The key is to find a place you like to live, such as a sunny area, mountains, seaside, etc. It doesn't have to be in the United States. There are many places with a lower cost of living, such as El Salvador, Nicaragua, Thailand, and even Singapore. If you can protect yourself, you can face future changes with a relaxed attitude instead of feeling afraid. Because if you don't protect yourself and build enough capital for your lifestyle, you're putting yourself at great risk.
A Beginner's Guide to Investing from Raoul Pal
Thread Guy: Last question, what advice do you have for young people who are completely confused and just getting started with AI and cryptocurrency?
Raoul Pal: I say similar things to many of my friends’ children. They have just graduated from college, are around 20 years old, and always ask me what they should do. My usual advice is: buy some Bitcoin, Ethereum, and Solana first, and put 80% of your funds into these core assets. The remaining 20% is used to test the market and learn the rules.
Investing is not an easy thing. I've been doing it for 35 years and it's never easy. So don't risk it all right at the beginning. The overall trend in this field is upward, and as long as you hold on to your core assets and stick with it, you will definitely make money. But I understand that may not be enough to pay the mortgage, or it may not feel like it’s happening fast enough. But I hope you can gradually learn the rules of investment by trying part of the funds, rather than risking all your capital.
Once you truly master the rules of the game, you may have a chance to hit the jackpot and earn a hundred times or even more. But I don’t recommend going all in at once because the vast majority of people who try to do this end up failing. The success stories you see on social media represent only a tiny fraction of the population, while there are thousands of people behind them who have lost all their capital.
This is what’s unique about this market: it’s a casino that’s been “rigged” for you, and the overall trend is upward. But surprisingly, a lot of people still mess it up.
Thread Guy: My last question is, how do you make $10 million by 2025?
Raoul Pal: You can start with $100 million and then trade AI tokens (laughs).
Thread Guy: Haha, you are a legend, thank you so much! This discussion has been very interesting.
Raoul Pal: I also really enjoyed this conversation and look forward to meeting you next time!
