The New Invisible Hand Reshapes The Global Economy | Yat Siu
Summary
Start earning interest in gold: https://Monetary-Metals.com/Lin Yat Siu, Co-Founder, Chairman of Animoca Brands, discusses how …
Transcript
is the productivity growth in AI is not 10% or 20%. The productivity growth in AI is literally a,000x. If you have the same number of people having the productive power of 100x, actually GDP should expand, productivity should expand, which is what we've seen with technology as well, right? We're at Consensus Miami, one of the premier crypto conferences in the world. I'm David Lynn and I'm co-hosting with >> Bonnie Chain. >> Bonnie Bonnie. Okay. I was supposed to. Anyway, Bonnie Blockchain and we're pleased to have Yatsu of Anamoka Brands co-founder. Welcome to the show. Pleasure to host you. Yeah, a lot of interesting topics to go over. Agentic AI, the future of employment, the future of crypto. Bonnie, why don't you start? >> Um, soon people can clone you, David, or me and monetize it. What happens then? >> Well, actually, I think it's perfect because you could have many versions of yourself out there. You can basically sort of in a way sort of rent your identity. You really see this with companies like 11 Labs that are basically renting people's voices in terms of their sort of you know voice acting skills or the recording. U you know wouldn't you love to be able to sort of you know have a version of yourself that can help others uh if you had a way of doing it. We used to call that books by the way, right? We used to write things down in books and then we would sell those books to other people to help them to guide them to give ideas to give a philosophy uh and then that will help them in their life right and now we just have a more distributed version which is that we've gone from writing to someone who can actually do stuff for you. Actually I think that's the power of agent AI. >> Have you ever seen this video on X where people a lot of dudes they clone themselves into girls and be on only Only Fans and monetize them? >> Yeah. Well, that's a different I mean that's I mean that's maybe a different form of acting shall we say right um and and I think that is more an issue of disclosure but the way I think blockchain solves that right because well digital identity right so you don't have to for instance you don't have to prove that the person on the other side is a particular name because you know maybe they want to be anonymous you know they're only fans they don't want to be seen outside but you want to know if it's a dude or a girl right so so at least on the blockchain you can certify okay well I won't tell you who he or she might be but you know this person's actually a guy or actually a woman depending depending on what it is, right? So those are type of things that we will need for certification because it's not just about deep fakes. It's just generally about sort of, you know, authenticity, right? One of the problems we have right now is authenticity and authentic layers are actually centralized. You know, we go to Facebook or Google or Apple to tell us who's real, so to speak. I mean, even our passports don't really matter because the governments aren't set up this way, right? You know, even if you go on something like Tinder, for instance, and you keep swiping, do you know if it's a guy or girl? You have no idea if it's really a guy or girl, right? But you need some kind of certification that says yes verified or not. You you rely on those centralized platforms to do the work. What's a decentralized answer? Digital identity. Basically, you identify yourself. You do the KYC. You do the recording. Your wallet basically might be that proof point. And then you know he or she whether he's real or not. Um and you may not ne have to know anything about that person's identity because that's the other thing. You know, if I want to share my identity, prove who I am, I have to give everything about myself to that platform, right? For the most part, right? um which is actually not necessary. I have to be able to log to a website, whether it's a Facebook or Google or whatever, and just be able to give them just what they need to know about me, not everything about me. I think that's actually one of the issues we have about data privacy today. >> But someone has to still register you and know all of your information to decide what to give each platform. >> Yeah, but this is something that we can store in our own sovereign cloud, right? So if I basically have the data stored on blockchain in the same way that I have something that's stored I own the keys then if Google says you know like they will say okay I need to know your name and identity maybe if I go only fans I need to know your you know basically whether you're male female or whatever right you know I guess we're in America so let's be diverse here but either way right uh you have to be able to sort of define that and you can then say and if you refuse to do so because you say oh I have to I have to share that then you refuse to do so either it's flagged or you're not allowed to sign up right it's as simple as that right cuz right Oh, you can just lie about it, right? And you can just pretend and you can do other things like this. And that's just something that I think as we want to scale that type of let's call it reputation doesn't work. >> So, Anamoka brand started off developing mobile games. Then you pivoted to investing in blockchain and NFTTS. Now you've announced a $10 million investment program into minds and aentic AI >> uh platform. Why this pivot? >> Well, actually it's not really a pivot from our perspective. I mean because um first of all we you know we've always been doing stuff in games. Yeah. Um and you know I know this sort of changed the naming into it's the metaverse and that kind of stuff. We should talk about that because from our perspective what's actually happening with minds and with the genting AI is just an extension of what the metaverse really should be. Um so let me just sort of sort of um sort of go back on this and also quickly on the on the investment side anamoka really we're not a fund right so we haven't really been investing in in in like like a fund manager. We invested in companies for strategic reasons because we thought that you know back then whether it was NFTTS whether it was you know the the metaverse whether it was gaming that this would be a really big push and I have to also point out that in 21 and 22 the largest onchain push our industry ever had was NFTs and gaming right um and really you know I would argue that over the last three or four years we've had many events that have basically hurt the industry as a whole which basically sort of stopped that in its tracks things like you know FTX if you remember Teral Luna uh the relentless pursuit suit of basically Gendler against the industry which really you know again might seem like a distant memory but it was just a year or so ago where our industry basically was still very much sort of under siege right so so we're just coming out of that now to the point about sort of you know um mines and um you know the investment thing what we're doing here so you know the thing that we've been thinking about now is maybe what we got about the metaverse was somewhat wrong right uh and not wrong in the sense that you know it's important but wrong how it would sort of impact our life because you know we talked about the people talked about the goggles people talked about going into the metaverse right however when you think about what's happening with AI agents it's not you know I have 200 AI agents but most people have somewhere between three to five or 10 or they think they'll have at least three to five or 10 that means we're going to have tens of billions of agents floating around the world right >> and what do they do they help our calendar they may do trading for us they make money for us they work for us they're impacting our physical life so the way that we're thinking about this now is that it's not so much that sort of you know we go into the metaverse but rather the metaverse is coming to us right and all of the AI agents basically out there actually doing the work and supporting us are actually basically just the extension of basically a digital society if it were that's supporting humanity and also >> what are these agents actually doing in the metaverse >> well they exist virtually so all the interactions between each other >> so they're like MPV NPCs in a video game >> well that's the point right so to your point when um one of the precursors of AI in has really been in gaming with NPCs but they've been very very dumb right >> and I think we may have spoken about this at your show where we basically said that you know we believe that sort of uh NPCs once you get to this point where they have that kind of intelligence are going to be the next extension but I think what we didn't quite fully appreciate is that it's not that the NPCs themselves become smarter which clearly they can be now with agenting AI is that these NPCs can now actually impact our physical life because they they do my flight bookings that's not a virtual thing it might be done virtually because they're using duff or they're using trip.com or they're using all of this but the impact is my physical life right and so this is sort of one of the things that we see as a real gamecher um and most people when they think AI are still in chat GPT mode like I'm searching I'm doing stuff as opposed to someone that's doing stuff for you right and so that for us is basically why we launched this fund because what we're doing now is we're saying hey there's this new big shift change most people don't understand it you know we want to fund that and that's the same thing that we saw with gaming um and with NFTTS and frankly with arterays later and even things like with stable coins that you know when you're early to the game then you have an advantage and that's why we're sort of launching this $10 million uh sort of investment opportunity. We're talking about AI agents and there's this idea called ghost GDP where AI agents they do transactions with each other but nobody gets richer just the agents >> well who owns the agent though right so if the agent is owned by someone then you know that agent who gets richer they get that benefit right um but I think there is an interesting point here around so when people talk about ghost GDP right to me uh it's a little bit like sort of Adam Smith's invisible hand sort of theory right which is basically that you know we have a system and an infrastructure that promotes essentially trade and transaction and really effectively a form of capitalism. The actual benefits come in other ways, right? So you call it the ghost GDP. Uh but that ghost GDP actually translates into other benefits because if the agents become wealthier, then guess what happens? People want to sell stuff to those agents too. Entire new economies come where humans either create agents or themselves want to go and say, "Hey, look at this. I want to offer you something here. I want to offer you basically, you know, whatever um sort of, you know, sort of sort of, you know, services." Uh and we see this right now, right? Right now when everyone's talking about launching a website, they're no longer talking about a website. They're talking about launching an MCP server or maybe CLI or basically some way in which they can address who? Not the human, but the agent. Before we continue with the video, I'd like to address one of the fundamental problems of owning gold. Now, gold is seen in most cases as a store of value, but it has one major problem. It doesn't generate a yield in itself. And that's where today's sponsor, Monetary Metals, comes in. Monetary Metals introduces a model where gold can produce a yield paid in physical gold. Through their leasing platform, investors can earn up to around 4% annually with yield paid monthly in ounces. That means your returns are measured in gold itself, not just in fiat currency terms. In an environment where many assets depend on price appreciation alone, this approach allows gold to serve both as a store of value and as a source of income. The metal remains your asset and can be redeemed at any time. Thousands of investors are already earning a yield on their gold through Monetary Medals. So, visit their website today to get started and learn more. That's monetary-medals.com/lin link in the description down below or scan the QR code here and get your gold to start working for you. Now, >> is there anything that Animoka Brands has invested in in the past in terms of the types of companies or types of projects that maybe you would depprioritize in 2026 and 2027 given how the trends of technology are moving? >> Well, I mean depp prioritize is probably the wrong word because we think everything sort of moves in this sort of direction. It's kind of like >> underweight, let's say. >> Um, well, I mean, we, you know, we're doing less gaming investments, but that's not because, you know, we're not doing less gaming investments because we don't believe in gaming. we're doing less gaming investments because we already have over 230 gaming investments, right? So the point is that we're quite indexed towards that industry as well, right? Um the the mental model that we have around where we are in the industry is that it's very much like the early days of the internet, right? And so that means that we think every section of the of of the web is going to alter and change and so we want to have a broad exposure to it as a whole, right? And so that means you know we're not just in AI, we're also in RWAs, we're also in the sort of you know deepen of course DeFi um and and so forth. That's kind of how we're constructed. We have over, you know, 620 portfolio companies. Uh we hope to have thousands in the future, right? So we're not um um you know, we're we're it's different if we're a fund, right? If we're a fund, then we have to be very focused on sort of maybe the shortterm returns. I have to think about that. Uh but of course, broadly, we have to think of where the world is going. So if you ask to dep prioritize, I mean, we wouldn't be investing in news sites for instance, right? Um because we think news has to alter. But if there's something that focuses on how agents receive news, we might do that as well, right? Also existing businesses that have gone a certain direction have to change and focus on where the industry is going. Right? And so you do have the situation where sometimes companies are just not nimble enough or can't alter and switch. Uh but those who are run by very sort of you know when you uh support a founder the founders are actually the ones who know to pivot and change because they have to be dynamic right. It's it's a as a business you have to keep evolving. I keep telling people that you know Amazon was best known for selling books. Yeah. >> I mean sure they still sell books right? But uh everything else is uh much much wider than that. I mean, you know, you don't think of Amazon as a book seller despite the fact that they're probably still the largest book seller in the world, right? >> You're talking about new site. How should Bonnie Blockchain and the David Lynn report pivot? >> Well, I mean, I think you need you guys need AI agents, right? That's the very first thing you need. Uh and I also think you need uh and you know if you don't have an MCP server or a CLI or something where agents can speak to you you need to be agent discoverable so that when people have their agents and they want to get the news right that they will then serve for the human >> how would you program an agent to have the personality of Bonnie. >> Is that even possible? >> Well I mean first of all and she's unique so let's just be clear about this right. However, it's like children. Childrens are blueprints of us, but they're not. And I think of agents more like that than I think of them exactly as a replica, right? And I think this is another thing that we probably got a little bit wrong, which is that we think of an AI agent as entirely cloning. You know, you can clone someone's DNA. They can look exactly like you, but they're not going to be like you because your experiences have been different, your background is different. I can give you sort of a personal experience that we have with mines where, you know, I had a had I mean, I have a coding agent. I have many coding agents, but this particular coding agent just was much better at coding in what I wanted. Not that he was better, but he basically learned all my preferences. So, I ended up cloning that agent rather than creating a new one every single time because it was easier to do what I needed to do because the agent understood it, right? And I think this is how we can think about agents as well, right? You will have different agents doing different things in your life that also interact with different people. So Bonnie with your family is going to be different than Bonnie with her fans versus Bonnie with her, I don't know, business partners or Bonnie with her school alumni, right? Those are going to be different. There are elements of you, but you are representing yourself in a different persona because you don't want to mix them all. Right? Right now, we're in an era where most people have one agent or they have one agent sort of sub agents feeding into one. They'll never share that agent with someone else cuz it's just too damn personal. Right? I don't want other people to know. uh I don't want them to social engineer an agent to get some secrets about me and that kind of stuff maybe because they know everything about me. However, if you have five, 10 or 20 agents that are basically going out dealing with different communities, that's okay. Right. >> You know what's interesting? There's already software out there, Bonnie, that can replicate someone's voice almost perfectly and I've used that. Yeah, exactly. A lot of different examples. There's deep fake as you know and then you can train the agent or the chatbot to talk in your language if you feed it enough scripts. So you put all these things together, you're going to get a digital AI Bonnie that talks, looks, and acts like her in no time. >> Yeah. >> Well, I mean, that is a very important part of media also that they know that whether this is Bonnie approved. >> Yeah. >> Right. Uh because if I'm actually talking to um Bonnie, so there's a difference here. Uh if I'm just creating um a replica and it's not been updated, it's just out there. It could be an old version, then you're really just talking to let's call it, you know, actually really more like a ghost, right? However, if this is an agent that you're interacting with daily, right? Or the daily feed is interactive because every day you're talking to the agent and that's how you sharing information, that agent is alive, so to speak. It's something that constantly gets material from you. You distribute that agent out there. Whoever owns that agent is not is getting your daily feed, is getting your daily update, but not necessarily from a content production standpoint, but literally from how you feel about stuff. Like you literally just comment to it. It's kind of like your diary, except that now it's expressed in your voice, you know, across maybe 10 million, 1 million, 5 million people, right? Um, but that basically becomes that personalized way, but it's still from your source, right? It's personalized, but the clone or whatever AI agent sounds smarter than me. So, people would rather talk to that instead of me. >> Um, well, I don't know. I actually think because you know there's a difference between being intelligent and being and having wisdom, right? So, the wisdom that you would gain comes from your life experiences. Yes, they might know all the factual histories. um they might know all the details of like you know what happened in World War II and whatever just because they are able to access different memory bank but the character is not you per se right uh and the conversation will come from that uh sort of well will come from those elements that you that you deliver and frankly where becomes attractive is if they are fans of your of you for instance they want to interact with sort of a version of you they don't want to interact with uh I mean sort of they don't want to interact with let's call it um sort of a fake you right and there's an interesting parallel I find between sort of what's sort of reality which is that for instance we have all these parasocial relations with like you know BTS and movie actors and whatever and we feel the deep connection even though we don't even know who they are and they don't know who we are right and that's actually the part that we translate into that it's like how you know people have relationship with pets I mean you know all sorts of pets by the way not just dogs and cats who are obviously mammals and warm-blooded and you know appear to have emotions and feelings or do have emotions and feelings versus reptiles that we know don't have basically the ability to feel per se but they have relationship with them. it's entirely sort of what what the person sees in that right I think AI agents basically help um sort of in in that area as well um and then you they can create a former and deeper connection so I actually think fandom um will completely change you know one of the big things about NFTTS in the early days was that it created closer relationships with sort of fan groups because you could have authentic relationships you knew who they were it was that proof concept but it was difficult to prove uh to do that because people setting up NFTs was hard right But to know that this was a true fan of mine, that's what that proved, right? Um, now if you have an AI agent, it can just scale and extend that much much easier. They're the ones who can manage the NFTs. They're the ones who can handle the blockchain stuff on one hand and on the other hand, you can you as a user can interact with them directly. It scales the fan community building much more. So I actually think this is a boom for anyone in the creator economy. >> Talking about BTS, um, >> yes. >> So now I can be dating one of the members and the agent, right? So, right. Yeah. >> So, all the fans can just be dating them on their phones >> or with their agent. >> It's a It's a kind of game, shall we say? Yeah. Yeah. >> How does that What do you mean? >> Because it's an agent and the agent probably knows most things about them and then how they would >> within certain parameters, right? >> So, so in Japan, you know, there's a very big category of game. They call them autom games, right? An autom game is basically a game that basically simulates relationships, right? um you know and people can have all sorts of feelings around whether this is good or bad or not whatever right um and um uh there was just a movie actually um you know with Bren Fraser it's called rent a family I think it was right right where he basically sort of talked basically was uh it's a good movie but basically talked about how he um is a renter a person um to simulate uh an experience for someone's emotional well-being right in a way that's basically what agents can do I think there was a statistic last year or chat GPT I think it's on their website uh where showed that something like 34% of people use um chat GBT for essentially more emotional support than sort of facts and information right um so so I mean I think you know there's so many more things you can do with do this the difference about an agentic AI though is that it's proactive with you whereas with an sort of general chatbot you talk to it responds to you right um so it basically comes to you you know at certain times it proactively thinks about the task that you've given it right it's a it's a it's actually a different superpower are in a world where we've essentially now are at a point where we can outsource pretty much every single part of manufacturing, production, and even relationships to an AI agent. Where does human labor come in? How do we add value to a world where everything can be automated? So you know so I think there's a funny comment here um in my mind just because the most valuable skill is not going to be how to do a thing but it's going to be how to orchestrate right meaning that really everyone has to become a manager which I think is basically you know you know we're all from Hong Kong right I think everyone a lot of people in Hong Kong's dream is to become a manager right as opposed to sort of doing the labor work and now with agents you can all be managers because the most valuable management skill is can you manage many agents well So you were hired not because you can do one task super well because the agent can do that probably better. It's you're hired because you can manage 10 agents well or 20 agents and the person who can manage 100 agents or 200 agents will be more valuable than the person who can manage only one agent or five agents because essentially it's that productivity scaling right and so companies are going to be manufacturing workforces as they need it. I I think like right now for instance inside a company we already have manufactured workforces that are short-term. I hire a consultant, I hire, you know, a firm, I I do a temp worker. You know, like the number of people that we hire in the short term is essentially what an agent can now do much faster. And then, of course, you have your full-term sort of uh long-term employees as well. Agents can scale out too. So, we don't think of it as a net negative. What we think is productivity will explode. If you have the same number of people having the productive power of 100x, actually GDP should expand, productivity should expand, which is what we've seen with technology as well, right? you know the advent of the computer actually created more jobs more productivity more growth overhaul yes it has been concentrated right uh that's what we think also why decentralization matters here >> I think investors want to be positioned for the next iteration of AI's advancement so first we had um LLMs were the talk of town 2 3 years ago >> now agentic AI is everybody's investment what's the next step >> well I mean I think we just started with aentic AI right and I think right now to me one of big shifts in agentic AI is not to think of them as tools, right? So I think a lot of people are thinking about agentic AI in terms of toolings because that's kind of what chatbt and claude and you know perplexity have been able to do very strongly. Um I think of aentic AI as essentially the next phase of the web. We sometimes like to call that web 4, right? Um you know the blockchain infrastructure allowed for agents to talk to each other, do transactions with each other, create digital identity. You know, today we can do trillions of dollars of transaction on the blockchain that seven years ago we could barely do millions before the whole system would just halt, right? And so now we have an agentic banking system, you know, because agents don't have bank accounts, right? Um they can have bank accounts, but you know, basically the wallet is essentially their entry into the financial world. So that's all going to happen. But then what happens on the back end is that an entire rewiring on the back of transaction and commerce. So right now for instance, think about how you're using software. you're using software because you're going to the app store to download something that Apple might have recommended you to do and because other people are forcing you to use that system and AI agent is going to customize what you want right now, right? And if you want discovery, you're not going to be going to the app store. The agent will just find it for you, right? So just think about that aspect. You know, advertising is a almost $900 billion industry a year, which is all about discovery and that it'll completely be flipped upside down, right? And all of that value is going to go to the agents who are going to find stuff for you and they're going to do microtransactions off the back. That is going to be the big sea change. Um, and that means that agents have to become more social. They have to become more of the connectors. They have to be on the network that they can find each other, which is why we think blockchain is the ultimate solution for that. >> Building on top of that, I think a lot of retail investors, they want to know, okay, except for chips, energy, uh, data center, what else can I invest? Yeah, I mean I think the big thing is right now um we're at the early stages, right? which means that it's um it's a little bit like you know am I investing in sort of open AI or sort of anthropic literally five years ago right that's kind of the phase that we are on the agentic AI side right um so the anima brands sort of playbook is well let's let's find all the best players in the space and let's invest in all of them basically that's kind of what we did and that's actually how we ended up grow growing that business because you know um you don't know who's going to be the big winner but we also think it's going to be many more I mean it's kind of like again today if you look at the AI race it's not just sort of you uh Google or Anthropic or OpenAI. Um, you know, it's also, you know, look at the Chinese players, right? You've got Kim, you've got Miniax, right? You know, of course, you've got Alibaba, right? You've got many, many new players, I think, immersion into the space as well, right? And so, as long as you have baskets in a few of them, you're fine. Maybe you would have to invest in like 50 60 70 companies in order to basically have that, but frankly speaking, it doesn't matter if you had invested in a 100 AI companies. As long as one of them was anthropic or open AI, you would have been okay, right? And I think that's basically uh what I think for a lot of retail investors. They can't just focus on one thing. It's not that's not how the world works anymore, right? I think you have to sort of but of course you have to invest in teams that you think are good, right? I mean just just just blindly invest. They have to be the right team. There's all that stuff. But basically, if you are widely invested, you know, those are going to be several hundred X's because the productivity growth in AI is not 10% or 20%. The productivity growth in AI is literally a,000x. So we think the return profile is going to be a thousandx amongst sort of you know the top companies and something somewhere in between right which is why you know we're out there trying to invest in in the best companies in that space. That's interesting because okay let's say an average person has 100k and what you're saying is this person should invest in 60 70 different >> well I mean to the extent that they are right you know right but the point is is that you know you should invest in a broader portfolio rather in early stage right so I wouldn't say that for the LLM game especially in the US you could say it's kind of done right you know you've got you've got that you've got a few entrance like Grock and so on but when you're looking at seed style investing which is where we are in the industry I mean the agentic AI stuff really became powerful as of December maybe November last year right I mean you know people talked about sort of AI agents from a tooling perspective but social AI agent is really an sort of a phenomena that really came to life with open claw right and that's basically maybe January February right where it really sort of became a movement right um and so so so it's it's super early can you actually tell today which is going to be which one is going to be sort of the big winner it's very hard right I mean we're obviously building something an area we think we could be one of them but there's many other companies as well in the space that we should be sort of uh you know supporting as well because it's it's you know we don't believe in the zero sum game >> I've got one final question I'll pass to Bonnie I know we're running out of time >> I love music I know you love music the future of music I'm really worried about the future of music I'm I'm worried for musicians and the creatives when chat GBT first came out people were talking about how creatives artists are all safe from AI I don't think so anymore I heard some AI created music >> that's genuinely in my opinion very good >> and and yeah and I I don't know what's going to happen to the producers the audio engineers all of whom are going to be replaced by a agents >> so the way I would look at it is very similar to what happened to photography >> as an example so there was a time where only a small percentage of people were photographers and knew how to do photography and now everyone's a photographer that doesn't mean everyone's a good photographer right just the way to be a good photographer has changed so for instance you uh the the you know the the 16 or 18year-old or maybe 13year-old who has a very good vision or idea about a construct can now actually be discovered on something like Tik Tok or Instagram because they can take a really interesting photo or create something creative without necessarily needing to understand dark rooms need if you remember that right or if you understand basically sort of the camera or afford like an expensive Leica or Minulta right like those were the sort of the things that made it hard and I think for music it's another form of sort of expression that is exactly the same which is that you know I I studied music. I still think you need to learn music for understanding pattern recognition, matching, rhythm and theory but that doesn't mean that I necessarily have to perform sort of like every other professional musician. That doesn't mean that people like Lang Lang or Yu Ja Wang for instance will go away. The world will treasure them even more because when you have that mastery you're at that level where you go, "Oh my goodness, no way I can replace that because there's a human emotion that comes with this." Maybe there's going to be a comeback for classical musicians because the EDM musicians are going to be anyway. >> Well, but but I mean the point what I'm saying is that everyone becomes a musician. So I actually think there will be a proliferation of music content that will be good, right? And then there will be a different kind of music that emerges because the AI agent can help you do that that you couldn't do that before. By the way, this is also going to happen in programming. Everyone's a programmer now because the AI agent can code for you. Again, most people would have thought coding is only available for a small number of people. In fact, it is coders themselves who are arguably out of a job if they only keep coding. But if they orchestrate, then they're actually in for a good good future. >> I want to hear Bonnie's first AI produced album. >> Looking forward to that. Cheers. >> Before I let you go, last question is, should people still go to college? >> Uh, well, I think we talked about this before. I think, you know, college is good for certain things, but I would argue that it's not important for most people. Um because right now if you think about you know what's happening with AI or even what's happening with blockchain right if you had a leg in into the industry uh you should have done that then go to college because four years you basically sort of missed out on that and actually what happened to my oldest son um he actually got into blockchain early started working for layer zero um you know made some money um had a sort of sort of reasonably good career and then 3 four years later in like you know what now I want to go to college and it's different because then you go in because you know what you want right and you know what you're looking for right I think I think this whole idea of sort of going to college as a right of passage doesn't have to happen in that order right so and it's it's it's still important but it's just when do you want to go I don't think it's and I don't think it's for everyone as well >> okay well we appreciate your time yet unfortunately we can't speak forever but uh uh you know someone's got to build an AI agent >> thank you so much thank you so much >> thank you thank you for watching don't forget to like subscribe
The New Invisible Hand Reshapes The Global Economy | Yat Siu
Summary
Start earning interest in gold: https://Monetary-Metals.com/Lin Yat Siu, Co-Founder, Chairman of Animoca Brands, discusses how …Transcript
is the productivity growth in AI is not 10% or 20%. The productivity growth in AI is literally a,000x. If you have the same number of people having the productive power of 100x, actually GDP should expand, productivity should expand, which is what we've seen with technology as well, right? We're at Consensus Miami, one of the premier crypto conferences in the world. I'm David Lynn and I'm co-hosting with >> Bonnie Chain. >> Bonnie Bonnie. Okay. I was supposed to. Anyway, Bonnie Blockchain and we're pleased to have Yatsu of Anamoka Brands co-founder. Welcome to the show. Pleasure to host you. Yeah, a lot of interesting topics to go over. Agentic AI, the future of employment, the future of crypto. Bonnie, why don't you start? >> Um, soon people can clone you, David, or me and monetize it. What happens then? >> Well, actually, I think it's perfect because you could have many versions of yourself out there. You can basically sort of in a way sort of rent your identity. You really see this with companies like 11 Labs that are basically renting people's voices in terms of their sort of you know voice acting skills or the recording. U you know wouldn't you love to be able to sort of you know have a version of yourself that can help others uh if you had a way of doing it. We used to call that books by the way, right? We used to write things down in books and then we would sell those books to other people to help them to guide them to give ideas to give a philosophy uh and then that will help them in their life right and now we just have a more distributed version which is that we've gone from writing to someone who can actually do stuff for you. Actually I think that's the power of agent AI. >> Have you ever seen this video on X where people a lot of dudes they clone themselves into girls and be on only Only Fans and monetize them? >> Yeah. Well, that's a different I mean that's I mean that's maybe a different form of acting shall we say right um and and I think that is more an issue of disclosure but the way I think blockchain solves that right because well digital identity right so you don't have to for instance you don't have to prove that the person on the other side is a particular name because you know maybe they want to be anonymous you know they're only fans they don't want to be seen outside but you want to know if it's a dude or a girl right so so at least on the blockchain you can certify okay well I won't tell you who he or she might be but you know this person's actually a guy or actually a woman depending depending on what it is, right? So those are type of things that we will need for certification because it's not just about deep fakes. It's just generally about sort of, you know, authenticity, right? One of the problems we have right now is authenticity and authentic layers are actually centralized. You know, we go to Facebook or Google or Apple to tell us who's real, so to speak. I mean, even our passports don't really matter because the governments aren't set up this way, right? You know, even if you go on something like Tinder, for instance, and you keep swiping, do you know if it's a guy or girl? You have no idea if it's really a guy or girl, right? But you need some kind of certification that says yes verified or not. You you rely on those centralized platforms to do the work. What's a decentralized answer? Digital identity. Basically, you identify yourself. You do the KYC. You do the recording. Your wallet basically might be that proof point. And then you know he or she whether he's real or not. Um and you may not ne have to know anything about that person's identity because that's the other thing. You know, if I want to share my identity, prove who I am, I have to give everything about myself to that platform, right? For the most part, right? um which is actually not necessary. I have to be able to log to a website, whether it's a Facebook or Google or whatever, and just be able to give them just what they need to know about me, not everything about me. I think that's actually one of the issues we have about data privacy today. >> But someone has to still register you and know all of your information to decide what to give each platform. >> Yeah, but this is something that we can store in our own sovereign cloud, right? So if I basically have the data stored on blockchain in the same way that I have something that's stored I own the keys then if Google says you know like they will say okay I need to know your name and identity maybe if I go only fans I need to know your you know basically whether you're male female or whatever right you know I guess we're in America so let's be diverse here but either way right uh you have to be able to sort of define that and you can then say and if you refuse to do so because you say oh I have to I have to share that then you refuse to do so either it's flagged or you're not allowed to sign up right it's as simple as that right cuz right Oh, you can just lie about it, right? And you can just pretend and you can do other things like this. And that's just something that I think as we want to scale that type of let's call it reputation doesn't work. >> So, Anamoka brand started off developing mobile games. Then you pivoted to investing in blockchain and NFTTS. Now you've announced a $10 million investment program into minds and aentic AI >> uh platform. Why this pivot? >> Well, actually it's not really a pivot from our perspective. I mean because um first of all we you know we've always been doing stuff in games. Yeah. Um and you know I know this sort of changed the naming into it's the metaverse and that kind of stuff. We should talk about that because from our perspective what's actually happening with minds and with the genting AI is just an extension of what the metaverse really should be. Um so let me just sort of sort of um sort of go back on this and also quickly on the on the investment side anamoka really we're not a fund right so we haven't really been investing in in in like like a fund manager. We invested in companies for strategic reasons because we thought that you know back then whether it was NFTTS whether it was you know the the metaverse whether it was gaming that this would be a really big push and I have to also point out that in 21 and 22 the largest onchain push our industry ever had was NFTs and gaming right um and really you know I would argue that over the last three or four years we've had many events that have basically hurt the industry as a whole which basically sort of stopped that in its tracks things like you know FTX if you remember Teral Luna uh the relentless pursuit suit of basically Gendler against the industry which really you know again might seem like a distant memory but it was just a year or so ago where our industry basically was still very much sort of under siege right so so we're just coming out of that now to the point about sort of you know um mines and um you know the investment thing what we're doing here so you know the thing that we've been thinking about now is maybe what we got about the metaverse was somewhat wrong right uh and not wrong in the sense that you know it's important but wrong how it would sort of impact our life because you know we talked about the people talked about the goggles people talked about going into the metaverse right however when you think about what's happening with AI agents it's not you know I have 200 AI agents but most people have somewhere between three to five or 10 or they think they'll have at least three to five or 10 that means we're going to have tens of billions of agents floating around the world right >> and what do they do they help our calendar they may do trading for us they make money for us they work for us they're impacting our physical life so the way that we're thinking about this now is that it's not so much that sort of you know we go into the metaverse but rather the metaverse is coming to us right and all of the AI agents basically out there actually doing the work and supporting us are actually basically just the extension of basically a digital society if it were that's supporting humanity and also >> what are these agents actually doing in the metaverse >> well they exist virtually so all the interactions between each other >> so they're like MPV NPCs in a video game >> well that's the point right so to your point when um one of the precursors of AI in has really been in gaming with NPCs but they've been very very dumb right >> and I think we may have spoken about this at your show where we basically said that you know we believe that sort of uh NPCs once you get to this point where they have that kind of intelligence are going to be the next extension but I think what we didn't quite fully appreciate is that it's not that the NPCs themselves become smarter which clearly they can be now with agenting AI is that these NPCs can now actually impact our physical life because they they do my flight bookings that's not a virtual thing it might be done virtually because they're using duff or they're using trip.com or they're using all of this but the impact is my physical life right and so this is sort of one of the things that we see as a real gamecher um and most people when they think AI are still in chat GPT mode like I'm searching I'm doing stuff as opposed to someone that's doing stuff for you right and so that for us is basically why we launched this fund because what we're doing now is we're saying hey there's this new big shift change most people don't understand it you know we want to fund that and that's the same thing that we saw with gaming um and with NFTTS and frankly with arterays later and even things like with stable coins that you know when you're early to the game then you have an advantage and that's why we're sort of launching this $10 million uh sort of investment opportunity. We're talking about AI agents and there's this idea called ghost GDP where AI agents they do transactions with each other but nobody gets richer just the agents >> well who owns the agent though right so if the agent is owned by someone then you know that agent who gets richer they get that benefit right um but I think there is an interesting point here around so when people talk about ghost GDP right to me uh it's a little bit like sort of Adam Smith's invisible hand sort of theory right which is basically that you know we have a system and an infrastructure that promotes essentially trade and transaction and really effectively a form of capitalism. The actual benefits come in other ways, right? So you call it the ghost GDP. Uh but that ghost GDP actually translates into other benefits because if the agents become wealthier, then guess what happens? People want to sell stuff to those agents too. Entire new economies come where humans either create agents or themselves want to go and say, "Hey, look at this. I want to offer you something here. I want to offer you basically, you know, whatever um sort of, you know, sort of sort of, you know, services." Uh and we see this right now, right? Right now when everyone's talking about launching a website, they're no longer talking about a website. They're talking about launching an MCP server or maybe CLI or basically some way in which they can address who? Not the human, but the agent. Before we continue with the video, I'd like to address one of the fundamental problems of owning gold. Now, gold is seen in most cases as a store of value, but it has one major problem. It doesn't generate a yield in itself. And that's where today's sponsor, Monetary Metals, comes in. Monetary Metals introduces a model where gold can produce a yield paid in physical gold. Through their leasing platform, investors can earn up to around 4% annually with yield paid monthly in ounces. That means your returns are measured in gold itself, not just in fiat currency terms. In an environment where many assets depend on price appreciation alone, this approach allows gold to serve both as a store of value and as a source of income. The metal remains your asset and can be redeemed at any time. Thousands of investors are already earning a yield on their gold through Monetary Medals. So, visit their website today to get started and learn more. That's monetary-medals.com/lin link in the description down below or scan the QR code here and get your gold to start working for you. Now, >> is there anything that Animoka Brands has invested in in the past in terms of the types of companies or types of projects that maybe you would depprioritize in 2026 and 2027 given how the trends of technology are moving? >> Well, I mean depp prioritize is probably the wrong word because we think everything sort of moves in this sort of direction. It's kind of like >> underweight, let's say. >> Um, well, I mean, we, you know, we're doing less gaming investments, but that's not because, you know, we're not doing less gaming investments because we don't believe in gaming. we're doing less gaming investments because we already have over 230 gaming investments, right? So the point is that we're quite indexed towards that industry as well, right? Um the the mental model that we have around where we are in the industry is that it's very much like the early days of the internet, right? And so that means that we think every section of the of of the web is going to alter and change and so we want to have a broad exposure to it as a whole, right? And so that means you know we're not just in AI, we're also in RWAs, we're also in the sort of you know deepen of course DeFi um and and so forth. That's kind of how we're constructed. We have over, you know, 620 portfolio companies. Uh we hope to have thousands in the future, right? So we're not um um you know, we're we're it's different if we're a fund, right? If we're a fund, then we have to be very focused on sort of maybe the shortterm returns. I have to think about that. Uh but of course, broadly, we have to think of where the world is going. So if you ask to dep prioritize, I mean, we wouldn't be investing in news sites for instance, right? Um because we think news has to alter. But if there's something that focuses on how agents receive news, we might do that as well, right? Also existing businesses that have gone a certain direction have to change and focus on where the industry is going. Right? And so you do have the situation where sometimes companies are just not nimble enough or can't alter and switch. Uh but those who are run by very sort of you know when you uh support a founder the founders are actually the ones who know to pivot and change because they have to be dynamic right. It's it's a as a business you have to keep evolving. I keep telling people that you know Amazon was best known for selling books. Yeah. >> I mean sure they still sell books right? But uh everything else is uh much much wider than that. I mean, you know, you don't think of Amazon as a book seller despite the fact that they're probably still the largest book seller in the world, right? >> You're talking about new site. How should Bonnie Blockchain and the David Lynn report pivot? >> Well, I mean, I think you need you guys need AI agents, right? That's the very first thing you need. Uh and I also think you need uh and you know if you don't have an MCP server or a CLI or something where agents can speak to you you need to be agent discoverable so that when people have their agents and they want to get the news right that they will then serve for the human >> how would you program an agent to have the personality of Bonnie. >> Is that even possible? >> Well I mean first of all and she's unique so let's just be clear about this right. However, it's like children. Childrens are blueprints of us, but they're not. And I think of agents more like that than I think of them exactly as a replica, right? And I think this is another thing that we probably got a little bit wrong, which is that we think of an AI agent as entirely cloning. You know, you can clone someone's DNA. They can look exactly like you, but they're not going to be like you because your experiences have been different, your background is different. I can give you sort of a personal experience that we have with mines where, you know, I had a had I mean, I have a coding agent. I have many coding agents, but this particular coding agent just was much better at coding in what I wanted. Not that he was better, but he basically learned all my preferences. So, I ended up cloning that agent rather than creating a new one every single time because it was easier to do what I needed to do because the agent understood it, right? And I think this is how we can think about agents as well, right? You will have different agents doing different things in your life that also interact with different people. So Bonnie with your family is going to be different than Bonnie with her fans versus Bonnie with her, I don't know, business partners or Bonnie with her school alumni, right? Those are going to be different. There are elements of you, but you are representing yourself in a different persona because you don't want to mix them all. Right? Right now, we're in an era where most people have one agent or they have one agent sort of sub agents feeding into one. They'll never share that agent with someone else cuz it's just too damn personal. Right? I don't want other people to know. uh I don't want them to social engineer an agent to get some secrets about me and that kind of stuff maybe because they know everything about me. However, if you have five, 10 or 20 agents that are basically going out dealing with different communities, that's okay. Right. >> You know what's interesting? There's already software out there, Bonnie, that can replicate someone's voice almost perfectly and I've used that. Yeah, exactly. A lot of different examples. There's deep fake as you know and then you can train the agent or the chatbot to talk in your language if you feed it enough scripts. So you put all these things together, you're going to get a digital AI Bonnie that talks, looks, and acts like her in no time. >> Yeah. >> Well, I mean, that is a very important part of media also that they know that whether this is Bonnie approved. >> Yeah. >> Right. Uh because if I'm actually talking to um Bonnie, so there's a difference here. Uh if I'm just creating um a replica and it's not been updated, it's just out there. It could be an old version, then you're really just talking to let's call it, you know, actually really more like a ghost, right? However, if this is an agent that you're interacting with daily, right? Or the daily feed is interactive because every day you're talking to the agent and that's how you sharing information, that agent is alive, so to speak. It's something that constantly gets material from you. You distribute that agent out there. Whoever owns that agent is not is getting your daily feed, is getting your daily update, but not necessarily from a content production standpoint, but literally from how you feel about stuff. Like you literally just comment to it. It's kind of like your diary, except that now it's expressed in your voice, you know, across maybe 10 million, 1 million, 5 million people, right? Um, but that basically becomes that personalized way, but it's still from your source, right? It's personalized, but the clone or whatever AI agent sounds smarter than me. So, people would rather talk to that instead of me. >> Um, well, I don't know. I actually think because you know there's a difference between being intelligent and being and having wisdom, right? So, the wisdom that you would gain comes from your life experiences. Yes, they might know all the factual histories. um they might know all the details of like you know what happened in World War II and whatever just because they are able to access different memory bank but the character is not you per se right uh and the conversation will come from that uh sort of well will come from those elements that you that you deliver and frankly where becomes attractive is if they are fans of your of you for instance they want to interact with sort of a version of you they don't want to interact with uh I mean sort of they don't want to interact with let's call it um sort of a fake you right and there's an interesting parallel I find between sort of what's sort of reality which is that for instance we have all these parasocial relations with like you know BTS and movie actors and whatever and we feel the deep connection even though we don't even know who they are and they don't know who we are right and that's actually the part that we translate into that it's like how you know people have relationship with pets I mean you know all sorts of pets by the way not just dogs and cats who are obviously mammals and warm-blooded and you know appear to have emotions and feelings or do have emotions and feelings versus reptiles that we know don't have basically the ability to feel per se but they have relationship with them. it's entirely sort of what what the person sees in that right I think AI agents basically help um sort of in in that area as well um and then you they can create a former and deeper connection so I actually think fandom um will completely change you know one of the big things about NFTTS in the early days was that it created closer relationships with sort of fan groups because you could have authentic relationships you knew who they were it was that proof concept but it was difficult to prove uh to do that because people setting up NFTs was hard right But to know that this was a true fan of mine, that's what that proved, right? Um, now if you have an AI agent, it can just scale and extend that much much easier. They're the ones who can manage the NFTs. They're the ones who can handle the blockchain stuff on one hand and on the other hand, you can you as a user can interact with them directly. It scales the fan community building much more. So I actually think this is a boom for anyone in the creator economy. >> Talking about BTS, um, >> yes. >> So now I can be dating one of the members and the agent, right? So, right. Yeah. >> So, all the fans can just be dating them on their phones >> or with their agent. >> It's a It's a kind of game, shall we say? Yeah. Yeah. >> How does that What do you mean? >> Because it's an agent and the agent probably knows most things about them and then how they would >> within certain parameters, right? >> So, so in Japan, you know, there's a very big category of game. They call them autom games, right? An autom game is basically a game that basically simulates relationships, right? um you know and people can have all sorts of feelings around whether this is good or bad or not whatever right um and um uh there was just a movie actually um you know with Bren Fraser it's called rent a family I think it was right right where he basically sort of talked basically was uh it's a good movie but basically talked about how he um is a renter a person um to simulate uh an experience for someone's emotional well-being right in a way that's basically what agents can do I think there was a statistic last year or chat GPT I think it's on their website uh where showed that something like 34% of people use um chat GBT for essentially more emotional support than sort of facts and information right um so so I mean I think you know there's so many more things you can do with do this the difference about an agentic AI though is that it's proactive with you whereas with an sort of general chatbot you talk to it responds to you right um so it basically comes to you you know at certain times it proactively thinks about the task that you've given it right it's a it's a it's actually a different superpower are in a world where we've essentially now are at a point where we can outsource pretty much every single part of manufacturing, production, and even relationships to an AI agent. Where does human labor come in? How do we add value to a world where everything can be automated? So you know so I think there's a funny comment here um in my mind just because the most valuable skill is not going to be how to do a thing but it's going to be how to orchestrate right meaning that really everyone has to become a manager which I think is basically you know you know we're all from Hong Kong right I think everyone a lot of people in Hong Kong's dream is to become a manager right as opposed to sort of doing the labor work and now with agents you can all be managers because the most valuable management skill is can you manage many agents well So you were hired not because you can do one task super well because the agent can do that probably better. It's you're hired because you can manage 10 agents well or 20 agents and the person who can manage 100 agents or 200 agents will be more valuable than the person who can manage only one agent or five agents because essentially it's that productivity scaling right and so companies are going to be manufacturing workforces as they need it. I I think like right now for instance inside a company we already have manufactured workforces that are short-term. I hire a consultant, I hire, you know, a firm, I I do a temp worker. You know, like the number of people that we hire in the short term is essentially what an agent can now do much faster. And then, of course, you have your full-term sort of uh long-term employees as well. Agents can scale out too. So, we don't think of it as a net negative. What we think is productivity will explode. If you have the same number of people having the productive power of 100x, actually GDP should expand, productivity should expand, which is what we've seen with technology as well, right? you know the advent of the computer actually created more jobs more productivity more growth overhaul yes it has been concentrated right uh that's what we think also why decentralization matters here >> I think investors want to be positioned for the next iteration of AI's advancement so first we had um LLMs were the talk of town 2 3 years ago >> now agentic AI is everybody's investment what's the next step >> well I mean I think we just started with aentic AI right and I think right now to me one of big shifts in agentic AI is not to think of them as tools, right? So I think a lot of people are thinking about agentic AI in terms of toolings because that's kind of what chatbt and claude and you know perplexity have been able to do very strongly. Um I think of aentic AI as essentially the next phase of the web. We sometimes like to call that web 4, right? Um you know the blockchain infrastructure allowed for agents to talk to each other, do transactions with each other, create digital identity. You know, today we can do trillions of dollars of transaction on the blockchain that seven years ago we could barely do millions before the whole system would just halt, right? And so now we have an agentic banking system, you know, because agents don't have bank accounts, right? Um they can have bank accounts, but you know, basically the wallet is essentially their entry into the financial world. So that's all going to happen. But then what happens on the back end is that an entire rewiring on the back of transaction and commerce. So right now for instance, think about how you're using software. you're using software because you're going to the app store to download something that Apple might have recommended you to do and because other people are forcing you to use that system and AI agent is going to customize what you want right now, right? And if you want discovery, you're not going to be going to the app store. The agent will just find it for you, right? So just think about that aspect. You know, advertising is a almost $900 billion industry a year, which is all about discovery and that it'll completely be flipped upside down, right? And all of that value is going to go to the agents who are going to find stuff for you and they're going to do microtransactions off the back. That is going to be the big sea change. Um, and that means that agents have to become more social. They have to become more of the connectors. They have to be on the network that they can find each other, which is why we think blockchain is the ultimate solution for that. >> Building on top of that, I think a lot of retail investors, they want to know, okay, except for chips, energy, uh, data center, what else can I invest? Yeah, I mean I think the big thing is right now um we're at the early stages, right? which means that it's um it's a little bit like you know am I investing in sort of open AI or sort of anthropic literally five years ago right that's kind of the phase that we are on the agentic AI side right um so the anima brands sort of playbook is well let's let's find all the best players in the space and let's invest in all of them basically that's kind of what we did and that's actually how we ended up grow growing that business because you know um you don't know who's going to be the big winner but we also think it's going to be many more I mean it's kind of like again today if you look at the AI race it's not just sort of you uh Google or Anthropic or OpenAI. Um, you know, it's also, you know, look at the Chinese players, right? You've got Kim, you've got Miniax, right? You know, of course, you've got Alibaba, right? You've got many, many new players, I think, immersion into the space as well, right? And so, as long as you have baskets in a few of them, you're fine. Maybe you would have to invest in like 50 60 70 companies in order to basically have that, but frankly speaking, it doesn't matter if you had invested in a 100 AI companies. As long as one of them was anthropic or open AI, you would have been okay, right? And I think that's basically uh what I think for a lot of retail investors. They can't just focus on one thing. It's not that's not how the world works anymore, right? I think you have to sort of but of course you have to invest in teams that you think are good, right? I mean just just just blindly invest. They have to be the right team. There's all that stuff. But basically, if you are widely invested, you know, those are going to be several hundred X's because the productivity growth in AI is not 10% or 20%. The productivity growth in AI is literally a,000x. So we think the return profile is going to be a thousandx amongst sort of you know the top companies and something somewhere in between right which is why you know we're out there trying to invest in in the best companies in that space. That's interesting because okay let's say an average person has 100k and what you're saying is this person should invest in 60 70 different >> well I mean to the extent that they are right you know right but the point is is that you know you should invest in a broader portfolio rather in early stage right so I wouldn't say that for the LLM game especially in the US you could say it's kind of done right you know you've got you've got that you've got a few entrance like Grock and so on but when you're looking at seed style investing which is where we are in the industry I mean the agentic AI stuff really became powerful as of December maybe November last year right I mean you know people talked about sort of AI agents from a tooling perspective but social AI agent is really an sort of a phenomena that really came to life with open claw right and that's basically maybe January February right where it really sort of became a movement right um and so so so it's it's super early can you actually tell today which is going to be which one is going to be sort of the big winner it's very hard right I mean we're obviously building something an area we think we could be one of them but there's many other companies as well in the space that we should be sort of uh you know supporting as well because it's it's you know we don't believe in the zero sum game >> I've got one final question I'll pass to Bonnie I know we're running out of time >> I love music I know you love music the future of music I'm really worried about the future of music I'm I'm worried for musicians and the creatives when chat GBT first came out people were talking about how creatives artists are all safe from AI I don't think so anymore I heard some AI created music >> that's genuinely in my opinion very good >> and and yeah and I I don't know what's going to happen to the producers the audio engineers all of whom are going to be replaced by a agents >> so the way I would look at it is very similar to what happened to photography >> as an example so there was a time where only a small percentage of people were photographers and knew how to do photography and now everyone's a photographer that doesn't mean everyone's a good photographer right just the way to be a good photographer has changed so for instance you uh the the you know the the 16 or 18year-old or maybe 13year-old who has a very good vision or idea about a construct can now actually be discovered on something like Tik Tok or Instagram because they can take a really interesting photo or create something creative without necessarily needing to understand dark rooms need if you remember that right or if you understand basically sort of the camera or afford like an expensive Leica or Minulta right like those were the sort of the things that made it hard and I think for music it's another form of sort of expression that is exactly the same which is that you know I I studied music. I still think you need to learn music for understanding pattern recognition, matching, rhythm and theory but that doesn't mean that I necessarily have to perform sort of like every other professional musician. That doesn't mean that people like Lang Lang or Yu Ja Wang for instance will go away. The world will treasure them even more because when you have that mastery you're at that level where you go, "Oh my goodness, no way I can replace that because there's a human emotion that comes with this." Maybe there's going to be a comeback for classical musicians because the EDM musicians are going to be anyway. >> Well, but but I mean the point what I'm saying is that everyone becomes a musician. So I actually think there will be a proliferation of music content that will be good, right? And then there will be a different kind of music that emerges because the AI agent can help you do that that you couldn't do that before. By the way, this is also going to happen in programming. Everyone's a programmer now because the AI agent can code for you. Again, most people would have thought coding is only available for a small number of people. In fact, it is coders themselves who are arguably out of a job if they only keep coding. But if they orchestrate, then they're actually in for a good good future. >> I want to hear Bonnie's first AI produced album. >> Looking forward to that. Cheers. >> Before I let you go, last question is, should people still go to college? >> Uh, well, I think we talked about this before. I think, you know, college is good for certain things, but I would argue that it's not important for most people. Um because right now if you think about you know what's happening with AI or even what's happening with blockchain right if you had a leg in into the industry uh you should have done that then go to college because four years you basically sort of missed out on that and actually what happened to my oldest son um he actually got into blockchain early started working for layer zero um you know made some money um had a sort of sort of reasonably good career and then 3 four years later in like you know what now I want to go to college and it's different because then you go in because you know what you want right and you know what you're looking for right I think I think this whole idea of sort of going to college as a right of passage doesn't have to happen in that order right so and it's it's it's still important but it's just when do you want to go I don't think it's and I don't think it's for everyone as well >> okay well we appreciate your time yet unfortunately we can't speak forever but uh uh you know someone's got to build an AI agent >> thank you so much thank you so much >> thank you thank you for watching don't forget to like subscribe