Crypto Was Never Built For Humans | Yat Siu

Crypto is like Trump's tenth child, right? As in, you know, he loves us, but we're just not the most important thing, right? He has to deal with China. He has to deal with now Iran. We're not the top priority. So, we have to continue to fend for ourselves. Crypto and AI are probably the most important relation ever because they need each other. I think the issue isn't that we haven't built great technology. The issue is it wasn't built for humans. And that's actually where we think AI agents come in. Code is their natural language. You know, we talk about code as law. Well, guess what? For agents, it is law. In a future where agents are going to be transacting mostly with each other and with other platforms, it's all going to happen on chain. What really has value is...

Well, I mean, first of all, thank you for having me. I would say it's a little exaggerated. I think a little bit of dramatic effect here. Of course, you know, we have a conference. There are people here. The people who are here, the energy is good. I just, you know, was at the web festival in Hong Kong a couple weeks ago and also at sort of, you know, Japan had basically some events around there as well. And the sentiment is good. People are basically sort of excited about it. However, it's not sort of the same bullish feeling that some people had like say a year ago, a year and a half ago, particularly because people had high hopes because, you know, Trump was about to be coming into power. You know, regulation, you know, it was the end of Gensler, so everyone's like, yeah, great, great, great. But then we had a series of events. I tell people often that, you know, look, I mean, you know, sort of crypto is like Trump's tenth child, right? As in, you know, he loves us, but we're just not the most important thing, right? He has to deal with China, he has to deal with now Iran, whatever, so we're not the top priority. So we have to continue to fend for ourselves. And I think we're coming to that realization, whereas I think before a lot of people were like, 'Oh, puppy, help me.' Right? So, there's a little bit of that, right? So that's kind of one element, but I think the other thing also is that, you know, the other realization for us is that the entire industry has focused too much on things like memecoins. And we have to remember that memecoins was a reaction to a regulatory scenario that was hostile to crypto. So, memecoins had to launch into existence to say, 'Hey, this isn't real. This isn't whatever.' It certainly proved out the technology, but it basically put the wrong focus on it, right? In terms of it took away focus from builders. It took away people who were actually sort of really creating products that mattered. And in a way, I would argue also made us not focus on the thing that really matters, which is very much come to bearing now, which is that crypto and AI are probably the most important relation ever because they need each other. But we kind of, you know, we spoke about this years ago, but very few people paid attention to it because they were more like, you know, what's the next memecoin, what's the next thing to trade. And, you know, you talked about where does retail get an edge. I think retail's edge is understanding where the long-term future is and holding it and believing in it in the same way that if retail had bought Bitcoin early on, it would have been the best play rather than basically speculating on coins, which can be fun. By the way, I'm not against the fun and entertainment of meme coins. But if that's where you think you're going to make a big return, then it's basically gambling for the most part. And I think that's the part where we just have to be real that, you know, as we have seen afterwards, the disillusionment around, oh, looks like the house wins there as well, right?

I think Polymarket's a great example. I mean, frankly, even, you know, people might say, 'Oh my goodness, NFT is not working, whatever.' It's still $300 million of sales a month, right? So, it's not nothing. I mean, you know, you talk to most people, normal business people, and you say, 'Hey, this is an industry that is just a couple billion dollars a year. Are you interested?' Hell yeah, right? I think people in the crypto space sometimes are like, 'Where's my million dollar payout?' Come on, guys, right? Like most people don't think of it this way. I mean, you know, we were happy in the early days when Ethereum went from 100 to 200 and we were celebrating when it went to 500 and now we're just like, oh my goodness, like, you know, so it's all relative, I would say. There's that one thing. So, I think we need to sort of come back to our roots a little bit, understand that, you know, we've had humble beginnings. But I think the other thing is that, and I think this is the big point, is that the technology we built is incredible. I mean, stable coins, trillions of dollars of transactions. You know, seven, eight years ago when we sort of started in the space, actually nine years ago if you think about it. Wow. Crazy. When CryptoKitties first launched, the trading of digital cats melted Ethereum. Yeah. Could not handle it. Okay, we're talking millions of dollars here. And that's when people started talking about sort of, you know, how do we scale this? You know, do we do sharding? Do we do sort of, you know, L1, L2s and all that kind of stuff, right? And then Solana and all these other chains started to emerge and today we have an infrastructure that is good enough for the biggest financial institutions with BlackRock, Fidelity, Visa, Mastercard, all of that is coming in, right? And so I think the issue isn't that we haven't built great technology. The issue is it wasn't built for humans and that's actually where we think AI agents come in because basically, you know, an agent doesn't care about sort of the cryptography. In fact, they're comfortable with it. Code is their natural language. You know, we talk about code as law. Well, guess what? For agents, it is law. For humans, someone steals a bitcoin. Where's the police? Need to go to the government, right? That provenance is important for digital property, is important for agents, maybe less important for humans because we expect essentially, you know, our rights to be protected in a human legal framing, right? The other thing is of course efficiency. The banking system, an agent can have a bank account, but he has a crypto wallet. That crypto wallet is effectively his bank account. He can now transact with each other in the future where agents are going to be transacting mostly with each other and with other platforms. Basically, you need a wallet. They're not going to pay 1 or 2%. They're not going to pull Visa or Mastercard. It's all going to happen on chain. You know, on-chain credit, digital identity, all of that stuff matters. So, you know, we think that that's actually the real use case. And you know, maybe I can ask you this question and maybe the audience will think about it as well, which is how many agents do you think you'll have in the next three to five years?

Yeah. But it can also take context from other things and make it better. And you can share these skills and if you have a good trader, you can basically copy that skill for instance or import it. But I think the point is that you're going to have many agents, right? And if you have, you know, in your personal life and then you have many agents probably at work. And so you add them all up, you know, we think the world's going to have hundreds of billions of agents, right? The entire internet is going to be swarmed by agents. There's going to be more agents than humans, right? And the thing that I like to say to people, you know, just to be slightly controversial, is that, you know, we still think the metaverse is a big thing, except that maybe it wasn't about us going into the metaverse, but it's about the metaverse coming to us through agents because they're basically affecting our physical life. They do things for us, right? Everything we built on blockchain was the PC for agents to come, right? This entire agentic web, which you could call web 4, is essentially sort of we prepared for that moment. And we needed that kind of agentic intelligence that really only started to come into fruition probably late last year. Yeah. To be reality. And now that moment is here, right? And I think this is where our gaming roots is interesting too because I'd spoken much earlier around how NPCs are basically our primary interface originally, like gaming NPCs for AI agents, but you know, they were dumb and now you have that intelligence where essentially, you know, everything in the gaming world, everything in that area is going to interact with the AI agents, but on the back end, what are they going to be powered by? They're going to need a wallet, you know, it's going to be decentralized, they're the ones who basically find the most efficient way. I was just having a conversation with someone and I was like, you know, every application will change. You're not going to visit a website. You know, if you want to find a date, you're not going to go to Tinder. You're going to tell your agent, 'I'm looking for this,' right? And they're going to find you suitable ones for other people who have the same request and the agents are going to match each other because for humans, we have a friction of going to websites. You know, the attention, okay, I go to Tinder, I go to Bumble, I go to whatever the websites are out there. It's going to take like 10, 20, 30, 40 websites. That's too much. I can't do that. Agent doesn't care, right? It's just going to go out and look and look and look, right? And then it's going to give you an output from the best that's out there. So, I think we're going to have an incredible sort of hyperefficiency, hyper productivity kind of era. And to the retail guys, my suggestion is utilize it.

I actually think the opposite will happen, right? And the reason why I think the opposite will happen is because right now with the chat bots that we have like sort of OpenAI and Gemini, it's centralized so it is about basically consolidation of that top information, right? Just because it's in a centralized platform and the likelihood it's making assumption about what you're looking for in terms of why I want the best output and that output is democratized by the way, so that output is the same for everyone. However, when it talks about social agents, which is where we think AI agents have the real power, because you read agents with Hello Minds, for instance, what we can do is, you know, we can connect people over email, over Telegram, over chat, that actually enhances your human connection. It doesn't actually sort of just is just about the tooling. So most people when they think about AI, they think about the tools, they think about what it does, which is kind of how people talk about the early internet. But then actually what the internet became was completely changing how we connect, how we interact, how we relate, right? And that's what we think what agents are going to do. And so it doesn't matter that it's so-called the top 5% because I'm connecting with everyone else that's relevant to me in my location, in my area, in my areas of interest. Oh, you know, I'm looking for people who are into Pokemon cards in my neighborhood. Okay, fine, right? You're not going to go. It doesn't matter that, you know, he's a top Pokemon trader in Tokyo or in New York. I happen to live in Austin or happen to live somewhere else. I'm in that network, right? That's what you can do now. That's how you could connect as opposed to going to a Facebook group or going into Reddit and trying to figure out who's that. You could do it, but the agent does it much better. I actually think humans can connect faster and easier because now you're not dependent on a platform that tells you what to do. The agent does exactly what you wanted to do and will bridge it with someone else. So if everyone has several agents, they're going to make these connections. But right now, speaking of tooling, the reason why tooling mindset is holding it back is because you put everything into one agent. So that one agent, you won't share. I mean, do you, you know, if you have an agent that knows everything about you, are you going to give it to maybe even your wife, right? Well, you're certainly not, you know, your worker colleague, colleagues at work or maybe even friends for your alumni. You need agents for different communities as you do in real life, right? So that's the part where agents become social and they become your social agents and that to us is where it's going to change. But those social agents would still transact with each other. For instance, when friends come to Hong Kong, you know, they can ask me for restaurant tips, but then I bring my agent over email and connect them and they say, 'This is a restaurant you would like.' It's more than just an assistant. It actually creates a stronger relation that I have with that person because my agent helped them find a good restaurant that came from my data, right?

Yeah.

Sure. Yes.

So very, you know, multiple questions in that, but that is the big reason why, the big reason why tokens haven't appreciated, altcoins, which is a big part of our thesis, hasn't really appreciated the way that we had hoped it would, is because the user numbers on chain haven't grown. So let me just do it by the numbers. We now have around almost 800 million, like over 700 million people that own crypto on exchanges and we have less than 70 million people on chain, which is almost the same number as in the last three or four years ago. In fact, the largest number of on-chain growth that we've had for new users was actually during the NFT and metaverse boom. People came in and the people who, I remember, you know, when Sandbox, which is one of our projects, right, people joined blockchain because they were excited about the metaverse, they were excited about SAND, you know, the building land and virtual land. They were not from crypto, right? They were not finance guys. It was all the creative people, the artists, the musicians, they all came in, right? And then of course we had SBF and we had this and then we had Gensler, we had everything and then, you know, unfortunately Trump, you know, this year obviously wasn't sort of, some of the stuff wasn't that helpful either. As a result, basically killing off, you know, the sort of many people in that space. And so we basically not grown, right? That audience, that regular audience that's in there. And I think I do believe that if we didn't have those horrible events over the last three or four years that actually this would have been the base that holds, accelerate towards that. That's one thing. But that aside, right, that aside, I think it's about basically making it easy for people to basically go and use these tools, right? And basically add value and that to me is basically what AI agents will deliver, right? Because, but then the AI agents have to pay each other. So for instance, and in Minds, people are actually trading with each other on the back, right? And also when you build a skill, if I share that skill with you, I end up basically having a revenue share for that, the agent will just negotiate with that in real time, right? So online advertising today is a 900, almost $900 billion. It's actually $800 billion. This predicted to be $900 billion a year, right? That is a transaction, but it's a transaction of your time and attention. If that time and attention is no longer on TikTok and on Instagram because my AI agents are basically sort of making it more efficient, then I'm willing to pay for that in a different way. So that's basically where that transactional value will go. And that to your question is that's where tokens come in because I will give you a token of one form of your attention or other attention of value, but right now we're not on chain. So the value can't increase when the population doesn't increase.

That's the effect then, the network effect. So as more tokens launch, actually you have a dilution because the attention of the same number of users goes to all of them. Yeah. And so, this is the thing. The net value of the altcoin space is actually kind of the same.