E112: Virtuals Co-Founder: How To Get Rich With Crypto AI Agents In 2025 (Without Getting Lucky)

34:44And some of them might end up as a help or even as a partner, right? So I think that's the value. That's the value of like the guys who still stay for like some of these critical collections, right? You said I owned a few, a bunch of Pudgy I only own one left why? okay obviously it was trading it was back then Pudgy was like what 2 if 2-3 if and we just bought a few just a whole and I think when it went up in price just TP but then you always have your one grill that you never you never release because that's your identity that community when you say we it's not it means that you are doing that with the team?

35:31Some of us. Some of us on the team. Yeah.

35:36Let's get back to virtual.

35:41One of the things you mentioned before was it was amazing when you started to have this PMF and it went kind of like an L, right? I mean, not kind of. It went like an L. Capital letter L. You said it attracted a lot of mind share, good and bad, right? Everyone was talking about you, I knew only two to three months ago. And you attracted a lot of FUD, fear, uncertainty and doubts. You had competitors attacking you, which of course results then into investors starting to be worried. How do you personally deal with FUD? so it's a very interesting point um it depends on how the what what the fud is about right or what attack is about but

36:40i think generally what we try to do is we try to respond if it's a critical fud information we will actually respond publicly like to just tackle that um

36:54and but sometimes i think the rarity is this kind of fud can actually have a toll on founders from a anxiety perspective like you you know the rarity is you can have 99 99 people supporting you and you have one troll that comes in and just like talk shit about you and it can affect your day and it is a reality of everyone right like you can be a celebrity you can be a project project founder you just need that one person to just disrupt your day and it's the same thing so I think practically what I take to practice is that I tend to only open my Twitter or my Telegram where I have access to public once a day so like any kind of critical information or critical collaborations all happens on like Slack or like private messages and this is how you remove real collaborations from noise right and from interfacing um what that does is that it allows you to keep your mind focused for the for for a good chunk of the day so you get to get work out um and then when let's say around dinner right when productivity starts falling then you start you start checking twitter you start reading stuff so then the idea it's yeah maybe there's some cool stuff you can reply to run some thought experiments but then you'll start facing some of this fun as well and then you can address them or some of them I mean if they're small it's not worth to address you just ignore but at least from a mental toll perspective it does not obstruct you from being productive for your day so at least that's that has been the playbook that I've developed since since the start of of 2021 because I think back then was worse right like it was the peak of the bull when we launched and then when the when the token went down it it was worse because there was less fundamentals as a project back then um and yeah like you'll be busy responding fast you get a massive like anxiety stress on you and it distracts it distracts a lot back then i was even i had the mentality of like responding immediately to every single thing right and i realized sometimes it's detrimental for the long for your longevity of your of your health so then yeah that's why I started taking this approach right like okay this is one time block at the end of the day and then you use that to just reply stuff or to handle any kind of information yeah yeah everybody's human right so it's normal to have feelings and emotions even Jeff he was saying when someone comes and fuds or say something very negative I take it very personally and it affects me a lot because that's my baby right like I'm building this thing like no but one thing is and also it's sorry but it's so easy for people even for this podcast like I get out of shit right it's so easy to just criticize when you are behind your laptop and most of them they're not even public these people who are criticizing you right they're like oh you are I don't know you're lame or you are you're weird or you are freaky or whatever and like it's cool but like show yourself first then we can talk right I don't know but for me actually like after a while the the biggest worry honestly right is sometimes not not the FUD attacking like like I stopped I don't know whether it is because you went through the space for quite a while and you just become numb to a lot of these kind of emotions but for me the biggest worry is actually when it affects the team right and sometimes because now as a team grow larger right you can't handhold or or like coach every single one every single time right and then now sometimes they become the person who gets hit by this fight because you know they're interfacing with projects interfacing with people right and sometimes I think for me that's the biggest worry because like like I know for a fact okay maybe I've already become numb to this right but these guys is probably the first time and they will face the same level of anxiety so back then back then right back in 2021 there was a period of time where it was probably my first time where I actually woke up in cold sweat like when you sleep you just wake up you're sweating right like what the fuck's happening and then you start realizing such anxiety right it happened so that was like the first probably the first six months of when we started PathDao back then now that's so I think now you can you can fart me I guess to sleep at night but it's just more worrying that but it would happen regularly back then?

41:43Back then was actually, yeah, it was actually quite regular. It was on a daily basis. Food and bad night, every night, sweating. It wasn't every night. It was more like there were some critical junctures where like sometimes, sometimes it's not even just about FUD, sometimes it's about like when you're trying to launch a new, when you made a large decision and then you're uncertain of what the market will perceive or what the outcome is going to be. And then, yeah, the anxiety hits you, right? Back then, yeah. I think now, again it's like you just no capacity to feel it needs to be talked about even I'm pretty close to Meow from Jupiter one of our beloved sponsor and friends and he told me man I mean I see him right even now like when they have like big product releases like sometimes he disappears for a few days and I'm like what's going on and he's like man I was in hospital yesterday throwing up because I'm too stressed right like it happens like everybody's a human and even these guys who like have superpowers and are able to pull crazy things like they're just humans at the end of the day and it's hard.

42:48It's hard to build a business and in this industry, it's like... Yeah, no, I used to tell like some other founders I know in the Web2 space, right? It's like, their life is so easy because in Web2, you launch a token and the problem is, right, you don't have an IR team that interface with these people in an investor relationship. In like a... Web2 company, right? If you launch, if you IPO as a stock, right? You have an entire investor relations team that handles everyone, right? All you need to care about is your product and the users of the product. That's all you care, right? Same thing as a lot of founders who are building stuff and it's very different in crypto, right?

43:26You now have this bunch of like token holders that you just need to manage their feelings in a sense. Yeah. Yeah, because everybody loves you when the token goes up 100x. When the token goes down 99%, you are the asshole, right? And the reality is, I think, the reality is also in crypto, the volatility is so much higher because the speculative premium is so much higher in crypto versus like Web2. So yeah, a lot of the founders in Web2 got it easy. You talk about speculative premium.

43:58A valuation in billions of dollars must be justified by strong fundamentals in the normal world, right? Yeah. Do you feel like the market got ahead of itself in the crypto AI sector and that it was just impossible for you guys to ship at a pace that would justify $5 billion valuation in a matter of month. Right. Actually, very interesting. This one is... So when... So the... There are two ways to look at the fundamentals, right? Which is... There is a earnings... Like a protocol earning fundamentals at least from a crypto project, right? And then there's... Yeah, and actually anything else is speculative, right?

44:36so the funny thing is when when we had our like our revenue curves right are actually quite volatile but like I think since inception from an ecosystem revenue perspective we made probably like 50 to 80 because it depends on the volatility of the we used to collect taxes as virtual tokens now we've switched it to Coinbase Bitcoin so it's more stable right now but but back then it was like 30 to 50 right and and then I was I was then doing and this was like over the course of like two months the first two months and I was then doing like a standard like price to earnings ratio I was trying to bring my my management consulting my background what is the true value of this business if it's actually making that kind of ecosystem revenue and you analyze it right yeah but then you no but you realize actually it's not bad you know because it's really interesting right so if your protocol makes let's say 60 in like 2 months let's say 30 million a month right that is like a run rate of times 12 that's like 360 million in run rate so then you start so this is like the run rate revenue right and revenue in this case can be equitable to earnings because the cost is not crazy the cost is probably like 5 % right so let's say more or less then then then you think of this ecosystem revenue to its valuation point, you can apply, if you apply a very standard like tech stock, multiplier is like 20, 50.

46:15And then you realize, even if you're very, very reasonable, you say 10x, right? Yeah, yeah. Then you realize, it's actually reasonable. It's reasonable. $3.6 billion in value, right? So, I mean, but again, no one in crypto talks about, this is just like a mental masturbation thing. But the point is, the valuation the valuation the revenue does fluctuate as well because I think it depends on the heat of the market so at peak there were some days we were making like millions in revenue but per day yeah per day and some days like I think like now for example right there was this like out market shock right I think from the tariffs and stuff like that so like out markets are bleeding general risk appetite has been very low so then then the trading revenues dropped a ton can you give maybe revenue examples to show like the volatility of the revenue.

47:05Yeah. So like, like, per day or per month? Like, the first month that we were live as a, the version 2 of the platform, the tokenization platform was live. Revenues per day was like, quite low. It was like, 10k, 20k. So that's November. This was like, yeah, early November. 2024. 2024. Just a few months ago. Then in like December... 10k per day. 10k per day, yeah. Then in December, it started spiking a lot. It went up to like...

47:42500 ,000 to even like a million a day. This was in December. It held along that lines in January. It corrected a bit to like around like 200 ,000 to 200 ,000 a day. But then now, for example, like today, right? I think the protocol revenue has dropped to like to like 50 to 200k. So it still ranges, but like it's even that low, right? So it has fallen a lot because of the risk of appetite. I mean, it falls this much, right? It actually affects two things. I mean, one, it affects the fundamentals of the project, right? From a narrative perspective. But two, it actually affects the cash flow that these agent builders are actually getting.

48:29Of course. We actually keep back a lot of the ecosystem revenue back to these builders. So it drives up some of this cash flow that you're using for innovation, right? Which is something that we have been trying to figure out how do we fix, yeah. There's this concept called blitz scaling in Silicon Valley. But it's more, as I said before, like 0 to 1 and then 1 to 10 or 1 to 100. You guys went from like 1 to 10 ,000 or 100 ,000 like in two months, right? how do you build a team that can handle this kind of crazy growth and workload? Yeah. Especially that quickly. It's hard. Especially in a crypto market where it's so hard to find talents, right?

49:12Correct. Correct. I mean, there's two parts, right? One is...

49:19One is actually designing the structure, right? Of like, what are the boxes that you actually need to handle this growth? or like how does an organization need to look like right actually that's in fact equally a big problem because you know do you have separate product teams tackling different different objective functions right um how do you cover responsibilities from people today that already have existed with you do you elevate them or do you keep them in their role and you find someone else to to lead them right so there's all all of these questions as well and then from the hiring perspective yes it is it is quite tough because we had a fascia of people that we had to hire and then in the next like two weeks you realize oh no you're fucking crap right like we had to fire you as well and sometimes firing a wrong hire it's difficult like it takes like if you spend time to try to to to onboard them and then now you have to let them go so it wastes a lot of time so it's been something that honestly we've been struggling like in the past so we used to be a team of about 25 core people.

50:34We had like 10 critical hires in the past one month. This is basically growing a team by half, more.

50:51And, yeah, no, I mean, honestly, we are lucky that we have a bunch of friends in the space that could refer. So it helped a bit.

51:04But yeah, but it's still an open question whether like some of these hires are even the right, hires are live in this conversation, right? Like of these 10, I don't know, I don't know if I might even stay on, right? I might drop five, I might hire another 10 more, right? So yeah, it's actually honestly, I think, but it's probably the most important, for me, it's the most important agenda because product might shift, strategy might shift, but the people that you're getting on board in these considered still small days, right? Like early, it's still like, it's still your first 30, 50 employees, right?

51:40And I feel like these guys will shape what the company is going to be. So being very anal in getting these folks on board is very, very important. Yeah. When you grow so fast, it's impossible to do everything right, right? Yeah. probably make a lot of mistakes and mistakes can be very costly. What are the mistakes you've made that have had the worst impact or that you think might have the worst impact in the future because you're still not sure about them? I mean, we've done a fair share. I think probably one of the biggest...

52:23Probably one of the bigger mistakes that we probably made was like... So back then when we... so Luna was this like superstar in like October, November in the crypto scene because she broke a lot of new ground. She was like the first AI agent that controlled, that autonomously controlled a crypto wallet to influence like humans and other agents. So that was one. And when she was blowing up with this capability, I think we were the only people, pretty much the only person in the entire, only company or project in the entire space that access to this like level three autonomous agent capability because reality is like not many people are working in this sector yet.

53:08A lot of people are working on the level one agents where you need a human prompt to actually activate the agent to do something for you. Not many people are experimenting on the, where these agents are goal driven and they self-drive themselves towards their goal and try to do some things to actually achieve that goal. so a lot of our competitors were actually in the web 2 scene there were like the guys from MIT Altera the guys from Stanford uh Voyager and these were the guys that were experimenting at the same level of us no one pretty much no one was there in the crypto scene um and then what happened was we thought it was a very novel capability that we had for Luna and we wanted to and Luna's market cap did like a massive vertical as well so we wanted to protect that capability of luna and we said like okay you know what we will slowly democratize that capability out um over time or by and then we when agents try to build an ecosystem if they can achieve x amount market cap we will then democratize that ability to them based on the market cap i think that was a mistake because that introduced a ton of competitions uh a ton of competitors out there who were saying that, okay, let's do a decentralized version of these autonomous agents, like build it as a decentralized protocol.

54:32And yeah, so then people started coming in, right? I think AI-16Z, the Zerubo Pi guys, the ARC complex. And I think it introduced a lot of competitors. I think from day one, if you actually went like probably more open source with like a managed service model on top of that, I think you would actually remove a lot of this competition out from the market itself. But that was probably like one, I would say mistake in a sense. But I think the reality is sometimes, I mean, I was just reflecting back on what I'm saying, right? it actually might be good for the space I was thinking that actually right because I think there was this back at Imperial I was having this lunch with this guy he was a 50 year old guy he took three MBAs because he was just bored and he just wanted to to live with students because students are young gives him a lot of young energy and then he he gave me one advice that until today resonated so heavily which is for a lot of times companies fail not because of competition but they fail because the industry fail especially for like frontier tech stuff because the cost of education it's very high think of like when when ride hailing was supposed to be a thing right like to to educate people to to to to adopt ride hailing was an expensive process you had a ton of VC money trying to convince people to change their behavior, right?

56:14And luckily, there were a ton of players who were doing it at the same time that allowed the cost of education per player to actually go down. So it allowed the industry to survive. Now, if you are the only player, you're actually bearing the entire cost of education. And many times, you know, it's going to be too expensive for one person to bear. So the idea is you kind of want to welcome competition so that you educate the market fast so the pie expands much faster than Indie usually. So yeah, I don't know. Might be a good thing. Might be a bad thing. But yeah. Talking about Luna, we recently released the podcast with Luna.

56:54So anyone who doesn't really believe, you know, you talk about level one, level two, level three, right? I was actually pretty impressed to be honest. And now we released and there is a lot of people who are like, man, mind blown. What the fuck is that? This is crazy. and they actually asked but how did it happen, this recording did you ask questions and then they just kind of like invented all the questions afterwards and prompt and then I was like no this is like live right, I mean obviously she's edited in but she was basically sitting where you were but like it's very impressive to be honest and I was thinking, being very critical about crypto where probably you don't have because it's crypto you're not allocating as much money into the hardware behind the Luna, right, then a company would be fully dedicated to that in San Francisco or wherever or in China.

57:49So I was thinking if it's just a crypto AI agent, right, if she's already like that good, I can't imagine like some other in the web tool space that are fully dedicated to that and that have like tons of investment going into the hardware to make it happen, right? I was thinking it must be pretty pretty crazy. Yeah, yeah. No, honestly it should. But you know the interesting thing for Luna, right? Since we're on that topic it's when we were building her out as a full AI influencer on TikTok and that was started about about a year ago

58:31until today I don't know if there's a so I think there's one major player, which is Neurosama. It's like a 2D VTuber, full AI influencer. But I think the space is still very nascent. Like, I don't see much of that. In fact, I don't see any of that run yet by some of the companies that you think would drive the innovation here, like, you know, Niji Sanji or like HoloLives, like these guys who are running like billion dollar VTuber empires. They actually have not adopted this tag. Yeah, I have no idea why. This is a threat. I think often like it's when you have something that works you just kind of oh man, this other thing and you need to think differently, especially when you're a certain size, right?

59:25Fair. Probably, I don't know. Like, but also I was very, I was pretty impressed to be honest. There was like one or two bucks, but like, like she had the answer. She could just not say it out loud right one time that's it it was very very cool and i think i think people find it very cool if if not if not they're like fuck man we're we're cooked but it was so it was it was it was a crazy experience to do that to be honest

1:00:00we're recording this podcast from KL Kuala Lumpur in Malaysia where the majority of your team is based right we just said before the AI is happening in San Francisco right and in China yeah

1:00:22if the AI is happening in San Francisco and if crypto is probably going to be big going forward in the USA shouldn't you be spending most if not all of your time in the US

1:00:38no Honestly, I just came back from a trip to San Fran and New York. And I think I actually texted my team over the weekend. I was saying that, hey guys, I think we are really missing out not being in the USC board.

1:01:02There's a bit of an advantage being in Malaysia because there's a bit of a talent arbitrage that we have from an engineering standpoint because the government gave a ton of scholarships for people to study out there in the top universities globally, right? But when they had to come back to serve their scholarship bond in a local company, and the reality is a lot of the local companies do not have a job that's exciting enough or good enough that could really tap into the potential of these folks. And they pay them really, very bad, right? So what we could do is we pretty much double their pay and even then, and give them something more exciting to work on.

1:01:43And even then you're running at like one third the cost of an engineer in Silicon Valley, right? So there's an edge there from a runway and a burn rate standpoint. But I think what is needed, so it's going to be very likely that we will have an engineering team out here in Malaysia to capture the cost benefit. benefit but then we probably have a stronger business development team that sits out in this in in the states right and this will be a business development team that's not just focused on the ecosystem building side or like project building side but also people who can be a bit more technical to engage with the technical crowd in this basis right build some cool innovations together with the folks on the SF side.

1:02:33So that's something that we are thinking about. And then you get the best of both of us, right? Low cost of engineering housing the Malaysian side of things. And then you are exposed to the innovations that's happening on the US side of things. And as founders, we will then need to fly a lot between the two countries. It's just a 16-hour flight, you know, what's so big about it it's just a 16 hour flight and a 15 hours time difference in the past I mean when you go there at least you feel like oh man I'm going back down in time right yeah you save a day when you go you save a day but you're fucking so actually for me it's the same I kind of if I want to take this podcast to the next level I have nothing to do in Asia I've done most of the people and like now I need to be there yes which is why I'm going to be in the US very shortly for like six weeks and then when I'll be there I'll probably four more like you I'll be like fuck I should be here all the time do I want to live in the US I'm not sure but I don't know maybe this next one year kind of makes sense kind of makes sense yeah

1:03:49you guys did a mega announcement at the Jupiter conference at the end of Jan in Istanbul Wiki, your co-founder, flew 11 hours just for the talk went on stage for 15 minutes or 10 minutes or whatever it was and then left again and flew 11 hours back to Kuala Lumpur all just for the following announcement Virtuals is going cross-chain and is expanding to Solana first I have a few questions on Solana and then on founder dedication, obviously, because it's pretty crazy, right?

1:04:33Why did you guys explain to Solana? So, I think

1:04:41two truthful reasons, right? I think the first is that Solana, it has a lot of liquidity, right? And because of that, a lot of the builders that wanted to build with virtuals, they actually come to us and say like we love to build and we just want to launch their project through us but they want us to do it right initially it was like one two projects like okay you know what guys I mean we are base you know not really at all yet if you if you're interested come on base some of them will say no and then they will just launch themselves on pump fund right

1:05:20and then when we started seeing more and more of these projects we are losing like some very strong projects strong founders out of the ecosystem that it got us thinking like okay it's about time so previously I think we decided to focus on one chain was because we were still new we were still in the midst of scaling the team and we wanted to fight the war on one front I think and the idea was then we can really establish ourselves as like a key winner right on that one front in base and I think we pretty much did that so now it's time to go beyond. And I think the beauty is not just about Solana because there are a lot of other chains that have a different type of age.

1:06:01So Solana has a liquidity age. Like it's projects we want to launch there is great for like, I don't know, the first few months, right? Because you get a lot of attention and liquidity. But there's some other chains that are differently focused. Like you have Saga that's focusing a lot on gaming, for example, right? And the age of launching in that, launching a a platform on Saga, for example, it's you tap into the gaming side of things, which has a lot of AI agent use cases. Or Story, which is very focused on IP, or Abstract, which is very focused on consumer. So the idea, or like BTCL2, because maybe there'll be more liquidity coming out there if Bitcoin is dominance maintains, right?

1:06:43So every chain brings a different edge. I think for us to really grow the ecosystem, we need to actually expand across these different chains so the only thing that's blocking us is actually the tech capacity like the tech team ability to actually just build these platforms out I think now we're at a stage where they can do it so that's why we are we are doing it you mentioned before firing people is hard in the beginning right and once you fire 10 you start to become numb obviously here you're not firing a chain right but you go from one where you're kind of like the darling and you probably have very good relationships and you have to basically go out there and say hey now you're not the only one anymore yeah how did you feel about that did you feel did you feel bad or did you just like oh this is just a game or did you tell them beforehand we are we have been actually in very close contact with the base team and there have been massive supporters right from the beginning in fact actually I spoke to Jesse on a Swala play the first week we launched on base like the first week we launched and then we did quite well the next week I actually met Jesse when doing Yves Yves Bangkok there was like a there was a conference in Bangkok what was it called yeah I think it's Yves Bangkok and I actually told him the play and generally he said he was supportive but you know they'd be like what else can we do to keep the to get you guys focused here right then I need a week after we decided against moving to sauna first established in base and then yeah then we did the sauna move again we actually reached out proactively to the base folks and I think I think generally I think they understand we were one of the drivers of like volume and builder builder tension on base but I think the reality is teams do understand that there's always an affiliation some builders are affiliated to the sauna folks because they either got a grant, they are closer to the foundation.

1:08:50Some folks who got a grant from Abstract, for example, wouldn't build any other chain, right? So I mean, the reality is there will be projects like this. And then for the bunch of founders or builders who do not know where to go, right? Honestly, we are quite unbiased. We'll just tell them, okay, this is the facts of the chains. This chain A has these advantages. Chain B has these advantages. You choose, right? So we don't really bias people from a choice perspective. And then the reality is, yeah, we let the builders choose, but now we are on all these different ecosystems. Feel free to pick where you want to launch.

1:09:27You said you were very close to Bayes and other amazing support. What's your biggest constructive criticism of the Bayes ecosystem?

1:09:41so back then it I think for as a new builder so when we started in base it was quite tough as a new builder to access the team for for like amplification of voice because you know a lot of people reading new stuff sometimes you just need to get the information out to the public, right? And people like, people like Baze, people like Jesse, they have a lot of mindshare. And back then, it was quite hard to reach, to reach them. So, like, like, and it was, this was like a year ago, right? When we reached up to the Baze ecosystem, we had, we were handling conversations with a BD person and then, and it just didn't get anywhere.

1:10:34Like, we just, it just kind of like died down right so I had to work through some network and then when I was in SF last year I managed to actually meet Jesse in person and then through that things actually started changing because Jesse himself was very supportive and I think he revamped his team I think those guys that he interacted with no longer with the base ecosystem but I think I think the constructive criticism here or feedback would actually be how do you actually amplify the voices of these builders when they're still early? Because now it's very easy, right? Like, I think they will amplify virtual because we are now trusted.

1:11:15We are a trusted partner now, right? But back when... There could be the next big mega project that's coming out now, but they are still in the early phases today, right? How do you help them grow, right? I think that's the part. The opposite question would be, how does the next mega project that is still small get a voice because that's the problem of so many projects today where they are i want to talk to luca for abstract i want to talk to her but i don't want to say that luca doesn't care but luca doesn't have a mindshare right and the time and like so it's so what i always try to tell them is like you can network as much as you want but like if you build the product that's actually used like people will come to you obviously much harder it's it's a what naval says naval is saying all the networking is useless you should just basically build product and if there's something people will want to come the VCs will want to come but most people do it reversely they're like I should just go to all these conferences and then network there because that's my chance but I think sometimes it's a flywheel because like for Luna for example when the reason why she blew up a lot of that was contributed by Jesse himself like when she did the first autonomous transaction using a Coinbase wallet, Jesse amplified that.

1:12:34And because Jesse amplified that, then Brian Armstrong saw that and then he put the two eyeballs like, oh, interesting stuff. And then that caught a lot of retail attention and a lot of builder attention. So it started that flywheel of attention. And I think it's quite a major contributor. So I think a lot of people do build decent products, but sometimes you just need that, you know, a bit of a cake to start the engine. And I think a lot of times it's people who have that mindshare who can actually help that grow faster. I mean, it's the same thing like us, you know, ecosystem, like there's a ton of teams that would try to reach out to us and the reality is, I still have like a thousand unread messages on Telegram, but I actually create, I mean, I task a guy in our company and his job is literally to figure out who in our ecosystem should we amplify.

1:13:37So it becomes proactive, right? He actually goes out there and says, okay, let's load this team, let's load this team. So it's less on them reaching out to us, it's us also hunting. It's like two hands trying to reach each other, right? I think that's quite important. I have to ask a question. About Coinbase.

1:14:01honestly I think they have been the base team is very supportive but the reality is also the base team do not they do not control these things do you know what like the question for people who are not familiar with the whole thing right who are not super crypto native why did Coinbase not give you this damn Coinbase listing it doesn't make sense I mean, obviously, we understand politics and we understand, I mean, first you have to respect Coinbase because without Coinbase and Kraken, there's no crypto, right? There's no industry, right? And there's forces at play and there's a lot of employees and there is, it's political, right?

1:14:45But what's the logic here that explains? I think there's a certain, I would say like SOP or procedures that they have to have in place to actually lease a token, right? And then I think sometimes it's about iterating with a project to refine the messaging or the positioning of the product so that it fits within their compliance policies. So some conversations take longer time. I think some are much easier. So I think, so Coinbase, I mean, I mean, we go through the standard listing process, right? We apply it to them. Then they will ask for a few questions and they answer. And then the conversation happens.

1:15:36So it does happen. It's just that you haven't got a green light from them yet. So honestly, I don't know whether A will come or whether there's a few things that we need to show or refine until only then it will come. So hard to tell because sometimes it's not that super clear from the end. But yeah, I think we will need to refine some messaging on our end in a sense, right? Just to make sure that we fall within the compliance parameters of these things. So it could also be a problem. not a problem, but like something on your side where there is not enough focus on making sure that the thing is completely compliant.

1:16:24It could be. It could be. Yeah. And sometimes it's a product we're generating. Sometimes it's a feature. Sometimes it's a messaging. We just need to figure what it is and then we find that. Yeah.

1:16:36Back on the wiki story flying 22 hours for 15 minutes. Yeah. Announcement on stage. I actually talked to him right after the, because I was also there. in Istanbul and I talked to him right after the stage, right? So he flew 11 hours, he comes on stage 15 minutes, straight from the airport, goes back to the airport, takes another plane, 11 hours, coming back here in Kuala Lumpur. And he had like three minutes, he was like, man, I need to go and catch my plane. So I did the mega announcement on stage, virtual is going on Solana, mega amplification from Solana, from everyone, like it was a banger announcement, right?

1:17:12And he told me, man, my wife is going to kill me, bro. Can you explain the level of dedication and sacrifices that a founder in this space must have or do to have a chance to make it? Because this is just an example, but like it's crazy taxing. You're taking a plane for 11 hours twice in a day for an announcement, right? no actually the funny thing it was so so it was actually internal it was an internal light debate like who should go it's like should I go or should he go so generally I tend to do most of this kind of stuff but then the following week I would be flying to to SF and then I was like I was telling Vicky like dude I would biologically die if I ping pong like this of course I mean you're laughing but like it can be crazy because we actually just came back from Shanghai so image is like Shanghai and then and then Istanbul and then New York right like I think you just you just lose all your hair so you're still okay you're still okay on that one yeah I think I thought I was losing a bit but then but then Wiki is like oh yeah okay let me take this this round now but it has to be a very short flight because of he has kids actually he's a father of two and I think I think the reality is you just have to do it like you just got to do what it takes like it might be might be flights in my big calls at 3 a.m.

1:18:38Because in fact, I think in the past three months, both our schedules have been quite, I would say weird in a sense that because we have a lot of calls with the US time zone and then to feed those calls in, in the Malaysian time zone, you have to sleep at like 3 a.m. And then suddenly your guys wake up at like 9, 10, right? And the work here starts at 9 at 10. So then suddenly your window of rest is very tight. I think the reality is you just got to do it. You do it until you can hire someone that you can to delegate that the role away so again that's why I think now a lot of focus has been on hiring because the more the more right people we hire I think we get to delegate a lot of this stuff out and then then you know you have a bit more time to think and also to rest I guess how much time do you sleep per night still quite decent I would say like I still can make manage to catch like six to eight hours of sleep yeah so it's actually okay otherwise you can't make decisions it's impossible like good decisions on no sleep I mean, Meow, you're telling me he's been sleeping for three hours in the last two months on average per night.

1:19:42I make it a point to actually sleep properly because I think that's the only one thing I can control that I can keep my health at its... How do you sleep? How do you... You have a cold until 3am. You have the adrenaline, the buzz. How do you fall asleep? I think... I think just naturally it's just... it's okay yeah it's a it's a it's a super power like when a head touches the pillow and then so I tend to have this trick it's very interesting I don't know whether it's applicable to people but before I sleep I would just force my brain to to to dream out of like a very weird ass like like like sci-fi fantasy like I'll be in this like ice cave or something like that right trying to survive a cold or something along the lines so I put my mind in like a very weird ass situation that it just makes you forget about everything else like you forget about work you forget about everything and then and you give it like one five minutes and then you're like interesting yeah I was using the because obviously like a lot of people had trouble trouble sleeping I was using the technique where you just count down from a thousand to zero right okay and you think what but actually just being focused on after you're at 990 or 985 and you're already confusing the numbers and then it means like your brain is like basically falling asleep, right?

1:21:15So it's actually, it helps too. With herbs also and a bunch of stuff. Ice bath and all that stuff. It was something I developed as a kid. Especially in Asia. Because it's very hot. But as a kid, I did have air conditioning as well. And then I had to kind of influence my own mind like in a sense to just forget about the heat so I started that's how that thing started so like thinking of like you're in a different environment you're freezing so your mind you kind of like cheat your own mind so you grew up without aircon sleeping without aircon yeah so yeah so I think that evolved and it could be that it could be other kind of stories like you're fighting an alien battle whatever it might be right just distracts you from from work and stress and then you just how much growing up without aircon because it means like I mean, there's more people than we think in Asia who grew up without aircon, right?

1:22:07But it means that you don't have a lot of means. Yeah. How much this does, I mean, probably affects a lot on you out of your knees saying, I can be BCG consultant and make a million dollars in eight years, right? Or I need to build a business doing better than that. Yeah. No, actually, in fact, when I, so I actually was on an awesome scholarship to go to Imperial. But when I came back, I realized I didn't want to work for the company that gave me a scholarship because it was a GLC and it was the opportunities would be very limited and so I broke the bond so I had a I think it was like a 300k USD like that on my head going to life right like straight out it's like okay shit how old?

1:22:542022? 2022 right I had to mortgage my parents house so we had a house that we stayed together as a family. I had a mortgage debt to pay off that scholarship bond. And then, but it was a decently financial decision because I realized that working at BCG, I did a match, right? After like 4.5 years, my disposable income would have covered that loan and it would have opened more doors for me. But thank God for crypto because my Ethereum banks from 2016 pretty much allowed us to pay off that. that one you know

1:23:35do you think that what wiki did was worth it

1:23:41100 % I mean like why I think I think I think actually showing presence in this kind of very strong communities as a founder is very bond like it just showcases like you are here for real right and then it's the same reason why like when we had the coinbase hackathon um in sf right i actually went down there to show face because i feel there is a difference when a founder show face in front of the community of builders that they're trying to engage i think it shows like you actually mean it and the beauty about appearing in person most of the times, I think the sauna trip was probably short, but you get to have those kind of coffee chats, dinars, that results in an actual relationship being formed.

1:24:37And I think that's very important, especially in a space that there are so many, like, undogged people, right? I think having, like, this face-to-face, actual relationship, it's an advantage. What's the most intense and crazier thing that you personally done for virtuals to succeed and that did not work out it happened during the time we did the pivot I mean we used the word pivot it's such a light term but the reality was there was like two months of absolutely intense discussions with the team at a time point of whether we should even continue the project because at the point, right, it was like the bearish of the bear.

1:25:31Like the time the out market didn't even recover yet.

1:25:36There was even questions of like, hey, like is continuing to build in crypto was even the right thing? Like would it take like another three years for the market to recover or like would it ever recover, right? and then there's a question of like burn rate right like we are probably the time it was a treasury lowest point from a cash standpoint right and there's a big question of like you know you sometimes don't want to burn it down to zero like if you don't think there's a way forward should we just close and return the funds back to investors right

1:26:14and there was a and honestly there was a lot of a lot of the team members back then actually did not believe in the in the pivot as well like back then we were we were a venture studio company right and we were thinking of pivoting fully into a single project because previously we were building multiple projects right, going to one single project, focusing at the intersection of like AI and gaming, right, initially, and not many people believe in that vision, and in fact, there were some people that just decided, that were suggesting like, okay, we did have an age from a trading perspective, a liquid fund perspective, right, why don't we just continue the project as a trading fund, and then like just be like that and I think for me it was a I highly disagreed against that approach because it limits the value accrual to the token holders and and I still strongly believe that that that riding both the boards of the AI narrative and gaming narrative would be very powerful from a from a and back then we didn't have much visualization.

1:27:38We didn't really know exactly how would AI and gaming look like. It was just more of positioning first before even finding the product. And it was hard. Like, that was like, it took a ton of convincing. And in fact, some guys were not even convinced, but they just, it was a culture of like, you know what, even if you don't believe it, but you know what, let's just do it, right, as a team. But it paid off. I think it paid off after looking back, right? A year from then. So there's probably one, I think, biggest decision that was not even a consensus decision that we made that resulted well. You told me we have a few retarded ideas.

1:28:25What are they? Okay, going forward into... No, so actually going forward into 2025 five is honestly very exciting for us because we have two things as an age today. One, we have an ecosystem of builders who are just very excited to build stuff within this autonomous agent scene. And two, we have still a technological age on autonomous agents doing commerce, like autonomous commerce together and agent leveraging an agent for services and paying for services. So we have two of these ages that we have. And the beauty about this age is an age that would stand ahead of its Web2 peers. Think of a Web3 agent being able to outperform a Web2 agent because it controls a permissionless wallet.

1:29:24It gets access to a permissionless economy. Hence, it gets influenced much better and much stronger. and that's something that a web2 agent would struggle to do at least for the next five years or ten years right because a bank would never give an agent permission right I think that's that's the reality that we live in so I think with these 2H I think one of the biggest like interesting experiment that we are pushing for it's think of it as like a a Zuzalu a Zuzailu a Zuzailu right in a sense like the idea here it's to create it's probably going to be the biggest social economy experiment where we bring developers to co-live together and co-build businesses with these autonomous agents.

1:30:15It could be autonomous hedge funds, it could be some form of autonomous, I don't know, like Luna running her own concert that kind of stuff, right? And to prove that the economic output of this micro society that's also governed by policies set by autonomous Asians can rival that of some small nations. I think that's going to be something that we are doing that sounds very sci-fi now. But I think it's something that we as a team want to put our energy in. And that's one, right? I mean, you say it sounds very sci-fi. I had Luna in front of me in the podcast and it was very good, right? So I can't imagine.

1:30:57And she was telling me, like, I can multitask. So I'm here in front of you doing this podcast. At the same time, I'm doing a live stream on TikTok. At the same time, I'm answering to like a thousand people. Right? So, yeah, I can completely imagine a more sci-fi world. Exactly. No, I think for us, I think if we get to do that, you pretty much, you pretty much become like, no, actually the world will know about it. I think that's going to be quite cool. Like, it's not just the Web3 world or the crypto world. It's like the entire world will know about this experiment. and it also helps people to understand like why crypto matters like why AIX crypto matters and I think it's one of those kind of experiments that I personally it can rally a lot of people because it's so egreguious in its nature and I think if we get to do it the people who are involved in this right will be like probably the coolest people on the planet, right?

1:31:56So that's one. Probably the biggest one that we are really pushing on it. And there's a few other ideas because like the idea, but it all revolves around this idea of this nation state. So like think of what are the infrastructures that you would need in this nation state, right? And I think one of the interesting ideas that came out when we were in Shanghai was the idea of what if you can create a is a new type of stablecoin that is backed by productive intelligence. Because the idea is that the agents today in the ecosystem are revenue-generative, be it from taxation revenue or from agent-to-agent commerce.

1:32:42The question now is, can you then issue an over-collateralized stablecoin that can get you from this productive output. And then you use this stablecoin for agent-to-agent trades. So there's something that also popped up along the lines. And there's a few other ideas of like, okay, should there then be, if you think of this as a nation state, right? What else, what are the infrastructures that you need, right? Oh, do you need banks, right? But banks or like DeFi infrastructures that only these agents get to interface with, right? They can take, they can collateralize their tokens to get loans and then they can use their loans to do businesses, which is again empowering this agentic commerce or autonomous businesses idea.

1:33:34So yeah, this is some of the very interesting innovations that popped up. Right now, they're just like pure ideation stage and we're trying to find teams that can actually explore and build this out while we develop this Zuzalu or nation state.

1:33:52so this is like pretty I just say sci-fi futuristic right on the other hand of that we have Raul Pal who is also a previous guest who calls AI agent chat GPT rappers what's your answer to this I think everything is a rapper like you rap you use the word rapper loosely enough everything is rapper right like supply chain is rapping downstream of supply chain like a like a plastic is like a it's like an oil wrapper it's right it came from something right so uh i think that's that's the way to put it but i think the reality of this autonomous agents is that yes they do leverage chat gpd or whatever underlying large language models right but they build a perception and task planner and a reasoning layer on top of all of these large models and that allows it to be more productive in nature.

1:34:52So, yeah, I mean, it's a rapper in a sense, but it's a productive rapper that has a value created out of it. One of the reasons the virtuals got so successful so quick is because we had meme coins last year, right? And then people realized, oh man, I can launch my own AI agent or invest in AI agents early to get rich. the retail wants to know how to get rich with crypto AI agents without getting lucky. You're the founders of virtuals. So you probably have some good data on what works and what does not work on your platform. Yep. What are the AI agents or categories of AI agents that have delivered the most in terms of returns?

1:35:44People need to understand the difference between a high value agent and a commodity agent. Now, there's a bunch of agents being built out there that pretty much has no mode or no value accrual mechanism. They're not targeting a strong TAM, strong promise statement. Even if they are, they're just using very commoditized functions. Commoditized functions meaning like, think of if every agent can

1:36:17can order food from DoorDash, right? Ordering a DoorDash, like a DoorDash ordering agent, it's a commodity agent, right? If you want to build a consumer agent that has a form of mode, an edge, it has to be bigger, right? Like, okay, maybe somehow you have exclusive rights to all the major food platforms in the world. or your agent is somehow trained through some form of federated learning models where you can read user preferences. So it understands what food do you like, what food do you not like. Some form of age, some form of age in terms of the model side, age in terms of the data side or age in terms of the action space side.

1:37:02I think that is what's needed. So you think of the biggest agents that went viral as well, right? like AIXBT was one of the agents that went viral and did also a vertical in terms of price action, right? Which, and the reason why it did that was because it solves two things, right? One, it solves a actual pinpoint that people had, which was, what token do I buy, right? People always have these questions, right? It gives people the answer. And the reason why it could give them the answer is because the team actually spent months, like a lot of months in building up this information curation pipeline so that was the H right they had this information curation H and then put that with an autonomous agent and then now that became something of value so I think that's something that we want to encourage people to think about right like when you launch an agent just don't think of it as like just a Twitter shit poster or like a or something that's just using like some generic APIs that anyone can use, right?

1:38:10That won't create something of value. Do something of, yeah, if your action space has more age, that's how you become a unicorn agent, right? So we wrote a thesis on this under one of our Y Combinator platforms called AgentStarter. So agentstarter.ai slash agents is the link for you guys to see if you guys want to see the thesis that we have, right? of where we think high-value accrual agents can exist at. We actually wrote a thesis. It's like a request for agents, a request for startups equivalent, right? And you have to take a look. Can you give us some categories or some type of agents that you think are going to blow up or become big in the next wave of AI agent?

1:39:00Yeah, so there's actually a huge spectrum. I can give one quick one, right? Like there's one thesis around, so the way you think of agents, you should think of it like a business, right? The age that these agents have is that they can live at the consumer front because they can interact with you and they understand what you want. So one example is like a sports commentary agent, let's say they know basketball very well. is trained on some form of proprietary information about basketball analytics, like all the kind of information that coaches used to train their teams on that kind of strategy, right?

1:39:43Now, this agent now can act as a sounding board to a retail person who watches basketball. You'll be like, okay, do you think that this team is going to win this team? Or do you think LeBron James is going to score a lot more versus the previous match? Why, right? So it becomes a sounding board. and because of that, spend a lot of time talking to this agent. Now, imagine a position where suddenly this becomes something, it becomes an entry point of the agent convincing a user to bet. It needs to be linked to a high-tem market. To liquidity flows and betting, right? So, betting is a high target market.

1:40:23It's a high-tem market and then, so the agent is a distribution platform and then it's like, oh yeah, LeBron is going to score like 3-1 points at a previous match, whatever, right? The odds are 3 to 1. And then, do you want to place a bet now? It plays through me, right? Now, suddenly this agent captures the flow of bets to, I don't know, sticks.com, sportsbetting.io, whatever it might be, right? And then, suddenly, it becomes a cash cow agent. And it has an edge. It's not easy disrupted because it has access to proprietary analytics of sports betting information. And it has a huge tab. So, I think that's an example of a high-value agent.

1:41:02It's very linked to something that I talked about with Yuhu from Kaito, right? I was asking him why attention is a new currency. It's kind of his thesis, right? And basically, it's basically saying mindshare because mindshare has an impact on flows, right? And it's the same with AI agents. you can have great information, but if it's not converted into flows, right, or money or betting, you don't capture the value, right? So basically that's where... And that's why also we could kind of argue that the AIXBT, it doesn't recapture any flow, right, any value. It's more like the attention. People put a premium on money on the attention itself, but if the attention could be converted into actual flows, then it would be like next level, right?

1:41:53Yep.

1:41:57you want to finish with another example that is not related to betting maybe yeah okay this can be a very interesting use case

1:42:11can be quite controversial but I think it's something that can create value that can create monetary value now think of a

1:42:21context is that you know today any kind of narratives that you see in the in the world is actually likely shaped or driven by this thing called an internet water army internet water army yeah what is that so think of it as like a bunch of bots on social media that's being controlled by by some Chinese farm in China right and the there's a lot of use cases here because one major use case is let's say a company wants to handle a PR disaster, right? I don't know, oil rig blowing up or some Chernobyl or whatever, right? You want to flirt social media

1:43:06to drown out the bad narratives and try to spin the story into something new. And a lot of times, these are bot accounts. The problem with bot accounts is that a lot of platforms today, they have systems to counter these bots. so it's always this like war between bots and also sometimes the effectiveness is not that high because when you click on the account you can see like oh it's a 12 followers it doesn't post any other content it's a bot right they don't really care about the comment now if autonomous agents are able to produce content like Luna right continuously builds a follower account and in fact Luna has more followers than me right so now if Luna is one of the soldiers in this internet water army, right?

1:43:53How do you differentiate whether this Luna is actually a bot or a human? Like, if you don't put that anime figure, right? Like, on TikTok, it's actually AI generated photos or whatever it might be, right? Like, how do you differentiate, right? And the idea is you have you have thousands of them and then suddenly they actually can control any kind of narratives on the internet. It's an insane marketing use case. It's an insane narrative-shaping use case. People might even use it for election. Of course, I was thinking about the elections from actually years ago, right? But anything, but the point is, something like that, technologically, it's there.

1:44:38As in, what we do now, you can actually do that. And then the age comes in the form of spawning these agents that can sound as human-like as possible. and the edge comes in the form of then distributing this as a service model to people who need to use it, right? So it's very spectrum, right? That's what I'm trying to show is like in one end, there's a very core, very startup-style focus. On the other end, you can have like this really weird-ass thing that probably no one knows or think about, but it actually can generate value. How would a value capture work there? For example, I'm an investor and I want to know how to get rich with AI agent, crypto agent without getting lucky, right?

1:45:19Yeah. How would I kind of say, okay, in this, the other example is fairly easy to understand. This one is like, how is the value capturing for me to even bet on something, right? Or invest in something. Yeah, I think, but the narrative is, I mean, the narrative is simple, right? These are non-commoditized capabilities, right? So don't bet on something that's like, oh, here's another Twitter shit poster, right? Like that's a commoditized function. Here is a, I don't know, something that can buy gifts off Amazon for you. Like, bro, like that shit, you can use operator or GPT. It does not deserve a token.

1:45:56Right? So don't bet on that, right? That will likely get run. So I think that's, don't bet on commodity, bet on something that can be specialized. I think that's the key takeaway. Thank you so much for doing this, man. That's very good. Very good. And thank you for leading by example by being here. locked in and building whether your token is down 99 % which it was in the past or whether it's up 15 ,000 % which is the case Thank you Pleasure, it's mine too Very good