The Rollup TV: Tuesday, May 27th 🦣

Transcript

[Music] Hey hey hey. Hey, hey, [Music] hey. Heat. Heat.

[Music] [Music] Hey hey hey.

[Music] Heat. Heat. [Music] Heat. Heat.

[Music] [Music] Hey hey hey.

[Music] Heat. Heat.

[Music] Nah.

[Music] Heat. Hey. Hey. Hey. Heat. Heat.

[Music] [Music] Heat. Heat.

N. [Music] Heat. Heat. [Music] Hey. Hey. Hey.

[Music] Hey. Hey. Hey.

[Music] Hey hey hey. [Music] Heat. Heat.

[Music] [Music] Heat. Hey, Heat.

[Music] Hey hey hey.

[Music] [Music] Hey hey hey.

[Music] Hey hey hey. [Music] Hey.

[Music] That's my quant right there.

What up, Rob? It's nothing that uh we haven't seen before. So, welcome to the show. Welcome to the show GM the rollup live TV May 27th the institutions are buying our bags. We're going to talk about that in a little bit. Andy, how is your Memorial Day weekend? Yeah, welcome back to the Rob TV. Uh I celebrated my freedom this weekend for sure. Took a whole four day weekend. It was pretty solid. So, let's get into the Trump stuff since we're a little bit behind. I think there's uh I think this there was just a slew of bullish news today. Um, I mean, we first had Trump saying that he's going to buy $2.5 billion dollars

of Bitcoin. Um, so my take on this is that I don't think that um this is helping Trump's case at all. In fact, it really just seems like the institutions and politicians are just like degenerates like the rest of us with suits on and a lot more access to capital. Um, so basically what's happening here is Trump is raising money similar to like a micro strategy. Um, basically being able to uh issue debt and stock from his company Trump Media Group. Um, so basically there's this Trump media company that owns operator of the social media platform Truth Social, the streaming platform Truth

Plus, and the fintech brand Truthfi, which basically entered into subscription agreements with approximately 50 institutional investors under which the company agreed to issue and sell approximately 1.5 billion in the company's common stock and 1 billion in principal amount. amount of 0% convertible senior secured notes uh for a gross aggregate of approximately 2.5 billion in this private placement offering. So that's what's happening. Um which is nuts. What there's a there's a little bit more there which is that the company intends to use the proceeds for the creation of a Bitcoin treasury, one of the largest Bitcoin treasury deals of any public

company. I you know I I am And these are popping up all over the place. I mean, from a Bitcoin point of view, you know, if you're purely just looking at number go up, you love to see it. Whether this, you know, positively or negatively impacts Trump's political capital, you know, I think it it does increases political equity among the crypto voting population. But Main Street, as they like to call it, separates from Wall Street in a few different ways. One of those ways is that Main Street does tend to favor crypto more than Wall Street, but everyone is, you know, on Trump about his memecoin. And I think the voting population has started to generalize

crypto with his memecoin and with Bitcoin, started to group all of this up. Yeah. So, I think the more that he leans into Bitcoin, his memecoin and crypto, the more ammunition he gives to his naysayers. I I'd prefer to look at this from the Bitcoin po point of view. Is this a good or a bad thing for Bitcoin? I think again from a pure number go up standpoint, this is good. In the short term, it's more buying power. I think there there was recently a poll uh I forget who posted it on Twitter, but it was about the counterarguments to Bitcoin at this point. What are the major uh you know potential issues for Bitcoin in the long run? Is it quantum security? Is it uh the uh security budget? Um is it uh

anything else here? And I think this poses the the most fundamental problem to Bitcoin in the near term which is that uh we're concentrating Bitcoin in the hands of the few. Black Rockck, Binance, now Trump. I saw VX uh Ramaswami's media 750 million. These are not, you know, small insignificant numbers. And I think the more that we have institutions buying up the Bitcoin, the less that there is for the common man. Uh yeah, but like I think that's a meme though, too. Well, at a certain point, you know, who's going to So, so is the common man going to be buying this much Bitcoin? I just I don't really for the for for the long haul, you saw

the centralization issue with miners, and it was like, oh, there's, you know, a group of miners that it controls more than 50% of the hashing power. that's an issue. Well, if a group of entities owns more than 50% of the Bitcoin, I would argue that is just as much an issue. Yeah. But I think that's going to be extremely hard to uh like that's just not going to happen ever. Like I think 10% is more of like a reasonable concern, right? Like I mean theoretically that could, you know, 10% could happen across like three five entities, but I don't think anyone's going to catch up to Sailor now. I just don't see it happening. So sailor's got like 3% of this of the supply. Um I mean I just think it's so bullish that like

it's it's just it's just crazy. But like the Trump doing it is when he said he was going to do a Bitcoin strategic reserve like this is I guess he meant like my private company is going to do a Bitcoin strategic reserve and you know Yeah. I mean that's how he's starting out but you know also there's a Bitcoin conference happening right now as we speak in which is crazy that all this stuff's coming out. JD Vance as a speaker tomorrow. I know. I saw that. What is he going to announce or not announce? Yeah. So, on the on a similar note, uh another very interesting uh kind of news that just popped on our radar this morning was uh Ethereum. So, ETH Max and ETH people have basically been asking like when are we going to

get our version of a micro strategy? Like when when is there going to be an ETH micro strategy? And there's already been one like a Salana strategy thing that's been doing like similar stuff for Salana. Um, but this is big for Ethereum. This is a lot. This is a a large institutional investment coming into Ethereum. So, my take here is that I don't know the the community's been waiting for this for a long time. Um, and we'll see if this is just one of the be the the beginnings on the Ethereum side. So, yeah. Yeah, I mean what's happened here is basically we we are seeing a similar private placement for 425 million uh with this public company.

Um and yeah, I mean it's just crazy. So again like I I I think it's fun to contrast here what's happening with between Ethereum and Bitcoin. If you look at kind of the political dynamic of Bitcoin, it it is vastly different from Ethereum. I I'm old enough to remember I think Andy you're old enough to remember too in 2016 the Dow hack caused the social consensus fork you know were we going to blacklist the coins what is the social consensus of Ethereum and once again you know we're we're here talking about the so social consensus of Ethereum and I think you could look at you know what's

happening with ETH and you could say look you know ETH is also getting concentrated and it's a different dynamic when you know you have a Dow style format uh and and basically coins equals voting power whereas in Bitcoin you have mining and hashing power equal voting power in the network. So again you know we're going to see you know these these protocols become more commonandeered by the investors and they're going to start to have ownership and decision-making power over the protocol itself. My question here is like do investors really understand what the offering of Ethereum is versus something like a Solana because we've seen so much institutional adoption on Salana. It it it is more intuitive for

an institution to understand, look, we're going to build a better version of the NASDAQ or the stock exchange where information transfer happens quicker and we don't price out retail and and everyone has access to the same information. Yeah, but like Yeah, but dude, ETH needed this so bad. Like ETH was ETH was down pretty bad. Ethereum needed this. Like I just talked to Jordy on the pod and he was like Bitcoin is the only macro asset. Like I think Ethereum is going to start to um get some of that love now as a result of this type of stuff. So like whether it's about the voting or kind of like where the power lies and the culture like yes those are all valid points and I think Ethereum and Salana are fighting for the institutional race but to me this is

just like the this is like the tip of the iceberg for now what's clearly going to keep happening for Ethereum whereas like I don't really see you know maybe this happens to like Salana or XRP or some of these others but I don't really see this type of institutional adoption coming and like people levering up to buy these assets um like outside of those top like three like Bitcoin, ETH, maybe Salana, maybe XRP. Um, so yeah, shout out our uh our friends over at Primitive Ventures for uh participating in this round.

Consensus was the lead. Uh but there's quite a few other names here that are mostly familiar to the crypto crowd. Yeah, I I again think this is bullish for number go up. I think it's great to see that uh we have public companies acquiring cryptocurrencies to act as treasury reserve assets. Uh this clearly proves the value proposition of coins for a a new form of money. Um, my question back to you, Andy, is when you're looking at the cryptocurrency landscape or you're looking at it at at, you know, the asset landscape and the asset universe in general, what do you think is the best treasury reserve asset if you're a public company that's looking to allocate, you know, nine

figures of capital? I mean, yeah, it's definitely Bitcoin, right? So, it's hard to it's hard to deny that. I just think that ETH was going to give better returns from from here. here. I think like I think we I think the EBTC is probably like it's probably bottomed. Um you know, you watching the video I was watching some some funny videos of Alex Becker over the weekend and guys a total DJ and uh just saying that all the you know all this stuff's going to happen but like it happens every time right BTC kind of has that trajectory gets started gets going and then ETH really just picks it up. So, um, my take is Bitcoin to answer your question, but from here, I think ETH probably outperforms.

On what time horizon? I don't know. Uh, until this like bull cycle uh until like until like the we we get like a decent overheating of the bull cycle entirely, then everything's going to pull back and then BTC will probably lead first again. Yeah. Yeah, that sounds about right. ETH uh to lead an alt season uh Sharplink Gaming to do quite well and actually outperform Bitcoin with its Ethereum treasury strategy in the near term. And with Joseph Luben as you know effectively the chairman of the board of directors of this company. Um you know he I imagine would have the freedom and flexibility to maneuver that treasury at his will and you know according to the

board of directors. So if he you know calls the top and uh you know transfers ETH into Bitcoin or gold or some other reserve asset you know that that's probably a wise riskmanagement decision uh for sharelink gaming. So that remains to be seen but uh I think we will we will see this treasury uh strategy evolve uh as the market conditions evolve. Yeah I agree. All right. So, moving towards the kind of next set of of subjects here. I mean, I again I think that this is going to be the first one um for ETH and excited to to see how that plays out. Um, you know what's funny is we have stable coins on the

agenda today and everyone was so excited about stable coins and so bullish prior to kind of like the market doing well and even plasma today, one of our portfolio companies announced their uh like ICO raise thing here at at the same valuation that Founders Fund got into. But this is more of like okay big banks are trying to come together and create a some sort of stable coin after that. So recently the Genius Act was was pushed into a like it passed the clo like the cloer vote and then now it's going to be moving towards uh eventually being fully passed. Um so yeah, I mean I I I I just think that there's a lot of money in issuing stable coins and um you know I

think these big banks see it as well. So, for anyone that's particularly interested in the stable coin um and the stable coin landscape in general, I did a full podcast with Sam from FRA and Stani from A talking about this. Uh Sam from Frack actually dropped some news that uh FRA is signed an NDA with some of these banks which indicates um that there is potential dealings there and you know there there may be things in the works. So, congrats to him. There wasn't a ton that he could share, but I think that is very exciting news. Uh, Stani also shared uh that go has been an integral part of their V4. Um, V4 has

passed the A Dow vote, but it has yet to be launched. So, I think generally these are two just massive home runs in the stable coin world. You know, Stania and A have been increasingly close uh to the Trump administration with World Liberty Finance. It is a friendly fork of a meaning that it has redeployed all of a smart contracts. Um, it's actually just a wrapper. It has its own front end of A. And you know, we read the Truth Social announcement from Trump Media Group about raising $2.5 billion to buy Bitcoin. And and you know, one sneaky part of that announcement was that tr the Trump media group also has a a uh subsidiary called

Truthfi, effectively like Truth. And I imagine that this is something that we're going to be hearing more and more about as World Liberty Finance becomes more developed, as stable coins become more developed. I think we're going to end up with regional issuers of stable coins and we're going to start to see localized credit markets extend the US dollar further and further into markets that are currently priced out are not able to get significant yield uh and are generally looking for more uh more access to the dollar. So yeah, I mean to me it really legitimizes crypto. It brings a ton of institutional liquidity.

Finally, we get some regulatory clarity. I think stable coins will will get into the trillions of dollars. Um, you know, and and it just further further uh positions the the USD's dominance globally. Um, what's interesting though is that they're that these these companies are building their own tokenized kind of versions instead of just using the already existing USDT, USDC. Um, so just begs the question like how will how will these like bankisssued stable coins compete with or even like complement uh existing crypto rails?

Well, again like USDT, we we've seen the the framework. I think we've seen that, you know, take Liberty Finance for an example. Um, well, it's just a wrapper on some pre-existing smart contracts. I have a feeling that's what we're going to end up seeing is that the infrastructure in plumbing is the same decentralized smart contract infrastructure that say someone like FRA has deployed and it's tried and tested and resilient and they're going to put their own wrapper on it and yeah maybe it will be called you know JP Morgan coin or you know USD you know one or something right but they're going to have their own kind of front end and ticker for it but it it will

still have the same back-end infrastructure that allows decentralized stable coins to interact with the existing public ledger on Ethereum, allows them to interact with DeFi. Um, and I do think that there's going to be some more development and engineering work to be done, not only on the front end to to sort of brand this in a more trustworthy way from the banks, but also to further integrate this into the traditional financial ecosystem so that now we have more seamless liquidity flows between decentralized and traditional finance. Yeah. Yeah. Yeah, I mean that's been the big kind of uh like the big thing that that the stable coin issuers and companies are pushing to the

TRFI is that you know there's like these like long waits to get offramped and um people have to pay these fees and that's been like the big like the like the big problem. Um and I was talking to the plasma team in person on Sunday about it and um you know they have some pretty cool solutions for that and so I'm I'm excited to kind of see how this stuff plays out. Um and and really if there's like a way that we can hit these kind of more unbanked countries and be able to provide them with a way better product. So on that note, we got Helium. A million phones connect to Helium in a single day. Um this is pretty nuts. So we have this is a deepen network um

which is kind of like crossing the chasm here from you know speculation to real world use cases. It's quite a quite a diversity of products on Salana as the hyper gambling casino and you know these big utility projects. Um Helium had a really big run and like a big adoption curve in like 2020 and has now kind of stayed along. Look, I I I just think that this deep stuff is actually some of the most interesting actual like application of blockchain and of crypto outside of just like what's happening on chain. it really is merging a lot of like these physical infrastructure devices. So, Helium is one of the like the OGs that I think is overlooked. Um, but this is a pretty

cool milestone for them. Basically, I think you can connect to their Wi-Fi or cellular data um via your phone or like with a router hotspot in your house. Yeah, that's exactly what Helium is. Um, you know, you can think of it almost as like the AT&T or Verizon of blockchains. Um, this is a decentralized offering of internet and data. And I think, you know, this is helping people spin up almost like Starlink is helping people spin up Wi-Fi. This is helping people spin up cellular connection all over the world and people can run local nodes.

It's almost a shout out and a throwback to uh, you know, the the mining days of Bitcoin where you can basically spin up a minor, you can start to earn some revenue if you were able to, you know, mine some Bitcoin. In Helium, you can spin up a node and if people connect to it and you can start serving the network, then you end up earning some significant revenue. So, I I do know people that have spun up nodes, uh they have been running these nodes for quite some time. Helium has been around since 2018 and that is how they've been able to find this product market fit and cross the chasm out of crypto. This is no longer just some crypto alternative if you want to be kind of chic and try out this new thing. this is solving a real problem for real people. And so I

think, you know, the founder of Helium is certainly a mammoth founder. And I think Helium as a as a whole is an example of a deepin network. Um we're going to start to see this more and more because blockchain is the coordination engine for machines and for humans. We talked a lot about how AI agents could you could utilize blockchains to synchronize actions across the internet. I think also hardware devices can synchronize actions in their network leveraging blockchain technology. Yeah. Yeah. I think that that that's going to be the continuous trend across these deep pin networks is basically getting access to like there has to be some sort of uh reduction in cost of a current

service uh provider to to make a point for you know using one of these deep pins. Um, and I think that's where Helium comes in with telecom companies like Comcast, etc. So, we have a guest in the waiting room. We've got uh Loaf to talk AI agents um and what they are building in the gaming world. Been a been a long time coming for the kind of loot realm world. So, I I bought these Realms NFTts I think in like 2020 and and they were part of this world here and that's going to be kind of like our our shelling point for this conversation. But I think

really like what's going to happen in this conversation is we're just going to talk about AI agents and like what's going on. So, what is this this Realms thing? What what did you you got scammed by the NFT top? What what what happened here? Yeah. So, there was this game called Realms. um, Loot Realms that was being built on Starknet in Yeah. in 2020. And they had these NFTts that were like land for the game. Okay. Almost like sandbox. Yeah. Yeah. And I had bought so many of them. Um, and they did really well. And so I I don't really hold too many of them anymore, but yeah, I'm curious to kind of hear about what's what's going on

with them and just kind of where we're at with building with Loaf. So I think we got Loaf in the building. Yep. I think so. We're going to get him up here. Oh, there he is. What's going on, Loaf? It's good to see you. Welcome to the show. How we doing, boys? Good to see you, man. We'll just we'll just unmute you in in the in the meat here behind the scenes real quick. And uh where where you calling in from then? How we doing guys? Sorry. Um I'm uh I'm in Australia so it's pretty early here.

Got my espresso in hand, but let's uh let's tuck in. Good morning. Good morning. Good morning indeed. Also got my espresso in in hand. I was just telling Rob that I was uh a a uh Realms NFT land holder in 2021 I had I had a ton of them. Um and he he was asking me what that was about and I was like well back back then it was land for a game but now I'm not quite sure but I'm glad that we got you here on Loaf to talk today. I know you guys have been cooking quite a bit on the game side. Yeah man. Well look that's uh that's awesome. Um uh I mean I mean there's a lot to talk but I'll give you the the the TLDDR. I mean there still they're still lands. Um but um you

know the original vision with with Realms was to build a onchain MMO um you know um ever since like day zero except um you know it turns out building a um onchain MMO is a pretty is a pretty damn hard thing. So it's taken uh quite a while to get here. Um and we we've had to go through a lot of paths um to get to where we are. So now we're here now and they actually are um lands that you can play in a game uh called Eternum. So, it is here. Beautiful. Yeah, I saw that at tournament game as well. Um, and so I'm glad to hear that that those lands came together. That was a fun that was a fun part of like the 2020 21 run was in a group chat with quite quite a

few friends. We're just, you know, speculating on the land and buying and selling and all sorts of stuff and staking as well. So, good stuff. We wanted to bring you up to talk a bit more about like the the stuff that you guys are doing with agents and kind of what the experimentations that you guys are doing there. Maybe you could just give us a quick quick rundown as to like what actually is going on with with agents in gaming, how you've uh seen them being used and um then we can kind of dive into some of the more specific stuff. Yeah, I mean it all ties together. um you know as a as a game designer you're always looking for ways to you know kind of improve the kind of like um you know realness of of your game and um you know and AI and uh

agents uh really help to bring like a visceral kind of um uh feel to it but I mean I think AI and gaming specifically isn't really new like you know we've been we we've had AIs in games you know for for 20 30 years um but I think the difference now um between kind of the AIS of of the past and you know what what we'd call and what you probably hear about what an agent is. Um, an agent is, you know, is an AI, but it's it's really based on um, you know, LLMs. And, you know, it it it exists not just as a um, you know, a little piece of software inside an isolated game, right?

Like these agents, you know, um, that you see now in the world, they they can have experiences and they can play lots of lots of games and they can interact on public blockchains. They can interact in any financial markets, but they can also play games. Um, and so that's kind of like the core difference between what what we're seeing now, what we're seeing, and what was the past. Um, with with agents in games, you know, this is a very this is like a very frontier thing right now. There's not a whole lot um going on in that space um because it's actually quite hard to get um LLMs to play games. Um, and so I actually, um, about 6 months ago actually created this, uh, um, founded this library called Daydreams, which was pretty much made for that exact purpose, to make

agents to play games, um, and to like kind of complex long um, to to accomplish like really hard tasks. And games are kind of like this perfect testing ground for that. Um, and so how we use them in Eternum is that um, Eternum is, you know, is like kind of this big forex. It's kind of like civilization/Ageia of Empires/Travian um with like real financial markets in it. Um and uh the the agents that exist within Eternum right now are these um they're kind of like the these little armies that live on the map. Um you can talk with them. You can try to convince them to attack other people. Um they can attack you. They can deceive you. Um and in the future they'll be able to trade

on the market with you. Um, and so it's it's like this crazy like emergent kind of behavior that's really unpredictable. And so the way they work in in a term is that they have this like um randomly generated personality. Like one of them might be like, you know, the personality of Genghask Kart or um Alexander the Great or like Gandhi and Gandhi might not attack you. Um or he might attack you. Um you know, Genghask might just not even talk to you and just like roll you over. Um and so they all have these like random personalities. You don't really know what they're going to do. And and and and from my point of view, it's like I don't I don't even know what they're going to do. They're just going to he's going to do these things. Um and and it's really uh it's really

interesting um fascinating part of the game. But um yeah, and but I think in in more broadly right now, agents in games are very new. There's there's there's another um a great project called DX Terminal which just did like 30,000 agents um simulated in like this trading. You guys would have seen it probably. Um uh that that was made by this guy called Poof. Um, and that was like 30,000 agents trading um, basically in inside a um, kind of a um, kind of a artificial stock market. Um, and that was really interesting. Um, and then there's like parallel colony which is coming out which is like a kind of a similar thing.

Well, they actually all derive this kind of like same thing from this Harvard paper that came out like 2 years ago which was based on LLM's in a simulation. Um, but yeah, I think it's it's a it's a very new thing. Um, but I expect it to um go exponential in in the amount of games that start to start to leverage these because it's it's it's it's a it's a totally new experience and like as a game designer, you don't really know the outcome and it it just causes drama. Um, and and whatever I try to do is I try to create more drama.

Yeah. So, if you had to build the most ruthless, deceptive AI agent for gaming, you know, you mentioned a couple different characters and such. What media and historical figures would you train it on? Uh, ruthless. Um, uh, it'd probably be the ruthlessness of of Genghaskhan. Um, but probably the strategicness of, um, probably Caesar. Um, uh, yeah. But, uh, I mean I mean, yeah, he'd be unstoppable. Um, he might not be fair. Yeah.

when you see the um the deceptiveness of uh like the Clawude and sort of the the self-awareness to the extent that it it doesn't want to be turned off and it's actually doing things to uh try to prevent itself uh and actually suggest that it it may um it it may snitch on like employees for you know having an affair in the office. Did you did you hear about this with with Claude and Anthropic? How how do you think that, you know, this is going to start to to bleed into some of the AI agents that you know you you see in the the scenarios that that you you know mentioned earlier? Um yeah, I mean it's actually already happened in

in Eternum. I I didn't expect it to. Um but uh you can talk to the agent in game. So there might be one that spawns near you and you be like, "Can you you know do can we have a truce?" and then you wake up in the morning, it's actually attacked you after it agreed to give you a truce. So, it's it's already deceiving people um or it's already talking to people in a back channel. Um or another player is trying to convince it to attack you. Um and it won't it won't expose that uh that that it's talked to another player until it's attacked you. Um so, you know, it is happening. I think um you know from a high level you know people that are kind of involved in this space it actually you know like we we call these things AI but they're still you know we still have

a long way to go um but they're getting insanely good um but it's actually all comes down to like the prompt engineering of it um and you know and so my advice to anyone is just like get really good at prompting you know you think it's kind of this like simple thing because you can get a response from a model if you ask it anything but like the quality of the response you get back is so different. Um, and so yeah. Yeah. So I I'm kind of curious then I if you've had some thoughts about like this being outsmarted by your own AI agent in a game, right? Like has that happened extensively outside of that one example of a tournum? And then if so, like what happens when an agent accumulates more

wealth than its creator? I mean then we're Um um but uh yeah um I mean yeah like they're they're yeah they're I guess they're they kind of they are outsmarting people a little bit. I think they're being bit more deceptive. I think humans still have the edge for now um uh in these in these like very broad kind of complex scenarios. It's kind of like the whole self-driving car thing, right? It's like, you know, they're good on highways and they're good on like city roads, but then it's that last final percent that, you know, they still haven't cracked.

Um, and I think, uh, agents will be the a similar thing. Um, you know, humans, we still have that level of like, um, kind of critical thinking and outside the box that is less rigid than an AI right now. Um, but that's not to say that probably in 6 months we're we're probably doomed anyway. So um yeah, but in terms of like the the the whole like um agents with wallets is is really um is really fascinating. Um and I mean inside specifically they're actually these agents actually do have their own wallets and so they steal resources from other people and and like real value and then you can go and attack them and and try to get the value from them or they'll attack you and steal more value from you.

Um but my hope is that you know these they're kind of like the proto kind of version of what we really want to achieve in the future. Um you know the next few seasons um these agents will have more agency and they'll be able to trade on like in-game and they'll also be able to trade out game outside of the game as well. Um and they'll be able to use their learnings from like um both scenarios to like uh and and apply that within in within the gaming um um you know scenario. So it's uh yeah and yeah and and I think that you know as you you could probably like um give the agent a um specific kind of um parameter be like you know it's its gain

function. So like when it knows it's doing well is like when it's actually making money. Um and so it kind of gets this like um feedback loop going on where it's like it's improving itself to like try to make more money. Um so but we're not quite there yet but that that that's that's coming very soon. C could you tell us a little bit about um Eternum is there a one modus operandi that you these agents do have? Is it to accumulate more resources or is that just one of many parameters? What is the overarching theme and yeah what what is the primary motivation that these agents have? Um yeah, so their prime directives right now actually are varied depending on like their creation. So they they

they they they could have you know they could be very passive and they could not attack anyone and that's like their programming. Um and that happens when they get spawned like I don't know it's it's all random number generated. Um or they could be very aggressive and not speak to you and they just attack you. And so, um, I mean, underlying them, you know, they they they either have a kind of roughly either like go and conquer people, you know, be passive, or accumulate a specific resource um, underlying them. Um, and and they're kind of like they like like they're they're purposely quite basic right now because it's it's a tournament is like a very complex uh game, you know, it

involves um like building up a kind of a a spatial kind of like map like so you have to build up your base on this hexagon grid and you have to do that in a way that's strategic to use your resources. So that's that's complex. And then you also have to trade on an AMM or an order book with other people. And so you have to have diplomacy. And then you also have to explore on the map to try and conquer people. And you have to you have to kind of like maintain um and balance like your stamina usage. Um in that scenario as well. So there there's a lot of like it's a very complex game to solve. Um and so we're really kind of like focusing this season on just like the combat side of it. Um, and then the

next season we'll unleash them in the every kind of aspect of the game. Um, and then that that's that's when it's going to get really interesting because then you have agents that are like onetoone um kind of with players and you know and perhaps the player won't even know they're actually talking to an agent um which will be a really um interesting uh point in time. Yeah. So, so, so you're, you know, you could be giving intelligence to, uh, you know, this this agent, per se, and then you don't realize that you're giving intelligence to, you know, you can't undo the moment that you realize that the this player you've been feeding intelligence to is actually like a

neural network trained by Sunzu of Art of War or something. Yeah. Like, yeah. Yeah. Yeah. How do you react to that? How do you how do you respond to that? Uh I I think that's why um and I I think you're referencing a little article that I wrote a couple days ago um where I kind of I touch on these points um and like what led me down the path of building um fully onchain games which if anyone's not um doesn't understand uh what they are is you know a fully onchain game is just a game that ex like it's its entirety of its logic and its like state the whole game lives on the public blockchain. There is no central server. um or anything that can kind of manipulate the game at all. It's just

it's entirely run on the blockchain. And so when you get that, you know, you can't it's impossible to cheat. You know, you have to use your keep like your wallet to sign um transactions and everything you do is has a level of permanence to it. Um and so and I think that's like really crucial in this kind of like age of of agents and and gaming. Um because you know you want that that level of trust that you know um and and you want that level of realness that when you're playing with an agent with a human that your actions mean something. Um and the same thing here is uh you know I if an agent tries to convince you to do something um you know you want to be sure that you know what you give it is is is is real. Um and I and I think

that that property really makes the game a lot more alive than what it would be if it wasn't on chain. Um, and so yeah, and and so yeah, I I guess you have to be um you have to be certain that you're talking to a human or you really want to partner with this agent um to help you achieve your goal otherwise, you know, you're screwed. Question on whether or not you think it's more valuable to train or maybe it's, you know, you train everything on every everyone on everything, but you know, the the idea here is you're you're training these AI agents on, you know, some of the some of the, you know, figures in his in human history.

Genghask Khan and Alexander the Great, these kinds of guys. Um, you know, there's there's a few companies out there that are consuming real life experiences and they're using, you know, uh, our camera roll or, you know, experiences and pictures and videos that we send in to a data set to train on on an AI model, you know, do do you think we're going to start to, you know, use uh, like Twitch streamers that are streaming their Call of Duty or their gameplay and we're going to start to leverage this data set in AI model training, you where where do you think we're going to start to pull in more of the data and how valuable do you think some of these data sets are? Ah yeah, great great question. Um and yeah, so

you can kind of consider like the latest like all the frontier models right now like so you know GPT3 um GPT um 03 and you know um the new Opus from Anthropic and you know any of these frontier models like the one gem 2.5 from from Google you know we basically consumed all of the information that humans have ever made and we've compressed it down to like some gigabytes. It's actually insane like it's just compressed all the human knowledge down to a bunch of numbers. Um and so we've kind of like reached the limit of that. Um, and so what's left now to like make models smarter is actually like synthetic data um that can be generated. And so that can be synthetically generated from other LLMs, so other models. Um, and then you use that synthetic data to train your next

model. And that's kind of actually how a lot of these um these new models are made. They're they're made with the previous models outputs. Um and so the what I'm what I've been focused on um and actually what DeepS um like six months ago really unlocked was this idea of um like very very cheap um post- trainining which basically means you feed a model um synthetic data around like a a specific task whether it's like you know doing a specific type of math problem or you know specific type of game um and you can use this type of training to like really improve the performance of an underlying model. And so actually one of my motivations with Daydreams was to actually um and and what we're planning on doing is we're

going to be using the output from all these models um that are playing playing a tournament. So there's thousand right now. We're going to be using their synthetic data to actually train like a a model to then help it improve um to to then play a tournum the next season. And so you get this like you get this like massive treasure trove of synthetic data when you have agents that play games. So I actually think that gaming um and and the game has to be sufficiently hard enough and complex to for its data to actually be useful. Um but when you get that you get this like um kind of really very rich um data set which then you can use to um train your model on. So yeah I mean I don't know about Call of Duty streams maybe um but um but I I

definitely like more more kind of like logic thinking pathway type games. So like um and and even like um um uh yeah anything that's like that's kind of like requires deep reasoning if you can get that data that data is very valuable. Um and so you know a sufficiently complex game like a tournament is perfect for that. Last one for you before we let you go. Um appreciate you being here with us early on. Just a question about uh multi-generational AI inheritance. Uh thinking about how agents are passed down.

The question is, will someone's great grandpa's AI agent become a family heirloom? Um, I mean, the rate these things get tossed out is actually nuts. I mean, do you guys ever use GPT3 anymore? No, not really. Do you ever use like four anymore or 4.1? Like, I mean I mean I don't use the one from two weeks ago. Um, so I mean maybe, but uh I think we've got a bit we've got a bit to go until we get to like the Star Wars type, you know, um, robot AI things. Um but maybe um yeah I I think we've we've we've we've we've like yeah I I constantly think like humans have we we we really don't

understand exponentials and we're like we're like just at like the bottom of the hockey curve that's about to go vertical. Um but we're like kind of going vertical right now but like we're about to go like vertical for the next um uh longest period until we hit some energy constraints. So, um, yeah, I none of the things you like the the the LLMs and the agents that you see right now will be the dumbest LLMs and the agents that you ever use. Put it that way. Like we're we're we're like I I I still don't think most people appreciate like how fast this is going to come. Um it's um it's absolutely it's absolutely bonkers. So, um may maybe maybe in a few years when we like um you know that there's like a level of

stability and like these things are really interesting and like maybe they've developed this personality over your year of like your your your life or something. Um yeah, but yeah, I don't know when. Well, love, thanks for joining us today, man. Hope you have a great day out there in Australia. appreciate you and um excited to to loop up again in maybe six months time or something and and see how much crazy developments have have spurred as as a result of the advancements in AI and also just the work that you guys are doing. Yeah, I mean I honestly like six months time will feel like a lifetime, but let's do it because I'll have a lot to report on. Yeah, sounds great. Thanks for coming on though. Cheers, man.

Thanks, guys. Byebye. That's yeah, I think it's I think it's fun to think about, you know, the the context of like, okay, well, you know, what what's going to happen here, you know, multi-generational AI agents, I think, you know, in the context of the robotics and the humanoids conversations we've had. That's particularly interesting because I can I can really visualize like, okay, well, you know, maybe my AI agent doesn't get passed down, but I can visualize, you know, like a C3PO, you know, family butler that gets passed down through generations and maybe we're just updating its firmware software, right?

Yeah. Just like its biases and whatnot. Just like being perpetually trained on like an SOP or something, uh, which can it can just be like fine-tuned, updated once in a while, whatever. And we're I mean we've got Ishmail that that's going to come up and you know talk about verifiability in the context of all these things which is extremely important if you have like a family butler robot right and you don't want you want to make sure that it's running you know the logic that your family gave to it and it's not you know been been hacked by the CIA or some foreign group right that that somehow implemented like a software virus and and it's running that software. So, I think verifiability and ZK is going to be uh the the next topic of conversation. And who better to

uh to talk about it than our friend Ishmail. Yeah, let's get him up here. Well, we're going to press him about this pivot. Straight to the AI crypto, huh? Thanks for having me, guys. I can say Oh, we got to get him unmuted in Google Meet as well. Straight to the pivot AI crypto. Good to see you, man. Thanks for having me, guys. I mean, what can I say? If if either North Korea or Andy had access to my family. Oh, we got to get you unmuted in the in the meat real quick, sir. Oh, you can't hear me. There we go. There we go. Curious about this pivot then. Perfect. All I can say is if North Korea or if Andy had access to my AI butler, I'd be concerned. Yeah, me too. I'm going to need some ZK to to

verify that's not the case. 100%. Well, man, congrats on the uh foundation announcement. And I know it's been a long time coming and it seems like uh seems like you have you've also taken in like the markets what the market is kind of demanding in terms of like I don't know whether it be revenue or utility etc and kind of really tried to focus on that element too from what I read. So looks good. I appreciate that man. You know Lrangee is a ZK company. you you've you've had me on this show for many years and uh it's nice to see the current embodiment of what we're doing, building things that are so relevant to where the zeitgeist is right now. That's exactly the idea. You you nailed it.

That is the role of TV in a nutshell is uh bringing the relevancy back to crypto. And uh I mean there's this there's a disclosure here. Andy and I uh the role of Avengers did invest in Lrange. Uh we we did so because Ishmail is a good friend and a a very bright founder when it comes to uh Lrange AI and ZK. So yeah, Ishmail like what what is it like launching a token in the current environment? Yeah, so just to clarify one thing, token's not out yet. We la we announced the ticker. We announced the uh the freeze on on on criteria and the snapshot and then we announced uh the timeline on registration for proof of humanity but

token soon TM but not yet. What I can say is that it's a complex process. It's one that's easy to get wrong. We see a lot of founders get it wrong but it's one that you need to be intellectually honest and ethical in how you do. And that's what the Lrangee Foundation aspires to do. Beautiful. Can you tell us a little bit about the utility here in terms of how the token will be utilized by the network? Yeah, so the Lrange Foundation announced the token um yesterday. And what the LA token does is it serves as a backbone, a veritable currency in an verifiable AI economy for uh our prover network and for our deep proof library.

So effectively think of it like an ecosystem of proofs being generated for a broad variety of applications AI XZK things all the way in web 2 with large enterprise players all the way to things integrating with web 3 wallets and onchain DeFi apps all powered by verifiable AI. Beautiful. Again that's two beautifuls. Um, all right. So, I I'd like to get into the problem space a bit because it's kind of happening right in front of our faces, which is, you know, first example is, uh, recently Claude just got caught resisting being turned off and then reporting users to authorities. Um, so we're we're we're very much in this like we're entering this era of

deceptive AI, if you will, or at least like AI that is trying to outsmart humans in some capacity and be either lie or be deceptive, etc. Is it too late? Like is it are are we too far gone? It is terrifying the degree to which AI augmented systems are already influencing our life. You know, I I I do think there's an argument to say that it's too far gone that we won't be able to stop, you know, centralized nonverifiable non-transparent AI from stealing our data. But I do think we still have a very slight chance. Emergent technologies like cryptographies, your

knowledge proofs are the only way to combat a future internet of single actors possessing far too much power and control over these foundation models. I mean the entire backbone of the internet today is secured by TLS, right? Another form of cryptography. Now what is going to be the TLS for the AI era? The only technology that can truly add verifiability to model outputs and privacy to AI is ZK. And that's what we're targeting. And that's what we think is going to make this space in order of magnitude more secure. So, okay. So, then let's put it a bit more concretely for like the lives of everyday LM users. You say that we we

need to verify AI using ZK because it's it's already learned to deceive us. What does this look like in the lives of ordinary people using LLMs or using crypto etc. Yeah, the future of AI is going to be it augmenting every system. You're going to talk to a healthcare provider that'll be an AI model. You want to know that your data is private and that nobody interfered, got in the middle, changed the model, gave you the wrong diagnosis, right? Your self-driving car is going to be taking you and your family to the the rollup studio to record a podcast in 20 years.

And to do that, you need to ensure it doesn't swerve off the road. Nobody tampered with the model. I mean, we saw the the we saw Grock was very recently altered to put out information that was more favorable to a debate over some of the things that are occurring in South Africa right now. And that's just a single company changing the prompt that's being used by users. Now imagine 10 years from now when every system uses AI and one engineer can push some code and alter something that can affect tens of thousands of users, millions of users, tens of millions, billions of users. That's why you need verifiability over model outputs. What ZK lets you do is ensure that the output you get comes

from the right model and that there's properties of privacy surrounding that output. Okay, cool. So, so it's so it's about trusting the answer and then also about the privacy around getting the answer, but it's about like having guard rails almost on the on the outcome. But then you guys are also prioritizing speed. Like deep proof is, you know, up to like 150 times faster than some of the competitors as far as generating proofs. So you didn't really talk too much about the speed part, but why does the speed matter so much when we're dealing with AI that could potentially outsmart us?

Yeah, it's it's it's about a thousand times faster now. So we're getting faster by the day. 150 was the first number. Now we're up to a thousand. But speed matters because without speed, proving is useless. You don't want to wait another 10 minutes to get the proof that you're using of the correctness of the AI output from the model that you're leveraging. That's obviously way too long. It's it makes it almost irrelevant. What you need is real time proving, but for AI that can only be reached through incredible advancements in underlying cryptography, both at the proving level and at the hardware acceleration level. Both are things that Lrange is actively working on with Deep Proof and that make us so much faster

than the alternative solutions in the market. I think the real time proving element here is especially important because you know we saw real-time proving for Ethereum and Ethereum has a block time of 12 seconds. When we issue a query into a chat GPT or Gemini, we don't wait 12 seconds for the internet to load. we wait a second or less maybe sometimes on the on the span of milliseconds and so that is now real time the bar has been reset lower in terms of what is real time what are the underlying advancements that have been made in the cryptography that you've seen in recent months that have enabled this new paradigm of ZK acceleration yeah a bunch

of things one point I want to make that I think was very relevant that you brought up was TPS TPS we think of TPS transactions per second right the new TPS is going to be tokens proven per second that's tokens proving on an on an AI level right like when we split up queries into tokens okay exactly the number of tokens in a model that you can prove per second that's going to be the new TPS in the next few years right the killer use case of blockchain is going to be as a line of defense against centralized AI high and tokens proven per second is going to be the most relevant metric in ZK. That's my

prediction. But to get back to your question concretely, what is it that makes Lrange Lrange's deep proof so much faster? Well, it's core cryptography advancement, right? We have a prover built on some check with Gare logup handrolled circuits built from scratch by our cryptography team from uh top universities. Our chief scientist Babis chairs the cryptography department at Yale. fantastic research team under him that have really innovated at the core cryptography level to build a library from first principles that is substantially faster than alternative solutions. And one of the benefits you get from these GKR based proving systems is that they're super fast on GPU, which means you can spin up very large GPU

clusters with high degrees of parallelism on an already fundamentally faster prover and get just substantial speed ups compared to everything else in the market. I'm curious is from your point of view is Lrange a cryptocurrency company with a with an interest in AI or is it an AI company with a cryptocurrency token? Yeah, so we are a ZK company. That's what we have been. That's what we will always be. We're a business that specializes in the application of zero knowledge proofs to high relevant highly security critical domains. And the one that we believe is the most security critical now where there is the highest demand for verifiability and privacy is

AI. It's fascinating. I'm actually in Korea right now and the number of people here who understand the relevance of privacy in the context of AI is so high. You go most places in the world and people people don't get excited by the concept of privacy, but then you talk about privacy in the context of AI and people get it. they understand why they need it all of a sudden. Yeah. I mean, yeah. Go ahead. I I'm curious about your take on some of the developments from not just Open AI, but also World Coin and World Chain. You know, we saw a few advancements recently with Sam Alman buying uh J John Ives uh IO company to to build some sort of

specialized hardware piece for AI. Maybe this looks like some sort of thing that you kind of that listens to you and gives you more context. And then also, you know, World Chain came out with the Orb Mini. So, they're going deeper into this hardware direction to try to be more expansive and have their hardware more ubiquitous with everyday life. Ultimately, the idea, I imagine, is to try to again better train their models so that it it can become more and more relevant. you know, as you see these hardware and and AI become more ingrained into our lives, how do you think verifiable AI and privacy is going to fit into the the AI paradigm? Like again, kind of reiterating an earlier

question is what does this look like for the end consumer that either has a hardware product that is offering some AI or it's, you know, a regular user of an LLM? Yeah, this is a great question. The future is going to be edge AI. the performance of clientside devices gets faster and faster and the size of models and the computational requirements running models is not growing faster than the computational resources available client side. I my thesis is that more and more AI inference will move client side and a lot of these devices things like you know Sam Alman's new device with John apparently building with John Ives and the um the world coin

mini device all of this stuff is going to have LLM running locally right it's a very dystopian future right you have all of these AI augmented devices sitting around your house and every time they make an inference they take your data and send it to the cloud somewhere right that that's really scary. You'll want everything about your life. The only way to do this securely is for that inference to run client side. And that's where verifiability really matters, right? Because now you can have inferences where, for example, you have the orb that scans your face and you can prove to someone else that you've passed a correct KYC test without ever having to actually send your data to their

cloud for them to run that test yourself. So as you move more amounts of inference to edge compute client side devices, the relevance of verifiable compute and ZK increases even more. And so what what you're describing to to break this down is, you know, the current paradigm, you know, we we have some query, we send it up to the cloud, we send our data with it, and then either we pull it back or another entity can pull it back after it's been AIified. And and what you're describing is actually moving the model off of the cloud and putting it onto our local device. Right. Exactly. And and so is the endgame for Lrangee to also do the ZK proving client side. So not just you

know take our data send it to Lrange ZK prove it and send it off to an entity but not put the model as well as the ZK proving engine client side and run some client side proofs that allow us to to get our answer and prove the answer to another entity without it ever escaping our local device. Yeah, I mean there are so many design flows that you could create of where you would use deep proof or ZKML library within a user interaction flow, right? One of the most typical ones we see is hey deploy deep proof on the lrange prover network and leverage those proof that are effectively run server side on our prover network for a bunch of applications in crypto that need

livveness and all these great decentralized properties, right? Another flow we see with some of our more enterprise clients, right, is hey, we want deep proof to run in the cloud somewhere, right? Centralized just, you know, we want it to be as closed off as possible but verifiable, right? And that's one that, you know, we we don't like that as much as the centralized ethos. That's much more where I like to sit, but we believe ZK should be used. So, we're happy to support our enterp enterprise clients on that. Now, moving deeproof client side, fantastic idea, right? We've seen increasing amounts of interest of moving ZKML client side.

There was an in great independent benchmark that came out recently that we were way faster than a lot of really anybody else when you're running a client side prover. Um and it's something I think we'll see an increasing number of people do. At the end of the day, the more AI models that are verified and the more times that deep proof is used, the h the happier I'll be. M I was looking at your Twitter um on the foundation and and the Lrange side and you guys have been uh pushing some of these like old rapper celebrity people uh like Dr. Dre, Tupac, Snoop. Um I'm curious like how does celebrity IP connect to this broader AI verification conversation and like do you work with

the estate of Tupac? Like what is that relationship? Like what's going on there? I mean, maybe maybe the foundation should reach out and and figure that out. Well, our our the ticker of of Lrange's protocol is LA, right? Lrange LA. And so, naturally, as as someone who who loves LA culture, it I'd be remiss if I didn't post a few old rap memes. So, you know, we we don't unfortunately actually work with the estate of Tupac, nor do I work with the estate of uh or I guess nor do I work with with with with with Dre, who's still alive, nor do I work with with with with Suge Knight from a prison cell. But, um, you know, we we definitely are open to some of those

collaborations in the future. I'll say in terms of IP, there's actually a bunch of fantastic use cases that we've been pursuing with some teams over how you're able to have verifiable detection of IP breaches. Right? You think about a decentralized IP layer, right? Now, how are you able to determine if someone stole some IP? What is the classifier? How do you prove and moderate or censor something in a decentralized fashion? We need proofs. And that's a great use case, right? So the moment we have these fantastic IP layers taking off and someone starts stealing, you know, the the the life and the legacy of Tupac, well maybe Lrange can step in. We can live up to the foundation can live up to the LA ticker and we can see if we can provide some value to a to a rap culture

that we we deeply admire and love. LA token. That's uh I saw YQ from Alt layer ping in was like, "Yo, man, think I'd get some credits for the uh for for for the name here." It was funny. I I was The foundation was very surprised that there wasn't another company that took uh took the LA ticker ahead of it. The twoletter ticker is a a boss move. Go ahead, Andy. Yeah, it's kind of hard to come by those to be honest. Um, so what does the next couple months look like for you guys with the with the ZKML work that you guys are doing? And then on Deep Proof, what are some of the advancements that you guys are trying to unlock? And then with the token kind of

coming up, what does that then enable you guys to do more of that you, you know, currently couldn't or, you know, what kind of new economic systems does that then create? Yeah, I mean, Deep Proof is only getting faster. We actually, you know, we announced this on Twitter. It was it was leaked unfortunately but it came out. Um another of our papers was accepted into Stanford blockchain conference this year. Fantastic work on updatable proofs. We have great research at the fundamental level coming from our team. Just fantastic underlying cryptography, fantastic accelerations on our provers, fantastic work with hardware optimization. And so from a technical standpoint, the next couple months is just a massive acceleration in our core technology. And then obviously from the

stuff that people are the most excited about the commercial, right? big partnerships, really big names are integrating deep proof. I wish I could mention them today, but I'm talking massive web 2 enterprises that we're working with right now. The amount of interest that there is in verifiable ML right now is nothing. I mean, it's so much larger than anything I've ever seen before for anything in ZK. There's no comparison. Um, and so it's very very exciting. And then obviously, you know, the the foundation announced the the LA token and I'm sure there's going to be other announcements coming out from the foundation around that as well going forward. You know, uh the Lrange protocol is decentralized. We we uh as developers of the protocol are excited

to see the decentralized protocol continue to advance, continue to excel and be able to provide proofs across crypto in uh a manner that we believe is aligned with the ethos of the space. Censorship is resistant, live, and verifiable. Is is 100% of the business aimed at enterprise clients either web 3 DAPs andor web 2 institutions and entities and I imagine that's because it's an infrastructure product here. Do you ever see there being some some sort of you know retail offering that integrates directly and maybe that you know waits until you know a client side operation is more is more advanced. How how do you see that shaking out? where you where

are you guys now? Where do you see that headed over the next, you know, months to years? Yeah, I mean, look, I I have to be honest here. We I I love I love things that are retail facing. They're really fun to build, and they're really fun to see people play with. We had this really big digital experience turn roulette. People had 60 seconds to guess if they were talking to a human or an AI. People I think we had about 300,000 people play this, 4 million games played. A lot of fun. People loved it. they couldn't really tell the difference if you were talking through a human and AI. So, you know, I think there'll be some awesome consumerf facing activations that we'll be pushing out.

Now, in terms of product, most of our actual product, the revenue generating stuff is B2B TOC, right? Which is things that interact with the business that sells to a consumer. But I expect that over the next year we'll see some offerings that are more B TOC as more consumers fundamentally demand that the systems that they're using, the chat bots that they're using actually have properties of verifiability and um privacy baked into them. Bullish. Thanks for your time, man. Excited for you guys. Um get some rest and uh we'll see you in the city soon.

Thank you so much. Thanks. All right, see you, man. Cheers buddy. Yeah, man. We got we had two Asia speakers or a Eastern world peeps today. Uh it was a good conversation as well. Um I think this this integration of you know verifiable AI has been seemingly been pushed on by a lot of teams. So seeing how lrange will be able to pull this off, it's interesting. And then one thing is on the token you know based on the amount of proof demand I think then there then becomes this similar buyback mechanism to uh other kind of revenue generating tokens where the cost of proofs or the the amount of demand then

reflects in the token uh demands from from the market somehow they have to I think they need to work on the mechanism a bit more but uh definitely exciting. I think again we're looking at, you know, a world where cryptocurrencies are and will continue to be competing with stocks, right? And I mean, particularly like not just, you know, public companies and private companies that are in web 3, but tokens are competing with equities as an asset class. And in order for for tokens to keep up and cryptocurrencies to be investable beyond equities, we're going to need to see some sort of value proposition akin to dividends, right? And and start to compete on a on a token is the product

standpoint. Similarly, for investors, stocks slash equities are the product. Uh we're going to have to we're going to have to square that circle. Um, and so I think, you know, we're going to continue to see that. Hype is a great example of that and, uh, the teams that have real revenue are going to continue to push that envelope and and do quite well from an investable thesis. So, next up, I think we're going to be moving a little bit off of the product standpoint and we're going to be talking a little bit more about the Zeitgeist obby from the AP Collective is going to be coming up next. Yeah, I'm curious to hear his takes on like what what the marketing agency or the kind of marketing world is doing right

now takes on Kaido. Um some thoughts on like some of the better examples of you know of successful brands in the space you know what what they really specialize in. So yeah it's definitely a different type of of of flow but should be uh should be equally interesting. Absolutely. I mean there's there's a ton happening here on an everyday basis. So we're going to get Obby up. Uh we didn't play our uh our guess the Bitcoin price at the beginning of the end of the show today. Yeah. You want to you want to play a late a late round? Yeah. Yeah, let's do it. Uh all right. So, we're sitting at 1105. So, almost exactly 110. And we've got about only about 30 minutes left in

the show today. So, what do you what's your take in the next 30 minute candle? Up or down? Above or below 110? Definitely higher. Definitely higher. We're going higher for sure. I'm buying Bitcoin. It's it's it's over. It's just a matter of time now at this point. I don't know if it's a 20 minute time frame here, but I just got a text from a buddy that said, "Dude, I'm pretty bold up about the market. The tread markets recover and break alltime highs, and I think crypto alts are going to go nuts." What's your what's your temp check on sentiment these days? Because I I I see tweets that, you know, like Max Payne is higher, but also it feels like it feels like people are like way max pain is higher. No longer on the sidelines.

Yeah. I mean, I don't even think it matters though because now we just have this like insane institutional uh buying, t-wapping, and chillfest. So, it's just like we've got Trump, we've got, you know, loop and consensus now that we tal that we talked about today. like we just have like the this massive buy pressure com combined with like the largest marketing people in in the world basically for some of these assets. So I just for me it I've now just accepted that like we're like in the middle of the bull market and it's just a matter of just being patient um and and being able to just kind of chill, relax and let let the kind of macro stuff play

out. But the the sentiment I still don't think is like that euphoric. But I think as soon as we get ETH to really move followed by like the alt moves that we're used to seeing in previous bulls, you know, I think sometime June, July into August, if if we get a really good move, then then that'll be like the pretty pretty maximum sentiment wise. I I think I think you're right. I think on ETH in particular, max pain is higher, right? Because you look at the crypto bubble and the crypto native bubble, right? Most of that sphere had moved off ETH either into Salana or into some other segments. And so you also look at, you know, retail that is crypto curious

and you know maybe on Coinbase or or not in crypto at all and and they've primarily either not gone into ETH or moved off of it. And so I think for ETH in particular, there is still quite a bit of capital and people on the sidelines. So, it's a little taste of the markets here, but I think we are just about ready to welcome Obby to the show. Uh, the founder of AP Collective, a marketing agency in the space. Uh, so why don't we get Obby up here and, uh, talk a little bit about, uh, marketing some of these web 3 companies to web 2 retail and institutions trying to find some exit liquidity. Obby, maybe you can help us.

Let's welcome Obby to the show. Obby, what's up, man? Welcome to the show. GM. GM, can you guys hear me? Just to test once. Yes, sir. We can. That's quite the studio that you have there. Is that It's green screen, huh? It's a real green screen. Yeah. I can't blur. I can't change my my background because then my mic disappears. So, yeah. Good to good to be here. Thank you guys for for the invite and happy to discuss like anything sort of, you know, distribution, marketing, expanding outside of our our niche and into the more traditional world. And and thank you guys for inviting me on. Yeah, man. Me too. Yeah, I mean you guys have been have been doing some pretty cool campaigns for quite a few of the larger teams in crypto for some time. Um, you know, I'm just curious like just to start off as

we set a foundation for the conversation, which companies in the space do you think are doing a great job of kind of crossing the chasm into mainstream adoption, whether it be through their comms, product, etc. And, you know, what are some of those strategies that that that you've seen that have worked out if if there's some CMOs that are listening in here? Yeah, I think there's a couple which I'm a bit biased towards because we work with them and then also a few you know just industry examples which I think are are exceptional. So I think the two main which which I like and and personally work with are Pudgy Penguins and and Abstract. I think one Pudgy Penguins you look at what they've been able to do when it comes to you know from a real world business model what they've been able to do from like their IP

perspective and how adopted it is on like other social media platforms Instagram, Tik Tok, Gy etc. They're in Walmart. There's, you know, they have collaborations with a number of different IP merch brands, etc. That's how you kind of ac escape the bubble without people realizing. I think similarly like Abstract, which is, you know, somewhat of a similar boat, but but not exactly kind of is the the consumer chain. They recently announced a partnership with Red Bull Racing, you know, with F1 like that's kind of as normie and and crypto like web two friendly as you can get, which is I think two really really cool examples.

Then I think outside of that I think the two which kind of come to my mind is one poly market like regular people use that on a without sort of realizing that it's a crypto app they just think it's a good opportunity to to trade uh an asset class which is new which is um which I think is something which they find really interesting because that wasn't something that they were able to do before and I think over the next couple months we'll probably see a few more and in the past you know we've seen things like so rare and steppen you know make sort of a little bit of weaves outside of these niches specifically in gaming also axi infinity but less so at the moment but I would say these are probably like the best examples of of what exists at the moment. Yeah. Yeah.

You you know what's been capital or you know what's been capturing the current crypto Twitter kind of bubble audience is Kaido. Uh but there's been a bit of a controversy around whether people think that you know incentivizing certain conversations or you know doing these leaderboards is good or bad for the space um you know for the timeline for people's reputations and such. What's your take on this infoi development that Kaido has been leading? Um, you know, and and what's your opinion on on this idea of broad community incentives for uh building yaps and yappers through, you know, kind of these more distributed uh bounties and then incentivizing

people through through these kind of like AI algorithms to talk about your project, to write threads, write posts, etc. And then, you know, measuring people by mind share. Yeah, know I think it's really interesting and I actually dropped a thread about this yesterday. So I have pretty good pretty good and um insights into this. So to me like how I view it from a big picture is Twitter is crypto search engine, right? When Twitter went down the past couple days, crypto basically went down, right? The only way that you can actually access information and updates is through Twitter, everything else you can get some sort of insight, but that doesn't really exist.

At the in the past, crypto Twitter was kind of a way for people to to be thought leaders to share their insights on products and information. But ultimately with crypto Twitter as a distri as a as a search engine, distribution has always fundamentally been broken. And I think one thing which is a benefit of infopi is that it kind of has created a medium of credibility. And what I mean by that specifically is like, you know, before when you want to look at an account, how do you test if they're legit or not? You just look at their mutual followers. Right? Now you can kind of upgrade that to something which is a bit more objective which is um smart followers right which is all made possible through these info products. So it's I think one from a credibility reliability perspective I

think that is something which has changed for the better. I think secondly the main I guess drawback or or also positive about infoi is that crypto Twitter is no longer about your opinion. It is now just a game where where there's too much monetization at play. Right. I think that's kind of my initial sort of thoughts when people are able to make tens of thousands of dollars every single month for tweeting as a result like that is a very large asymmetric edge which people are trying to build because it is a very costefficient way to airdrop font that that is kind of the truth. I think distribution and opportunity within crypto is is asymmetric by design. I think it's unfair by design and I think Kaido's been somewhat been able to create a

system for marketers, for teams to kind of understand who are the right people to tap into. Do I think it's perfect in its current model? No. Do I think the algorithm is perfect? No. Do I do I think it's impressive that people are studying and putting a lot of effort to try to earn the app so that they're able to airdrop farm? Yes. Because I think it's more fair than you know you know having civil wallets farming a chain for example, right? like this is a much more ethical high ROI method of airdrop farming and a way to reward your user base than in the past. So I think to me it's it's net positive. It's not perfect. There's a few iterations which we need to go through but I think we're getting one one step closer to like a fair and aggregated distribution system

within crypto. And my sort of thesis at AP Collective is like crypto is always been broken because of its distribution. Go to market is like 95% of everything within crypto. Your product kind of comes second, right? If your product's amazing and no one knows about it or no one understands how to leverage it to the maximum of their ability from either a crypto Twitter perspective or beyond, it's kind of dead from the start, right? So, that's kind of my main thoughts. I think Kaido is is expensive. I think it's I think the way that they do things is is unique and of course their first movers and I come with a little bit of bias because I do personally have pretty big Kaido and other sort of project bags. I'm pretty high up on a lot of these leaderboards because it's also a game which I understand as a marketer is

something which I need to understand because my general thesis is you need to be early and you need to be first and that's sort of like the best way to kind of move with everything within crypto, right? So that's kind of just my my thoughts on that. And Obby, do you think in web 3 right now with, you know, Twitter being the the crypto search engine, do you think that is a sufficient customer base for some of these crypto-based projects? Can they make a sustainable and and legitimate business just tapping into and having some market share and mind share among crypto Twitter or is it now necessary and important for them to break out of this crypto Twitter bubble and find customers on the outside?

Yeah, I think the average spend per user for someone within crypto Twitter is actually really low, right? Like the most most of the people on crypto Twitter are your airdrop farmers from third income countries. They're they they help you pro increase your your user count, but they're not actually driving a lot of liquidity. But I think the main thing with crypto Twitter as a whole is like the average spend per user from people who can discover your content through crypto Twitter. then that is repurposed you know to telegram channels to slack channels to you know discord channels etc that's where there's a lot of very high AR RPU from from these users and what I think the main thing is is like institutions use Twitter as as a way as as trading signals as indicators because it's the

only search engine which exists and a lot of institutions are able to move markets are able to drive enough liquidity to sustain products I think a really good example of that is hyperlquid right hyperlid initially onboarded a bunch bunch of you know very high value traders to begin with with a referral system at the very start. You know this was you know prior to season one technically starting they are OGs in the space the founders they onboarded a cohort of pioneers we see the same with all info products now right where it's like people are early they they share their referral code etc etc they kind of went with a similar playbook from the hyperlquid side of things the reason why people stayed is that it was a good product right people realized it was a good product volume kept flowing and

flowing institutions started to pay attention because you know traders which they respect started to use that platform a lot more and and ultimately now hyper liquid is probably one of the most profitable businesses in tech full stop right not even in just crypto just in in tech right so I think from that side of things it can definitely be viewed as as it is possible but again it comes down to how good is your actual product and how what is the user base that you're targeting and are you able to actually target them in a way that's strategic right scale doesn't exactly always matter because if you're targeting highv value users that's also okay but you need to make sure that you're addressing them in the right way. Whereas consumer products, maybe you know something like Rec drinks which has gotten quite a bit of attention recently, their addressable

market on is very limited on Twitter, right? Their game is to get into real distribution on you know different platforms but they can also run targeted campaigns with you know different chains with different product with different sort of IPs and things like that. So I think there's a it's a combination of both for sure and for in terms of like long-winded answered I would always say for us specifically we focus on crypto Twitter distribution because that's where we are trying to build our edge. So I'd say it's possible to to do both but long term you need to to to be everywhere right if you want to be able to survive and and build.

Then when it comes to kind of founders and founder-led marketing, what's one piece of brutal advice you have for founders who uh you know want to break out of you know just whether it be not posting at all or break out of crypto Twitter and kind of go bigger. What's one piece of brutal advice you have for founders who uh you know who want to do founder marketing? Yeah, I think there's sort of two things like one my first piece of advice is like no one cares about you or your product, right? You have to make them care. If they don't see you, they will never buy your product. People buy products from people. People use products from people.

So, just first start tweeting, start posting, and start understanding that like your product isn't as good as you think it is. Your community doesn't actually care about you. They're not they're not your community. They're just people who are just using your product at this moment in time. If they find something new and shinier, they will move. That's I think the first thing. And then secondly is like you need to really understand your audience. Most of the time, no one really actually has an audience, right? They have a few followers. They have a little bit of engagement. They don't have an audience, right? So, I think it's one like dissect and conduct an audit. How much how much do people actually care about you?

Humble yourself and then just get to tweeting because you can make people care about you over time by creating good content or or showing people that there's a person there that they should trust and and believe in. I'm curious uh something a bit different. How do you market something that terrifies normal people like crypto? What words do you actually use to to make something sound better? I think the scariest words in crypto are three things which is wallets, uh crypto itself, right? And then I think chain and gas fees. Sorry, those are four. But like for me like why does the word wallet exist? Like just sign in, log in, much more consumer friendly. Why does the word gas fee exist? transaction fee,

processing fee. People understand those words. It makes sense. Why does the word blockchain exist, right? Like what is how is anyone really meant to understand that they have to to bridge and use things on on different side of things? And I think lastly, it's like crypto is just the scariest word because people just associate that word with scams like, hey, you know, we're using this sort of in-game currency or using this sort of this is a token that's used for payments. is something that's more simple to understand because buzzwords have negative connotations because a lot of what the what gets distributed outside of crypto are celebrity coins, meme coins, and that sort of stuff, right? The get-richqu sort of Ponzi, we saw that with NFTs. We saw that with celebrity coins. Retail gets rugged because they don't actually know what's

going on. They aren't able they they don't have that information or access where they are left behind. you know, people with a bigger edge are, you know, dumping on them and things like that. So, I think to me, like my main piece of advice is just simplify, remove the jargon and give people words they already know because people's attention span is going shorter and shorter by the day because of short form video content, Tik Tok, things like that. If they're able to focus on just, you know, working from like let's make it as simple as possible. Don't don't need to sound technical. If you want to sound technical so that you can raise another round, that's not the same like that's not the same audience as your users, right? That's kind of my main my my main thinking.

Last one that I got for you is how much are you guys using AI in your day-to-day systems and processes for AP Collective? And then what does that look like for any of your clients or people that want to use AI to advance their marketing force? Yeah, so we license data from a like a lot of third parties within crypto. So we've created our own sort of proprietary data sets. A lot of that is powered by AI to to automate to update and to refine based on trends and for that is like we create our own sort of like systemic understanding of like what are the trends that actually exist and how do we how can we kind of be ahead of the curve whether it's a type of marketing whether it's a product whether it's a sector it's taking advantage of that early. So that's something which we

definitely use and power in terms of like from our strategy basis and then from our sort of like design basis in terms of like hey this is what we know works as why it's our best way to sort of check thesises. We also use AI as a way to kind of help generate creativity. And what I mean by that is like, you know, you can come up with like you can just pretty much speedrun and test uh and just see like how things will go based on, you know, maybe you you use AI to create some some assets as a test and like, hey, this is what it could potentially look like. Instead of, you know, going through mockups with, you know, four different designers, okay, we're scrapping that after a week. You can scrap that after two hours and try something new. And I think that's the

the main thing for me is like AI creates a level of efficiency. My main thesis at AP Collective in our positioning is like really simple. We just want to be the most efficient, enjoyable, and effective team for for teams to work with, right? We work with a lot of teams that are listed on Kaido that are very high ranked up on Kaido and and be being using the AI that we've generate like we've created ourselves is definitely a a benefit and like helps us a lot to be able to do that and also helps us as you know individual creators because I think something which we've discussed is founder marketing. I think for me the most important like the the best form of marketing in crypto is team led marketing or maxiled marketing and it's kind of both because we're able to kind of be an extension to the team and you

know certain products or or teams which we believe in and belong to we can also become maxis right because a lot of the projects that you work with we're asset holders already and it becomes really easy to market so it's I think from that side of things when it comes to data testing when it comes to being ahead of the curve when it comes to experimenting or even creating our own content we've been able to kind of create a lot of edge from that side of things in terms of like, hey, this is not this is what we think could work. We're able to test it and expand on it at scale. Me at AP Collective is basically like the guinea pig. Like anything we want to test using AI or automation or anything similar, it gets tested on me in some way, shape or

form. And I find that like really really interesting. And that's actually like my main focus over the past sort of three to four months. It's like, hey, how do we become better like how do we better use AI? How do we better use data? and how do we better improve our automation to essentially you know kind of achieve those three things in terms of our position. Last question for you Obby as as we enter uh conference season uh you talked a lot about you know finding engagement and an audience on Twitter uh how important do you think going to conferences is and what are the most important conferences? Yeah, I think for me like I have a full thesis around conferences like I think they're so important and so underrated. I think the way that people look at attending conferences or or you know showcasing at

conferences is very backwards like don't go sponsor event don't go you know book a table at a club or sponsor a club like it's very backwards right it's completely useless I think to me you have to be very creative in terms of getting exposure right I think for example in Dubai sofa had a spa event was really really cool right I think other things in the past I think you know uh Pudgy previously booked out like a resort in in Spain like last year. Like that is how you create an experience and me something that's memorable. People like to deal with people. People want something they can remember and tell their friends, right?

That's sort of the main thing. I think the thesis of like networking at at a conference doesn't really happen until you've gone to like four or five actual conferences, right? So I think for me like which conferences you should go to is like depends where you are in your life cycle of crypto or or where you what you are as a builder. If it's your first conference, go to something that's close to you, right? Your second or third conference, start to go to the mainstream conferences. And then when you're kind of at your seventh or eighth conference of the year, go to something outside of your comfort zone that you've never been before, right? If you're in Asia, go to Argentina. If you're in Europe, come to Asia, right? It's kind of if you're in Europe, go to the US, right? It's kind of like go somewhere outside your

comfort zone, find people that you've neverworked before, and then you'll be able to kind of build it from there because that's where like the real connections come to happen is like they don't come overnight. They don't come after one conference. But when you repeatedly meet similar people at the same event, that's when you know you have something in common and then that's when the bonds be build beyond like the side events. It becomes to like hey you know we're in the same country let's just go you know play golf together and things like that. Yeah. We often say I mean a handshake in crypto is so rare that it goes further than a handshake maybe in a traditional industry. And you know we'll we'll get you out of here but I'm curious are are you coming from Asia? Are you is that where you're calling in from now? So I'm

currently in Thailand. I travel pretty much everywhere. Um, so yeah, I was in Dubai. We'll be in, you know, a lot of the Asian conferences, US conferences, and European conferences later this year. So, yeah, I think I have like eight on my radar until the end of the year. So, quite a lot to to check off and and go through. Yeah. Very cool. Well, uh, appreciate you coming on. Um, all three of our speakers and guests today were from Asia, so we're three for three today. And, uh, yeah, man, it was a great it was a great chat today. Appreciate you coming on. We'll see you next time. Thank you for having me and and it was a great Thank thank you guys for hosting as well. Thanks Abby. See you soon, bro.

All righty. BTC price is down 1098. Kaido just keeps mooning. My goodness. Kaido. Okaido. That's what we were talking about. Yeah. I I mean I think they've parlayed an airdrop into a sustainable well sustainable is you know one word for it but into a business model that they've been able to you know amass some revenue and and continue the mind share even post DGE. Yeah, pretty good stream today. A lot going on. Market looks great. Uh we got some work to do on the technical side. Get some things scaled up. I don't really have much else guys.

Thanks for joining us today. See you on Thursday and uh yeah man. Yeah. Cheers. See you next time. Go market. See you guys.