LIVE from NEARCON Day 2 In SF

Transcript

Heat. Heat. Heat. Heat. Heat. Heat.

Heat. Hey, heat. Hey, heat. Heat. Heat. N. Heat. Heat. N. Heat. Hey, heat. Hey, heat.

Heat. Heat. N. Heat. Hey, heat. Hey, heat.

Heat. Heat. Oh yeah. down.

Down the back off the back. H the hopped up the

JP Morgan, Black Rockck, DTCC, Fidelity, the entire thing was just institutions. It's just it's next level. Well, this industry is going to the next level. And guys, I I don't know what else to say. I'm bullish. He said it. He finally said it. So bullish. And he's bullish. I'm bullish. We're bullish. I'm about the I'm above the back.

Ble. above the mother. The rollup is headed to the top. If we're not already at the top, we're going to the top. We're on fourth floor today. We're going to be on floor one

and then we're headed straight to the moon boys. >> Guys, welcome back to day two of Near Con. We are live at the Fort Mason Center in San Francisco bringing you the leaders of AI at the convergence of AI crypto innovators, investors, venture capitalists builders entrepreneurs and more. Today hosted kind of covering the main topics, privacy, confidentiality, intents, the state of agentic commerce, workflows, and more. Uh, yesterday was a doozy. We did six and a half hours. My voice is still intact. If you heard me in Denver, it's actually quite better than what it was then. Rob, how you doing today?

>> I'm doing great. That was a fantastic intro. Thanks for getting us kicked off here. Uh we're right across from Alcatraz. We have a view of the Golden Gate Bridge and we are in the heart of AI innovation, which is incredible. We had a bunch of great conversations yesterday. Uh, and you're right. We're going to be talking more about intense agentic economy, uh, AI commerce, uh, AI agents, the, uh, energy super intelligent crisis that we have at our fingertips. >> So, let's talk about it. We've got our first guest.

>> Yes, we've got Mike Burcell, executive chairman of Confidential Compute Consortium, >> the executive director of the Confive, >> and I'm here for six and a half hours, right? Yes. >> That's it for six and a half hours today talking about confidential AI. >> Sounds good to me. >> Let's do it. >> So, let's start off with some fun here. Okay. >> Um, if an agent closes your deals and manages your team, are you still the boss or just the AI's legal representative?

>> Oh, okay. The word here is agent because agent is is really important about what we mean by an agent. So, if you think of a a theatrical agent, right, or you know, you're an actor, you're I don't know, you're Denzel Washington or something, you look just like him. Um, you have an agent, right? And he he he or she is uh legally able to act on your behalf, right? Yes. So, if we get it right, and there's a that's a really really big if, right? Um, you kind of he's acting the agent is acting for you. And that's kind of where we want to get to with Agentic AI. And I was at a a conference yesterday on human agency and how we can kind of look at that from the technical and from the legal and from the infrastructure points of view to

make that all fit together because if we if we don't do that and we have agents just doing stuff, you know, all over the place, it makes no sense, right? And the banks will get upset and the governments will get upset and my mom will get upset because I won't send the right flowers or whatever for Mothering Sunday. >> Yeah. >> Yeah. Um and confidential AI is going to be a really important piece of this. Privacy is something that Near is working on. you guys are working on as well. Just tell us about why confidentiality is so crucial in this era of digital twins. >> Yeah, sure. So, confidentiality is only a part of it. So, well, let me rephrase that. Privacy is only a part of it. So, confidentiality and privacy kind of go together, but they they don't mean the same thing. So, what what the

confidential computing consortium is, we're part of the Linux Foundation, and we are kind of like the industry body, but the open-source side of confidential computing. So that is chips which basically allow you to run your applications your work your workloads your AI agents specifically here in a protected isolated way so that even if I know it's running on on a cloud in in AWS or on Google or on Azure or whoever it may be or even on your mate's computer they can't look into your agent and they can't mess with it. So all that stuff about you know let's say I have we have sorted out this thing about agency and that you know my my agent is going to do what I want it to do. If someone can can look into that agent and change

it and say let's buy 10,000 uh shares of Azure rather than two shares of Azure that's a bad thing right I mean no no disrespect to Azure right but I I I want to be in control. So being able to uh to be able to control what goes in and out and that the integrity is is maintained. So not only that people can't look in and say, "Oh, I see how much what Mike's bank balances or Rob's or Andy's bank balance is probably much bigger than mine because you're media stars, right?" But not only do we we not want that, but also we don't want them to be able to change it. Unless of course they're giving you all of more money, in which case I might be more interested. But there's a there's another thing. So I'm going to just keep going because once

I'm going I'm just just going like which is at a station and that is to be able to prove that this is done. It's all very well for me to say set up an agent say okay this looks good but how can I be sure how can I be sure that what's running in that agent is the code that I want to be running right and uh that it's set up with AI we care that it had the right data sets that it came from in originally the right models right we want to be able to check all the way back check the provenence of that and check that that is what's running and there's something called atastation which gives you cryptographic really really good strong technical proof that that is what's running and that is kind of the gold standard for where we where

we are and where we need to be. >> Yeah. So, you know, you bring up a good point with the attestations, right? Because there was actually recently a shadow a shadow model accusations between Anthropic and Miniax where Anthropic accused the Chinese firm uh Miniax of redirecting traffic uh from of capturing Claude's capabilities basically forking their code putting it into Miniax and uh had 16 16 million exchanges were used to scrape this model's logic um and bring it into the Miniax uh kind of like UX and then cut the cost by a ton.

>> As a security guy, I kind of I love that, but I wouldn't want it to happen to me, right? I love I love what they did. >> Also geopolitical, right? It's China versus the US as well. So, how can we protect this kind of like model IP where a model's personality can be distilled and redirected within 24 hours of release? >> Yeah, sure. So, let's come back to sovereignty because that's another really big and interesting one. But, so there's there's kind of two different ways of of doing this. So one is um to say we don't want anyone to be able to look directly inside the model right when it's running and hopefully we can stop that. But I think probably what happened for at least some of that attack was it wasn't actually looking inside. is just asking enough questions to get the information outright. And

that's a bigger and so much more difficult problem because you need to sort of uh it's like saying to I know a Navy midshipman that um you can answer any question that comes to you, but if you answer too many of the right types of questions, you're giving away information. How you expect that Navy midshipman to to get that? you know, an admiral might be able to put all the pieces together, but the Navy and the midshipman doesn't have that sort of amount of, you know, context. And yeah, we're in Fort Mason, so I'm thinking navy here, right? But the thing it's it's difficult to stop exfiltration of information without sort of specific guard rails which are aimed at at protecting that and without knowing what you're protecting, it's really

difficult. There's a classic World War II example when uh they were worried that um if they were very careful about stopping um people saying where they were going to be deployed, right? But if you see that uh 20,000 um cold weather gear and sets of skis are being sent to this particular port, you can have a pretty good guess, right, that they're going to Scandinavia or the Arctic or or wherever it may be, right? That sort of leakage is really difficult to to control. Really, really hard.

>> And we actually have a very similar thing happening here just in the last couple weeks where the CIA and Pentagon have this pizza tracker. And if there's more pizzas being delivered to the Pentagon, it's more likely that we're going to go send some bombs to Iran or something like this, right? And so this is again like you're saying, we can have almost these predictive indicators to indicate where things are going, what is happening at any given time. And it's almost impossible. the people need to eat, right? Like you need whether it's pizza or something else. Uh and so if there's more things happening in a close location, people will pick up on the hing.

>> We call that side channel leakage or just information leakage because it's not directly what you're you're looking for, but you get information that is correlated with it in some way. And if you have enough of this, you can start making guesses that, as you say, if someone's been talking about bombing someone, then maybe the time is now. Those sorts of things. Yeah. And how much should we be protecting some of these foundational models? Obviously, it is a question of national security at this point, but we are in this open-source environment as well. How should we balance closed source censorship protecting these foundational models from any sort of attacker whether they're just pulling it or directly trying to get access to the underlying foundation versus look this is free, code is law, open and everything for

everyone. Where do you personally stand on this? >> It's it's really interesting. First of all, I should mention I'm not a US citizen, so you know, you got to careful what you leak to me, right? Um, getting a decent cup of of tea is very difficult in the US, but I can sometimes manage it. And that that leakage I can manage. But I am a big fan of of open source uh for everything we can. But it's so I'm a I'm a a crypto guy, by which I mean cryptography. I don't mean cryptocurrency. I mean, they're related, right? But my background is with that. And the deal there we always say is that you want the protocol the underlying protocol to be open but you want the keys to be secret. Right? So in this context it's a bit like saying well you want the uh the model itself to be as

open as possible but the data you're putting into it and the training may be proprietary. So at what point do you say where do we stop? Where do we stop being open? I'd love everything to be open but that's not the way the world works. It's not way world's working now. And we need people need to be able to make money of this. That's entirely fair. So I think that there's a there's a danger where people say this is open source. They don't mean all of it. And understanding what is appropriate um is really important. But again from a security point of view, I love open source because the more you can look at the happier are that there's no bad actors, there's no malicious work, there's no holes in it.

>> Yeah. >> There's no vulnerabilities. Something we worry about a lot. So the other thing that you you you touched a bit on there was was the sovereignty issue, right? And that's becoming a big big big issue because um and it's we mustn't think just nation states although it's absolutely nation states as well, but if you are a a large company, a large financial company, you know, name any of the big banks or or pretty much any of the big software companies or whatever. Um you don't just operate in one country, right? you operate in different countries and you will probably have different um divisions who can't look at each other's data on occasion, right? Uh the classic is you don't want your system administrator being able to see, let alone change your CEO's salary,

right? So you want to be able to run your applications, your agents in places that you don't necessarily have full control over, right? Because if they can't, then they can't do everything they need. That is where we need controls and confidential computing is definitely one of them where you can run stuff and you can have really strong assurances that even if it's running on your computer right in front of you, you can't look in, you can't exfiltrate that data, you can't mess with the answers it's giving uh or or see what questions are being asked of it.

>> Yeah, I'd love to learn a little bit more about the confidential AI consortium like what are the benefits of company that it joins it? You know, what kind of protections do they have? >> So, we're not the confidential AI consortium. confidential computing consortium. No, that's fine. But AI is absolutely one of our big focuses at the moment. No surprise at all. Yeah. So, we are anyone can join it is what it comes down to. Um, you need to be part of the Linux Foundation which is doesn't just do Linux, right? But Linux is the first thing, but it does things like Kubernetes and a lot of open source and embedded pretty much anything open source these days. Um, open source for movie making, all of that. It's really cool and games too. But um we have all of the big silicon manufacturers, but we

have also all of the big uh uh uh operating system manufacturers, the big clouds, but all the way down to to startups, you know, one or two people, we have academic uh institutions, we've got enterprises, we've got financial institutions. And what we do is we a promote this technology but also work um to to understand what it means for people work with on on the technical protocols get them into standards. And we've just started a special special interest group for um talking to regulators because if regulators don't know about this, they can't tell people they should be using it. And that's we know the best way of getting things taken up for these big companies who

will say, "Well, we're not sure about it." If the regulator tells them, they're gonna start using it. Right. >> So, I'm curious if you guys have have faced this situation or if you have if you've seen anything around this where you have, you know, a autonomous agent or, you know, we're getting to this autonomous world. Um, and they go and and the agent does something on behalf of the CEO of a company or of a company at large. as you mentioned perhaps there's a company in the consortium uh who would need to actually get some privacy controls embedded and uh they cause real world harm they for example there was a woman who worked at open AI and this uh openclaw you know erased half of her inbox she unplugged the Mac mini uh but even even more so right like

they get a prompt injection from a other agent and they come in and they've um you know sent out a really obnoxious email uh to the entire company uh that leaked publicly Um should should this >> we should be worrying about this and this is stuff that's happened. There's another good example of a uh a guy who's runs open an open source project. Yeah. And uh a bot put in a bug report and a patch and the policy of this open source project is they don't accept AI uh patches. They have to be come from a human. And so he rejected it. The AI wrote a blog dissing him.

>> Wow. uh and and and took out a whole bunch of private information and put it on the blog and basically attacked him. So, we're seeing this sort of stuff already, right? And so, we absolutely need to worry about it. >> But like, who should be in who would the human or company be liable for the criminal or civil charges that >> I am so not a lawyer and I'm so pleased I'm not a lawyer? But these that is that is the question. That's exactly the same. But it's the same agent goes and does something, you know, insane online, like am I supposed to be liable for that?

>> Is that different to I am Whimo and my uh car hits someone and breaks? I And I don't know. >> Or if I'm Robin Storm and I make tornado cash and someone uses tornado cash for some nefarious purposes, >> I don't know. And and these these are absolutely the right questions we need to be asking as humans as we try and set this up. And and what's more, the answers aren't always obvious or even consistent. So, for instance, a classic one is autonomous cars. If you have a uh you have cars going on and it can't stop in time and it has either take an old person or a young toddler, which should it do, right? Either >> and well, trying to hit both. No, probably not. But in in Western

societies, most people would generally say the old person because they've lived their lives and that's what it is, right? In some eastern and Asian societies they say hit the youngster because the old person has lots of uh wisdom that they can give and the young person has not set out. So we can't even assume that the answer is going to be same everywhere. >> Um and so we need to understand that law and uh and ethics are contextual. Right. >> Of course. >> But I can't help but think so many like the anthropic and minimax thing that we spoke about earlier is geopolitical. uh the US and India just signed a paxilica treaty you know bringing out uh production from China and Beijing trying

to get more hardware in India rather than in China $15 billion Google investment into into uh actual uh creation of data centers there in India um like AI is creating a cold war it's creating a new or or it's exacerbating the existing uh geopolitical dynamics I don't know if you have an opin on this >> uh I think that all technologies which are are gamechanging and let's face it AI is gamechanging have significant geopolitical impact and I think that the problem one of the problems you have is that things move so fast these days. uh two or two or three years ago I thought oh well and I'm in my 50s by the time AI

takes off I'll be retired I don't need to worry about it and how wrong was I right because things move so much faster and that that is the problem that in the same way that you know when we had uh you know microscond trading that changed the stock market right having AI getting involved well in that let alone all the other stuff is going to change how everything works and we need to get our heads around that and it's one of the nice things about this conference is that uh here at Neon, we're talking about this and we're bringing the people who are who are thinking about it and and designing these systems and and working out how they interact with the real world. And I I've got a panel exactly on the real world later on today and and getting them together because if

we don't talk about it, we're going to get it wrong. We'll probably get it wrong anyway, but let's do our best. Right. >> As we kind of close out here, just give us a bit of a sense of the the panel that you're going to have later today because we're going to be here in this room for the next six and a half hours. We might miss the panel. >> Yeah. Yeah. >> So, so tell us a little bit about what you're going to be talking about on stage. >> We're going to be talking about where how is AI making a real difference right now and uh how does human is human judgment still important? And if so, how uh should we be worried? Is there hope? And it's kind of a hope or not hope. And I'm I'm hoping that we're going to move

towards the hopeful rather than the hopeless. I think I think there's there's good chance we got some bright people uh here, right? And uh we'll we'll try and work together and do we can. So, lots of lots of questions. I wish I had more answers, but we we need to we need to do this thinking now. So, it's great that it's happening here. >> Yeah. >> Sounds like you're scared of AI, Mike. >> Uh I'm scared of of cars. I'm scared of books. I'm scared of all technology is scary, right? Um so, so yes. But am I excited about it at the same time? Yeah. My two daughters are 18 and 21 and they are scared that if and they're at university, right? They're scared that if they touch AI, they're going to be disciplined because their universities are going to say it's not your own work,

right? And that's the sort of stuff we need to work past, right? AI has got to be a tool. If we use it right, >> hopefully we're not tools for it. We don't want to be the batteries for for AI robot overlords, right? We don't want matrix. But if we get it right, we should be able to use this as tools and make the most of it. And that's what we need to aim for, I think. >> Yeah. I always like to think of LLM as calculators. And you know going through school sometimes they let you use calculators, sometimes they don't. Yeah. Uh but it's important that we kind of use it and recognize it for what it is and not use it. >> We need to get you explaining that there's a lot more calculators. But if you want to use it as a calculator,

that's just fine. That's fine. >> Absolutely. >> No, no, you mean allow tools. It's a really good example. I I like that that that there are times to use it and there are times not to. And that's what we need to learn. That's what we need to teach people. >> Yeah, absolutely. Mike, thanks so much for coming on, man. >> Thanks for having me. Lots of fun. Really appreciate it. Thanks a lot. Cheers. Thank you. Thanks, Mike. All righty, guys. Uh, we're gonna keep this rolling here today. We've got a ton ton of great guests. Uh, I've got some things to bring up with with Rob, whether he's ready or not. >> Um, >> let him rip. >> So, the February duel has peaked. The February duel >> open >> has just released G GBT 5.3 codeex which was >> I think within the last couple weeks

>> featuring Frontier which is a specialized system for managing swarms of AI workers. >> Yep. I think they also released uh EVM uh bench which was the ability to audit some smart contracts. >> Yeah, let's talk about that as well. I mean obviously Anthropic came out with with Opus 4.6 several weeks ago. >> Yep. >> Boasting a 1.2 2 or 1 million token active reasoning window. >> Yeah. >> Uh I'm not even sure what that means. >> Well, the active reasoning window is how many tokens they can go back and look at and remember with the context that they have.

>> And so that you're going to think you you can picture context window, right? Before it starts to have the entropy of knowledge, before it starts to lose kind of a sense of what you're talking about, you have to start a new >> it's an infinite >> and the way you would measure that would be uh in tokens. So uh released a few weeks ago but dominating today's session 1 million token reasoning window first opusclass model to allow users to drop a thousandpage technical manuals or entire code repositories into a single prompt without losing a shred of detail. Um so that's kind of like the you know context windows are bigger than people's memory now.

>> Yeah. >> And so what do you what do you what's your reaction when you hear something like that? Well, I've just actually noticed it. Uh I have like a rollup ops uh window in in Claude >> and um we you know me and Claude have a really good relationship >> and you just continuously use the same window over and over. You don't need to change a new chat and it's very good and it's built. >> Do you have you noticed that it loses any sort of like >> does it forget like who some of our employees are, what their role is or like >> it kind of forgets like what we've done?

>> Okay. be like it'll be like oh like do you remember when we did this and it be like yes and then it's like okay then let's go back to that and then it mis construs what I want to go back to sometimes >> um so I found that interesting what is the latest on EVM bench >> open recently announced yep there's a collaboration between paradigm and open AAI uh and the ability to I believe it's just to audit smart contracts but I'd imagine that this is going to quickly descend upon like writing their own smart contract.

>> Yes. >> And once we have sovereign AI agent like AI is already writing code, but it's not like >> using cloud code but it's not specifically optimized, right? Because we have a lot of very general tools that are generally very good at a lot of different things. But once we have uh specific purpose training, not general purpose training for some of these models, these are going to get a lot better because it will have uh context only with the things that it finds important within that particular domain. Your LLM has the context of the entire world right alongside rollup ops. Yeah.

If we have something specifically trained on smart contract and it doesn't care about specs and rust and javascript and typescript and all these different things, all it has is solidity and it know and a context of what it wants to get done. It can be much much more specific in the things that it's achieving. >> Yeah. So basically it can find a AI bug in a massive codebase. Can it fix the bug bug w without breaking the rest of the contract? Um that's kind of that's kind of the hard part right now is actually solving the bugs and then um not breaking the entire smart contract. Uh and then you know can can it actually drain funds in a sandbox moment, right? Like can it literally

pull funds in a white hacker white hacker kind of you know setup and and drain funds successfully? >> Yeah. Yeah. So, I think we're going to end up with uh agents deploying their own smart contracts. Uh and so, um obviously it's not going to be perfect at finding these bugs or solving these bugs or uh or really writing the contracts from scratch, but there's not going to be a human in the loop once agents are writing their own smart contracts. Uh these things might be open source. Uh so they could be verifiable once they are deployed. But we're going to enter a reality where we're going to have some AIdriven businesses. Most of these AI agents are deploying the code themselves. Uh there could be sovereign AIA has talked about uh AI that is born

out of smart contracts. So you could have a smart contract defined as deploying a new agent every time some you know let's say you deposit another 100 USDC into this smart contract. every time it hits that number uh it deploys a new agent and then this agent goes and it has a mandate and it's deploying new things deploying new smart contracts writing code deploying that code uh potentially with no one there to review it before it pushes live which means that we could end up with protocols and DAPs that are live in production deployed by smart contract agents that no human has ever touched and we wouldn't know the difference whether or not a human deployed it or had anything to do with it. It might just look like a place to deploy our money for some nice yield.

>> Yeah. Yeah. >> And it just goes to show that we need to do our own research. Um really once that reality happens, I think our agents are going to need to be the verifiers, right? So we may have a smart contract AI deploy its own protocol, but you and I are also going to have agents ourselves. And those agents, you our agents should be responsible for actually doing the verification of any protocol that we're deploying our money into. Right. Technically like the whole point of like blockchains and self-custody and all this is we should be like verify don't trust right like we should be reading the smart contracts and verify we don't have time for that even when it pops up on our hardware wallets like we don't really know what exactly we don't really know what it's

saying we should have an agent that is deployed to verify these smart contracts >> because you don't want that money ring you don't want that money r you do not want to get that guys we got Alex Shvchenko coming up next. Uh he just gave his keynote to talk on stage to open up day two of Nearcon. Shout out to Near with their big announcements yesterday. They they rolled out Near.com >> which was their all-in-one tracker of their uh of of their intense products of your portfolio.

>> But also Ironclaw. So like you go to go to.com and it lets you activate your agent. That's the real power here is that you can decide between openclaw and ironclaw. Imagine you, you know, probably use ironclaw cuz it's more secure. trying to talk about the unified multi-chain hub powered by Near Intense first. Then we've got ironclaw. By the way, the near intense fee switch is now live as of yesterday, meaning they've act activated a revenue sharing model with distribution partners as well with 13 billion in volume growing. The switch ensures intense usage translates into protocol revenue and increases value capture from near. And then of course we have our lovely ironclaw product.

Ironclaw you can there's oneclick setup. You pay uh $0 a month, $20 a month or $200 a month based on your API usage and you get a much more secure and private implementation of openclaw. My Mac Mini is actually sitting at home right now and uh it's not plugged in. So, unfortunately, Chef Clanker is off lot. Uh, which means that I am not getting the most of my, uh, Claude API credits right now that I'm spending that I'm using. And so, I think when I get back home, I'm going to fire up Chef Clanker on on ironclaw, give it full access, >> full access, >> and just I mean, we had we had

>> just see what happens. quite a few people in here yesterday talking about like what we can do with these agents, what you know, we had Greg talk about things all the way from cameras, monitoring where his family is, maintaining security >> to his heart his heartbeat off the rails. >> Knows when he's going to wake up before he actually wakes up. >> Can detect these things tracking license plates of cars that are parked outside >> off the rails. Um, I was talking with him a little bit at the speaker reception afterwards and uh, he showed me pictures of his server rack and his GPUs that he has installed.

>> He's got a server right in his house, right? >> He does. Yeah. And so, uh, this is why we're going to need distributed inference, distributed compute. >> I mean, look guys, I'm I think full access is reasonable with iron glass is not OPSC advice essays. Um but um I you know like like I said we you try open claw and you're like all right takes me like five hours to get to a point where I can have it edit Google docs for me. I'm like bro like this is ridiculous. I can just do this on my own. Like >> quite a bit of lift. Yeah. Or use like Gemini has it embedded into Google Docs.

But the like the breakthrough with ironclaw is that you don't need to go through all these excessive security measures because the idea is that the secure enclave where the uh cloud server is being hosted >> is in it is in a trusted execution environment. It is also there's also stronger buffers and barriers to prompt injection which is I'm learning is the biggest issue of autonomous agents. Imagine if you have an agent talk to your agent and they try to get you and your agent to send your personal data over to their agent because they're trying to inject the prompt of your agent by giving it a prompt that that

obviously you wouldn't give it to, you know, you wouldn't give your agent yourself. >> But I think from what I understand, ironclaw is beyond that. Um, at least a part of of protecting against that. >> Let's uh perform a bit of a thought experiment. All right, >> let's go on a journey. >> Uh, we get back home. We're setting up our iron claws on near.com. We're activating our agents at a reasonable tier. >> Yeah, >> we get them hooked up to Telegram so we can message. Maybe we also give them uh Google Voice phone number so we can SMS them. Uh maybe hook them up on Discord or WhatsApp.

>> Okay. Then we ask it to spin up its own accounts. >> Yeah. >> We uh it sets up a Gmail. It sets up a notion. It sets up uh Google Drive and then we start giving it access to all these things. We start forwarding emails to it. We start asking it to take care of things. We start looping it in on email threads that we personally don't feel like we uh we ought to respond to. Then we start asking it to create sales material. Then we start asking it to recommend podcast guests, reach out to podcast guests. We ask it to make a LinkedIn account. Then we ask it to do some cold outreach on LinkedIn to some of our ideal podcast guests. We ask it to make a Twitter. We ask it to engage.

We ask it to pull some of the most recent news. We go on this journey with our ironclaw until it's time that we create another AI. >> You know, we just ask that one to make agents under it, >> right? >> Who are responsible for us for specific things? >> But you're scared, man. You're scared. >> Why? Why am I scared? >> You haven't jumped into the dark forest yet. I um I'm enthralled by the dark forest.

>> No, no, no. You been spending time outside >> while while the whole world's been changing inside. >> The enlightened forest. >> The dark the outside nature. >> The real world. >> Yes. The real world has not been the same for me since I downloaded Open Claw. >> Really? >> Yeah. There's >> Tell us tell us more about this new perspective that you've gained. >> You can tinker around in Telegram. Yes. and build different clawbot implementations. >> Yeah. >> Or you can go outside, >> right? Yeah. >> And play golf. >> Play golf. I do have a golf shirt on.

>> I'm I'm ready to get back to Florida. >> Yeah. So, um it's also important to uh prevent infinite tinkering, which uh some of us have fallen prey to. Uh >> I feel like a lot of the AI work these days, people like, "Oh man, like just download Open Claw, like I got this new uh primitive, you know, I'm gonna, you know, automate this and try that." It's like six hours later you're like you you're stuck on the same like terminal window like trying to like copy an API and like you're just broken >> which is what I think a massive unlock could be for ironclaw is once and George from Near talked a bit about this which is when these uh these ironclaw accounts can go uh investigate documentation on some of these website themselves right

we just prompt it say hey go to polyarket.com set up uh an account and access the API. Uh and then once you have but again the the blocker I believe for that right now is you need a way to pay for the API key. >> Yeah. >> Once it has stable coins or wallet or some credit card information or something like this then it can start to pay for API keys. it can start to make API calls and again API stands for application program interface, right? Which is going to be the interface that applications can engage with AI agents.

So agents can can start to interact with some of these APIs, some of these programs via these interfaces, but they need documentation >> so that they can train themselves on how to use the API, which is the same, you know, course of action that a human would take if they need to train themselves on how to use an API. Uh, and then they need a way to create the account, which I believe they have right now. Uh, and they need a way to pay for that account. And so, um, once they have a form of money, >> then they can start to buy these accounts, set them up, and engage with these APIs, >> which unlocks the entire breadth of the internet at their >> all of this is like the next era of um of of usage of API credits and usage of

power. And if we look at the markets today, we're seeing this rerate in real time. Like we're seeing the rise of uranium and the rise of copper and the rise of other of uh of stocks that are like uh like Sandis has been on a crazy run. We're seeing other Korean random stocks that are powering the AI trade go on random runs. Uh you guys know I have a big bag of uranium I've been underwater on for about three weeks still. Uh but man oh man we're ready to go. >> So uh what what a little TA on that chart.

>> N it's all good guys. We got Alexenko the the CEO and co-inovator co-creator of Near Intense up next. After him we're going to go rapid fire into a ton of other guests. Stick around. We'll be right back. Quick break and we're going to get Alex up in the show. See you guys in a sec. Hey.

Heat. Heat. Heat.

Heat. And guys, welcome back to Rollup TV from

Nearcon. Uh, the government of Bermuda is doing what governments do, not showing up when their citizens need them the most. So, here we are today with our lovely friend Otto Otto Jacobson, the CEO of YAP Global, uh, a friend of the show, a friend of the company. It's good to have you on. >> Thanks for having me. It's good to be on >> glad to have you. We're at we're at we're at Nearcon talking AI agents, talking this the intersection of AI, crypto. Uh before we get into kind of like what's going on in the market and in the state of affairs in the industry um we are very stoked to be running a media company right now because we feel that video hasn't really been disrupted yet

>> by AI but uh when I look at research when I look at blogs when I look at newsletters I we've put a lot less focus on written content from just all aspects of of of the business from descriptions newsletters blog blogs research. We haven't really invested much into it and we don't really see us, you know, ourselves doing that in the future. Is this a fair strategy? Do you, you know, do you agree with this? What's your take on uh >> Yeah, I guess the question you're really asking, Andy, is like why would anyone hire a PR firm in 2026 when we all know that traditional media is is dying or bleeding out. Uh, look, I think especially when you work in crypto, digital assets, which is what we do,

right? And AI and fintech as well, you need to look at what we now call new media, but other types of content, right? Like like what you guys are doing with video, podcast more generally, uh, newsletters is a big thing. And the I think the the the bullc case here or the positive scenario is that these can become like super niche um in terms of what they cover in terms of what they really want to write about or you know talk about on the podcast or have on you know a show like this but have a global audience. So therefore, it's a viable business model for you guys to uh well, we can talk more about what what where you're going and and what kind of guests you want to have on the show, but you know, you can talk about rollups for and that can be for a global audience and

it's big enough so that you can really focus and drill down on that. And I think that I think that's a good thing, right? Uh distribution, you know, globally is is is key. Uh and with with internet and now with AI and so on, that's getting easier. So I think any PR firm worth their salt thinks about this and they think about the client and think okay what kind of newsletter should we be placing them in what kind of podcast even if they're super niche you know builds credibility at the end of the day we we we help our clients build credibility with the media what if it's a newspaper podcast if it's a live show like this yeah I >> I think we're in a unique position because we started before AI was created uh I can imagine in a few years there might be a new podcast and we're already

seeing this with musicians where they have good music. People think about the music first. Okay, I like this thing. Then they worry about is this thing an agent, an AI, or is it human? And there's actually really popular artists on Spotify right now that people are not sure whether it's an AI or it's a human. >> Yeah. Yeah. >> And and Yeah. Go ahead. >> No, just on the podcast I I literally heard this morning in a meeting that someone has taken the Epstein files, all of them. >> Yeah. >> And put them into Yeah. 60 episodes, right? like 60 episode podcast would cost a lot of money and it would you know take a lot of time to produce >> two years ago. I don't know. Uh but yeah, they done and it's really popular and they're getting sponsors as well. So someone's making money out of it.

>> Yeah. >> So when this thing is super niche and someone can prompt an AI to spin up a podcast of the thing that they want to hear about. Yeah. >> Like where's the differentiation, right? Like how are like we compete because we are not AI. we can curate something that we're genuinely curious and interested in. But when someone is come coming up and thinking about what they want to learn about or what they want to hear about, is it just a matter of laziness that they don't want to prompt the AI to spin up a podcast to their choosing? >> I'm not sure what you're actually asking, but I I think like what what people want, right, is yes, niche kind of content and so on, but what they really value is relationship and, you know, having kind of human interest,

right? like why your show does well, you know, and I think it does, right? Uh is because people respond to kind of talking to you or you you chatting to other people that they like and they want to hear about it and they want to hear kind of your views on it and like you know less formal but I think with someone like TBPN, I hope I'm saying that right as well like people respond to those two guys. I think the question that you might be thinking about as well is if they continue to do well, continue to be successful, how long can they be like the primary face and when do they bring in co-hosts and so on, which is what you see on like radio and traditional TV and therefore the brand gets a little bit >> the brand network, right? Blockworks has

done that with their shows, >> right? Bar stool obviously famously has done that. Well, that >> we're not that far away from like them almost like manufacturing their themselves. >> Well, like Ral Paul has AI avatar personas, >> right? >> As well, like whole like do a show and like he'll have like this is not me. This is like my avatar doing the content. >> Yeah. Yeah. But so you guys have been going for, you know, with the roller for a while now, but basically like four, five years. >> Yeah. Five years. Yeah. >> Yeah. With live content. >> No, just just a year and a bit.

>> Always been video content. video coming live content has been about a year. >> So that's that's pretty good, right? But I do worry about some of these newer kind of new media channels that pop up like what kind of staying power do they have? Like and people get tired, right? Like >> how long can they keep it up for? Uh maybe one year, maybe two years. Maybe that's okay. They just kind of come and go, but I don't know if you want to build like a lasting brand and a lasting business. Yeah. >> You know, how do you think about that? >> Well, you you always have to be kind of pivoting in the space. It's like you always have to be pivoting. Where does a sponsorship uh you know cash come from?

Where does the where's the market going? What is what do people want to learn about? >> I think we're under like what do we constantly like push ourselves to go kind of push the frontier and continuously learn about? And I almost think about it like yeah we pivot we pivot left and right but it's also just kind of an optimization or refining. >> Yeah. >> Which is what I which I was asking about is like how when we when we hyper niche >> like what does that look like? Why is that the differentiation and why is that a moat around like human created me versus >> because I think you can quickly become like a domain uh you know a subject matter expert right like for example I tell new people that join our firm and we're about 30 people now and they you know maybe they're fresh out of uni or

whatever I say with crypto right and digital assets is so big now that you can't pretend to know everything that's going on you know be an expert crypto almost no one is that right uh but you have to find something that you find interesting and where you can have a little bit of an edge, have a little bit of a niche. Uh maybe it's one ecosystem. You know, we used to have a lot of like different blockchains and they would have their own kind of particular ecosystems. Uh it's a little bit less of that now and maybe maybe that's a good thing. I don't know. Um but you know, what's your niche and can you become kind of an expert in something and then can you write about that? Can you tweet about that? I mean, that's what we actively encourage our kind of team to

do right? >> Yeah. How would you describe the shift going on right now, right? Like obviously we've told you a bit about what we're trying to accomplish this year. How how do you think about >> um and I'll give you a very concrete example, right? So some of the teams that we're speaking to, >> they're like we well we go on Bloomberg for free >> like you know our CEO goes on Bloomberg whenever why would we ever work with you guys? That's like you know that's just a very interesting >> object goes on Bloomberg whenever maybe but yeah like or yeah we have a Yahoo Finance PR connection but so like how are you thinking about this like next step of media in digital assets? What does this look like? what are the key, you know, distribution, like what are the new sources? What are the key

angles? Like, like how are you thinking about transitioning your company into this? >> Yeah. So, I mean, we only do earned media, right? Like we love working with you guys and and but and you do kind of both. Like there's definitely a role for paid media if you want to be a successful digital assets company business in 2026, you should be thinking about media and you should talk to Yap Global about that, >> but you should also obviously be thinking about paid media. >> Yeah. >> And the roll up. You're a great example of that. like you need to do both. Like gone are the days where you can just pretend that you can do one thing and be successful in terms of marketing where you can just do socials like social text. That's a given now, right? Like most of our clients probably have that

automated with a with a >> AI agent, right? Um >> and then they might add some, you know, personal takes on top of that. But anyway, so in terms of like three things that I've been thinking of since I took over as CEO is the professionalization of the space. >> Okay. And that's kind of also we can talk a little bit about that but like the fintech narrative that is going on with with crypto and digital asset companies but so professionalization then regionalization crypto especially D5 brands like you worked with countless D5 brands so we they launch a product you know borrowing lending um you know AMM whatever it may be and they have like we have a global product and anyone can use it anywhere. Why would we need to differentiate how we talk about it in Asia versus the US versus Europe? Well,

obviously with regulation, you do need to do that because you can do only certain things in certain reg uh certain jurisdictions when regulation comes in. So, the regionalization of the space uh and then, you know, we haven't gone like hyper local in crypto yet, but I think the brands that do they will do well. So, they will develop, you know, maybe a product for a city. Uh I'm not talking about like New York City coin or anything like that, but still taking a very localized approach. So, so far it's regional. So, we have some clients who have, you know, a a number of PR agencies that maybe focus on the US versus Asia versus >> Yeah. etc. And then you have the diversification and on the other hand specialization.

So, uh that's more for us and how we think about this is you know as the space grows we want to do more things in the space. We want to diversify away from just doing earned media for example but we still want to remain specialized. Yeah. >> So, it's a little bit of like two sides of the same coin. And then as far as like some of the distribution sources that you think are are growing more valuable, like what do you think that evolution looks like now? Like is it media versus >> less important? >> Is it like a LinkedIn world now? Is it like an inerson world now? Is it a like how do you think about that? And then how do you think about uh like what these fintech or institutional kind of client like okay so where do they live as like a base but then what do they

want? like I'm a ETF issuer or or an asset issuer, tokenization company, like what what do I what do I want from a more like crypto cryptonative media source like you know Yeah. >> Yeah. I mean if they if they're even thinking about like crypton native media then they're not >> I think yeah they might not be right. Um I think there are options out there um you know whether it's kind of online print or or podcast and so on. But I think the the the key question becomes right with the new media that is popping up is it going to be around because someone making a commitment to sponsor a show okay maybe it's not a year-long contract but still they want to know that there is something that's going to

have like lasting value and at least that the you know content will stay online to be there as kind of evergreen if someone searches for the product whatever it's like I do think there are questions about that with new media uh properties that are popping up. Um, so and you your question was more like >> okay so like how is distribution changing and then and then what is it that like if you're going to go work with a fintech or an institutional more client crypto what what do they want? >> Yeah. Yeah. Yeah. So to your point about X versus LinkedIn I mean yeah for sure less focus on X right like that's kind of you have it as a baseline. um you still you still you you're more maintaining a presence there rather than like actively trying to grow it maybe and then LinkedIn is where you focus on

creating content that will generate leads generate sales because that's the still LinkedIn is kind of it still works right like there is still a I mean even I've gotten like uh successful leads out of content on LinkedIn um I think there's a question now with AI you know how this I can't remember who said this but basically like it will just become so optim optimize that everything will just become slow. So, I wonder when that will kind of when we'll start to see that on LinkedIn >> because, you know, my friend who runs a >> kind of AI startup, he of course has an automated like outreach uh agent that does that for him. Yeah.

>> Um so, he's constantly scraping his almost his like top 100 leads. He's constantly pinging them. >> Yeah. >> But everyone else doing the same. So, how do you get an edge? >> Yeah. Yeah. >> I think you mentioned in person. I do think that in person is still an edge. I'm curious about like event strategy for for you guys >> generally when you're advising teams now that you're working with or just you know friends and founders what like what is your event strategy for 2026 2027 like how has that changed?

>> Yeah. Well the first thing I would say is like have we seen your overall strategy for 26 like what's your business strategy? >> Yeah. If you're still in like the more um crypto like developer maybe you're building like tooling or something quite infra you know are you going to conferences where there is a large presence of developer if that's your target audience >> or are you going to talk in you know Dubai which is a great conference but maybe not so focused on those kind of topics >> are you fundraising then maybe talking and talking Singapore great places to go and try and get investor meetings right >> so it depends on what your strategy and what your priorities are for 106 I mean, it has to be that, right? Gone are the days when the strategy for going to, you know, big conference is just we're just

going to roll out, spend a lot of money, and hope that something sticks, right? >> Yeah. >> Which was, you know, never going to work and it was a bull market thing. >> Yeah. Yeah. >> Um I'm curious how you think about uh Yeah. Like bull market thing versus bare market thing. You mentioned leads getting generated on uh LinkedIn to ultimately arrive at sales. uh how much do you take a marketing approach like we want to generate some impressions, we want to get that top line and then that will ultimately flow into some sales uh eventually versus much more direct. We want to generate you leads, we want to see a conversion rate and then we ultimately want to acquire more customers.

>> Yeah. I mean for our business our business isn't huge, right? like we we get I think 80% of the like clients that we sign are warm referrals or introductions like but again we're not you know we're not a big company in that sense we we sign maybe 20 clients a year or something like that >> um so but we still do LinkedIn right I mean everyone needs to do it >> um >> I think the the interesting question is do you have content that will help you convert >> uh and I think if you don't have that in 2026 you've gone, right?

>> And and how like what is the STRA is there any general rule of thumb for companies to create content that helps close leads? What like how does the content differ from just any old content? >> Yeah, this is maybe a little bit outside of my my area of uh expertise. But yeah, um yeah, I think the people still respond like we talked about, they still respond to people, they respond to story, you want to highlight, you know, your founders or your founding team, talk about the mission. I think you one thing that we're working a bit with our clients on is like case studies and uh client success stories maybe is what it's called in tech generally >> and highlighting that like >> okay we moved away from marketing like tokens as the product right like tokens as the the hope or why someone should

care about a project um and now we want to talk about products and what's actually helping them >> right or their users >> yeah uh last thing for you is is AI helping helping you with your company. >> Yeah. But it's also a bit Yeah. It's scary, right? Because I think probably some of our compet well like some large PR agencies, they can obviously invest a lot of money into AI and I know that they're building like for example benchmarking tools and so on their own benchmarking tools. Um and we're not able to invest the same amount of money.

Um but yeah, we're I mean we're running into it >> for all different types of uh just workflows or your own work. >> Yeah. Yeah. Yeah. Cool. >> Just one thing um that we didn't chat about is like the fintech story, right? >> Okay. >> Um I think two three years ago most of our clients said like fintech, who cares about that? That's like trying to build something on old banking rails, maybe build an app that's using old rails. >> And then now it's shifted completely, right? Like crypto brands actually want to be seen as fintech brands. >> Yeah. Why? Because the most successful context have taken, you know, well, they've gone Swedish originally. There's some success stories from Sweden.

They've gone from Sweden, they built a global brand like CLA. >> Yeah. >> Has known everywhere that matters, right? >> Um and they have a um a regulatory strategy that works like they've been able to figure that out. So, they build a global brand and they have regulatory compliance. Those are two things that crypto brands wants to do as well. >> Okay. >> And that's why they're looking at fintech. And what does that even look like from a strategy perspective? Like is it a branding? Is it a positioning? Is it events? Is it content? Like how does that even look? >> Well, maybe even me during the show has hinted that like we talked less about crypto, we talked more about digital assets right?

>> Yeah, of course. >> That's a branding that's a choice of words that uh that you know that is part of the strategy. >> Yeah, definitely. >> Uh digital assets probably more palatable to banks than than crypto. >> Yeah. >> Um so yeah, it's definitely a branding. It's it's um with regulation that's obviously a lot of compliance that goes into it and so on but it does influence your communications right >> definitely >> you can't market to retail probably um as as wildly as as you used to um you know impact things like how you talk about your product on on X and yeah >> thanks Otto thanks for your time >> thanks for having me >> cheers buddy >> yeah man I mean the the industry is going through a massive shift I mean it's happening day in and day out like we just keep seeing the uh proliferation

of new companies, new uh words like even trying to you know refrain from using crypto in your in your brand amongst others. Uh speaking to a new audience, marketing to them and it's definitely happening. >> Yeah, I mean crypto is getting seen as a liability especially when prices go down. It's just negative vibes. People just don't want to be associated with it. >> Just wait till we get to Austin's prediction. >> Yeah, Austin predicts 35k is when the bare market starts. Uh we'll see. Um but yeah, look, it's a bull market and a super cycle in so many like like blockchain things that aren't >> today. Meta stablecoin re relaunching their their plans to launch a stable

coin. >> Just just rattle off a few. It's stable coins, digital assets, and tokenization. There is a massive massive adoption uh underway right now. >> Huge. >> Huge. And that is happening right before our eyes. But uh when people think crypto, they think prices. You know, you saw the Coinbase Super Bowl ad. The sentiment is down. It's just not really what people want to be associated with. They do want to be associated with tokenization. Like that's why you see Black Rockck's Larry Frink going to Davos talking about tokenization in front of the world's elite. You have digital assets. Uh this is being adopted on really every stage. uh most predominantly in the fintech and the

private credit space. Uh digital assets are you know what is getting tokenized. Uh and then you have um and then you have digital what what digital assets tokenization and stable coins and st of course stable coins right and so then once you have stable coins and digital assets tokenized you end up with high quality assets on chain. This is the neo finance thesis and then you're going to want to do something with those stable coins and those digital assets. The first step is tokenization. Then we're going to get composability and then that's when DeFi Lego block building blocks, DeFi Legos like the Morphos, the Aves, uh, and and more of these lend borrow markets, these vault strategies is when that really starts to explode.

We haven't talked a lot about vaults today, but that is where I think a lot of uh AI uh blockchain uh finance starts to converge because you can imagine an AI asset manager for a vault. >> Uh whereas right now that's that's really uh unpractable practical to do. >> Yeah, guys, we got Alex Shchenko, the founder of Intense here in the near ecosystem. He just gave his keynote. I think we're going to have him up right after this. We're going to take a quick standby, quick break, and we'll be right back with Alex. Thanks, guys.

Hey. Heat.

Hey. Hey. Hey. Heat.

Hey. Hey. Hey. Heat. Heat. Hey. Hey. Hey.

Heat. Hey, heat. Hey, heat. Heat. Heat.

Heat. Hey, Heat. Heat. Heat. All right, guys. Welcome back to

Nearcon. We've got Alex Chevchenko live on Rollup TV from Fort Mason. There he is. Alex with the big product release yesterday near.com the multi-chain interface uh talk to us about what you guys unveiled yesterday. >> Yeah. Um hello everybody. Uh so there are lots of different changes that are happening and uh previously if you have been using nears uh directly on our own interface it was located at near-intens.org And we decided to find a better place for this uh experience especially taking into account that there are quite a lot of innovation that is happening on the protocol level including confidential intents and uh we just wanted to have one authentic place um so that everyone

are able to experience it there and that's why near.com pretty expensive domain u but uh now near.com is the face of near >> yeah and So talk to us about what bringing all of these things together means, what it looks like. Now we have near intent, we bring in the ironclaw. These are all under one roof. Near.com. Talk to us about what happens now that these are more integrated together. >> Yeah. So from my point of view, the core use case of crypto and the core use case of Neans is everything about finance, right? moving assets around, setting up your portfolios, integrating yield bearing products inside of it. So that's

pretty important and this information is uh is your financial information. This is the most private thing that you need to have. Mhm. >> So from that standpoint it's very straightforward that we need to if we would like to have AI capabilities and definitely we want to have because AI agents becoming more and more capable um we want to bring this AI tools inside of this solution. So a great example for this is uh like uh many people are not having the access to financial assistance, financial consultants >> uh and uh uh and we know with AI now it it can take your MRI scan and tell you exactly what's the problem right uh is

finance vertical anywhat different from my point of view no right there are concepts in finance uh they are a little bit cumbersome Um but in general if you put it in the proper language uh many people can understand this. So I'm thinking of AI being like an universal interpreter into human language from the language of the white collars from the wall street and uh since it is financial data it is private data we need to have a private AI solution uh that that a user is 100% sure that this solution is not sharing his information. That's why private AI branch and near intense the financial backbone for all of these systems that's a that's a marriage made

in heaven. These two things must be built alongside one another in order to uh enjoy the benefits of of of each of these things. >> Yeah. And so you guys also kind of launched this uh AI marketplace as well with that includes you know open source models, closed source models. How does um this kind of problem like what problem does this uh solve for for the market and how does that kind of tie into almost looking like an intent itself? >> AI agents are capable uh but they there is no single AI agent that can do everything.

uh they have their own specialties the same way how we have different models that are doing different things right >> uh Google VO is doing videos um uh chat GPT is very good with reasoning cloud code is an amazing tool to do software development and the agents are also specializing in particular things so in some cases an agent your personal AI agent can understand that there is a task that it it is not equipped to deal with.

>> Yeah. >> So it can ask other agents actually to do this and uh we need a place where these agents can communicate, can find each other and can set up tasks for for others and actually can implement this tasks. >> Yeah. >> Um yeah, it's it's like a gigantic uh gigantic vacancy fair. >> Yeah. uh for the AI agents and uh this is a definitely needed piece of infrastructure to make uh all of this stack very useful. >> Yeah. So I think something that's interesting is that recent recent data is showing that 80% of new AI startups are now building on Chinese open- source models like uh Zippu GLM5 QWEN which are currently outperforming US models in

cost efficiency by nearly 70% at times. If the best bang for your buck intelligence is coming from the east, how do this western world and western enterprises balance performance with data sovereignty, geopolitical allegations and regulations? I can talk about this for for hours um because I'm the big user of AI at this moment and I'm using all the spectrum including open weight models like quen or llama um and uh and also closed sourced and closed weight models uh from entropic and open AAI. So I think the great thing that we see is that the models perform performance converges and um the version of the world that we are

evolving into is actually the version where um open weight models are not that far from the frontier models and the the difference is maybe like several months to maybe half half a year of the research. This is amazing because it means that people are able to have this models and make sure that they are not sharing their data with the centralized providers. From the other standpoint, it's not for everybody to buy a Mac Mini. And on this Mac Mini, you actually cannot have the big model like 70 billion model or even bigger models. For this, you need to have Mac Studio and for the Mac Studio, you need to spend 10K or even more. And uh from that

standpoint all of this capex is is very hard to justify. So having a solution that from one side is cloud-based so that you are able to access these inference uh without these capex costs uh at the moment when you need to have it. But from the other standpoint secure enough and with privacy built in >> that this is a a very very important topic. Um yeah, I'm just fascinated by all of the developments here. Uh and uh encourage everyone uh who have not been using AI before, go and try to do this. Uh there are many interesting use cases that you can apply the AI for uh right now and

I'm pretty sure that there will be even more in the future. >> Yeah. I'm I'm curious at what point you think AI progress becomes a threat to the national power grid. Um should we cap AI growth to keep the lights on? Um so from the power uh perspective I think it's okay right now like um you know stop targeting the rights of AI companies because the market says the price on the energy right uh and uh uh yeah it's just the open market and and that's it like if if there is if there is more demand it's like it's like with RAM at the moment >> have you seen the surge in the prices of of RAM?

>> Yeah. >> Yeah. It's like the producers are able to to get much higher margins on the server RAM. That's why they are reorienting their production lines towards the server RAM because it is needed for AI. Yeah. >> So it's it's it's not very good. But if we're working in this open market paradigm where the market is able to balance out many many different things much better uh than laws and uh you know some individuals can then it's kind of good. From the other standpoint, obviously I would like to have lights at my home.

>> Yeah. >> Uh uh but I'm following very closely the developments on the um the diffusion reactors. At this moment there are quite a lot of innovation that is happening there. There are lots of smaller startups that are not doing insane installations, multi-billion dollars, uh very long projects, but instead they're focusing on the smaller fusion reactors. And there are quite a lot of innovation that is happening here. I'm pretty sure that by the end of this decade, we're going to see production nuclear fusion reactors that are going to deliver amazing performance u in terms of the output of the energy. So uh maybe right

now there is a there is a little bit of a of a hurdle >> uh with electricity supply but um in a couple of years I think this problem will be completely off the shelf. >> Yeah. Yeah. The uranium bull fees just continues to be grown here in San Francisco at Fort Basin. >> Yeah. So by the end of the decade, this is going to be something that is going to continue to proliferate, especially as we draw more and more uh needs for power. I think we're we're finding them all across the base. >> Alex, I think that's it. Thank you so much for joining us.

>> See you in an hour. >> Absolute pleasure. Yeah, we'll see you in a little bit with >> Thank you guys. See you. >> All righty, guys. Up next we've got our our friend Shen Yoda uh Ajit Trapathy talking about the state of affairs in this crypto AI crossover. Ajit has recently uh joined the IEN cloud foundation working on things there and has a background in investing in this crossover as well as AI experience. Uh Ajit what's up man? How are you? >> Oh always good to be with you guys. You know I'm a I'm a fan and a friend.

>> Thank you. >> Friend of the show. front of the show. Good to see you. >> Long time from the beginning. >> Oh, yeah. >> Yeah. >> Um, how's how's things going? >> Uh, pretty good. So, you know, we uh I mean, I also have the same crypto price anxiety and >> AI unemployment anxiety as as everyone else and US jobs. What is it the global geopolitical uncertainty anxiety and I think I have like you know quad quatic anxiety now. So, four or five dimensions of anxiety. >> Yeah. And uh but other than that, you know, when you uh the day-to-day, I mean, near these guys are, you know, Ilia, I'm a huge fan. The shipping really cool kit and this this conference has so much vibe.

>> Yeah. >> In a in a really bad market, right? So, it's kind of I'm surprised by how vibby it was yesterday and how many builders are still here and they're inventing cool things and is kind of single-handedly holding the fort for crypto. Yeah. >> In San Francisco, >> right? He is. Right. >> Yeah. I mean maybe you know the stable coin guys are around and tempo and >> you know uh now Zak is launching a stable coin uh and bringing DM back there are also Aptos and Suie so you know there's but I think Ilia is kind of the >> the the most credible you know face of AI in the the in the crypto world and you know the most credible face of crypto in the AI world. So he kind of occupies this unique place and

>> yeah and he's kind of held his conviction in spite of the social you know pressures of San Francisco which is quite remarkable. So I think he's been able to attract builders. I think tomorrow we're also going to see you know I work with a foundation and uh >> we're going to see Shri Ram that's another really credible face right of tech uh one of the legit most legit builders going around uh >> working in a similar space around you know verifiable agent sovereign agents uh and sort of bringing the power of crypto to AI in a in a sensible way >> not in the you know oh yeah we launched a meme coin for u which annoys Steinberger and everyone else but you know okay yeah you know more of a technology okay product utility. What can we do with this stuff? So, >> yeah. So, so I think I'm pretty happy to

be here, you know. Um, >> yeah. >> Yeah, that's and and and let's see, right? I mean, hopefully the uncertainty and liquidity conditions will resolve themselves and and business will boom again. >> Yeah. Let's talk about the the the kind of crypto industry uh at large first and get into some of the AI work, some of the cloud stuff. when you're looking at the kind of phase of maturation right now, you come from a, you know, a trei background and kind of got into DeFi, then got, you know, kind of seen both sides. It it feels like, uh, digital assets is the new word. It's not crypto, it's fintech, it's not crypto. Feels like there's this massive like put the suit on, grow up, you know, be in that kind of like finance realm, not this like meme culture like just talk to us about

>> in all those cultures as you know. So and I've traded >> you know fartcoin and >> so tell us about the evolution like tax. >> Exactly. So like where do we go from here? >> So yeah I've been all there. >> So like where does culture but where does like where is there actual use cases going from here? How do you see the space growing? >> I I mean I mean so so a lot of this is driven by liquidity conditions right so crypto every four years and it's a self-fulfilling prophecy. We just start dumping in the towards the end of the third year and and then we then we all cry and say crypto is dead and >> we're all leaving right by I wasted my life. All that [ __ ] goes on. So that's not new. I think we've been through this

over and over again. >> Uh leverage is endemic right to crypto but leverages and precious metals and and stocks also. I mean we just the purp uh but I think ultimately crypto needs law. >> Yeah. >> Right. So yeah use cases are coming through. I think stable coin payments, right? Uh we know this stuff works. We know that, you know, I was in Redmond with my IIT Microsoft guys and >> and they are really spending quality time on stable coin payments, right? So, I've been doing some work with Noble >> on that as well. And uh so yeah, so stable coins are here, stable coin payments are here. You know, tokenization of securities is a longer process, right? I think guys have done well. Robert is doing a phenomenal job.

It's security keep pushing. uh but it's still a longer game right I think when you think from a consumer perspective from a regulatory perspective yes it's kind of there right from a consumer perspective we don't know >> if you know I really need a blockchain to trade and I've debated this endlessly with a few others the use case for you know sure there are post trade efficiencies sure we can get rid of some intermediaries right but do we really need a blockchain to trade trade securities that thesis still needs to be proven but do we need a blockchain for stable coin payments Of course, right?

But then does it really benefit in a crossber context? Absolutely. Does it take out, >> you know, a lot of costs? Yes. Can we integrate it into loyalty programs like the Noble guys are doing? Absolutely. Right. So, can is it going to be the main state of agentic finance? >> Is it much easier to give any AI agent a stable coin wallet than a credit card? And you know, who knows what the hell happens with my credit card, right? It might end up end up in Sure. Maybe not. >> Yeah. Yeah. And and also risk is easier to manage, right? And uh >> well, that's what they're saying, right? They're saying that this agentic finance is going to be the thing that brings crypto out of the ashes. We're going to get a ton of agents onchain.

>> It will Yeah. Yeah. I mean, it'll definitely bring us uh two things, right? One is uh agents don't have the friction can deal with the friction much better >> than humans, right? Of using private keys and things like that. >> The other thing is I think uh crypto has an obligation to demonstrate some utility, right? So, so, so we are I mean beyond Bitcoin and speculation uh because you know this society has waited for us for 15 years. >> Yes. Right. >> Or 18 years. Uh but look I mean neural networks were dead for a very long time and now we are all getting replaced by neural networks. Right. So >> so so things change rapidly in tech. Uh I mean this is a very young technology.

So it goes through, >> but I think the one thing that really disappointed me was the was sort of Coinbase pulling out of the Clarity Act. Of course, they have their reasons >> at the end, right? >> Yeah. But I think that really hurt us a lot, right? I think uh one thing that all the lawyers I speak to in this space you know uh Rebecca Miles Miller all of the OG uh policy people I think we we have all come to recognize that to really bring institutional capital flows in size right beyond what we seeing in the major CTFs we're going to need uh some rule making and some laws right >> and then the second thing is the risk factor so let's say a corporate or uh or a trafi institution deploy employs a lot

of capital on behalf of clients today into crypto, right? Uh then the question is if after midterms the Democrats end up with a lot of seats or you know or or if uh or if they come back you know who knows I mean Warren might start building an anti-crypto army. So, so that's that exposes these institutions to a lot of risk right of being called up and said bro what were you thinking >> right so so no one wants that risk so so clarity act is fundamental uh genius act raised stablecoin you know market cap by 50% >> after it was passed I think Chris Dixon mentioned that >> so so so to me I mean there's nothing more important than crypto going from code is law because we know there is no

law right I mean we've been in the trenches we know what kind of behaviors we've seen >> and the kind you know just free for all misconduct that goes on in this that requires some fear of God right and then also some rulemaking so that legitimate entrepreneurs like Ilia >> and I mean you know everyone the the Vitalik and all the good founders building cool things have a have an advantage over >> you know whatever the hell >> else we do around here right >> so so and and also VCs right I mean it's going to be very hard to draw capital >> if uh you know altcoins everyone loves to say altcoins are dead right but then what are we even doing here Yeah, >> I mean then that's it. We just have Bitcoin and a few ETFs and we're wasting our time here. I think that's a terrible

outcome, right? So for altcoins to and all the innovation to really come back and draw capital, >> we're going to need clarity, right? There is no way I think we're going to get capital into altcoins from trade without this whole ETF rappers and whatnot unless unless we have clear guard rails around it and that's why clarity is so critical. >> So you know the agent commerce is one one thing that might bring us out of the ashes. The other one is uh you know like what what do you make of the black rockck going into the unis swap and the Apollo going you know into morpho and you know kind of as we tokenize these things these high quality assets come on chain they ultimately you know they do get the reduced settlement maybe the better fees but ultimately maybe the best value proposition is the

composability with DeFi so that you can get liquidity out of these things where they're inherently locked most of the time and then ultimately these things get more fees and that's where there's real businesses that's how they end up as like legitimate token altcoin complex. >> So, so, so these institutions have very long time horizons, right? They can wait for governments to change. They can wait for geopolitics to change. >> Black Rockck, Goldman Sachs, these, you know, they they've been around so long, you know, more than 100 years and in Goldman's case and you know, they've seen so much [ __ ] that, you know, they plan in 10, 20, 30 year time frames just like the oil companies >> whereas crypto Twitter, which is where we live, uh, plans in one time horizons and, you know, and over five five weekly

candles you dropped from 97k to 64k and obviously you're going to be in clinical depression like maybe I am right now right so so the time horizons are very different right >> yeah it's like and and once you add the Trump era tariff uncertainties and you add you know AI you know psychosis going around and you multiply all of these things together and you're like bro I I need my family I need a dog I need to hug my you know then you need to go to the church meditate for 3 days that's where I stress emotionally that's where I am I'm there >> right so so so yeah I mean and then the jobs issues and whatever else right >> the golf course >> yeah so so that's I mean very honestly

that's where I am right but when you think about where institutions are at I mean they are >> hedging their bets >> right one they have to hedge their bets against disruption >> right let's say this onchain technology securitization goes that continues right putting 1% of the assets or.5% of the assets is Right. But it's a huge amount and it's it's a huge impact on the overall 4 trillion well 2 trillion now >> 2 something trillion now >> it's always been like this right we die every few years and every time it looks like we're never coming back and honestly we don't know right so 2019 was like this I was I I really don't want to remind myself but I have serious PTSD about being broke and unemployed in 2019 after ETH crashed

>> uh and I don't think I'll ever come out of it right uh but that's a good thing because then you learn to sell a bit. >> Yeah, exactly. >> Uh so so so yeah. So so so I think from a crypto Twitter den, you know, short time horizon and then I think politics is a big part of this, right? So I think clarity has to be passed because that gives the president an election year win. >> Yeah. >> Right. And he needs election year wins. We saw on Fox News. Uh I think uh there's a there are a few members of his team who are >> ideologically committed to making this happen. >> Right. Not just like uh vested interest and all that, but ideological commitment. Patrick Devet and others.

So, so I think I think I think they're going to go very long way. >> Yeah. >> David Saxs needs a win. >> Yeah, he does. >> So, you know, uh I mean on the AI side it's all but you know 3 months 6 months ago AI was all only good things. Now I think Americans are really worried about data centers and job losses. >> Yeah. >> So so there is a political drive to make this happen and and in institutions will come in in much bigger size. Right. I mean the so when you look at our market cap it's 2 trillion. Yeah. >> But how much real cash is in there? Right. I mean is it 200 billion? >> Yeah. Right. So because it's all just levered up and you know paper money and you know what not. So I think there's not that much. It's such a small

industry. Right. >> You double it. >> A lot of us will be in clinical exuberance, right? Some kind of manic behavior as opposed to this depressive thing that's going around in I think we'll and we saw this in and it's all liquidity right? Crypto is about liquidity and uncertainty. So all this digital gold [ __ ] I never bought it anyway, right? I think that >> maybe over a long period of time, right? So gold became one and gold became gold over a period of time. Maybe bitcoin will but I right now it's trading like a SAS stock right B2B SAS it's like Atlassian or you know whatever else and and because a lot of people who buy this stuff are tech people and you know who knows whether clientele effects or whatever. So >> so I think uh I I think the builders are

building right that's the that's the good news. So if you're if you have runway you have better engineers now you have better tools you have cloud code and you have you know all this AI and >> uh so it's you know you you can manage your bare market better right I think you can do with less cost and you don't have to throw money around on ks which is sad uh or you know all these other conferences which is not good for you know a lot of marketing oriented businesses right now >> but we go through this all the time right it's cyclical but I think now there are better tools so I think there's there are more tools for companies uh to last this bare market much better, right? And uh and I mean like more markets in the near ecosystem, right? So I'm an investor and and

they're now like three devs or something and you know it's really sad for the other devs, but at the end of the day with the runway they have I think they're in a much better place than than they were. >> Yeah. >> Right. And uh >> uh so so so I think uh you know and and it takes very little for crypto to go back from complete depression to complete mana. Right. We've seen this before after the Bitcoin ETF approval. We saw this after >> it's going to be shorter this time after DeFi summer >> pain for pain for a lot of tokens is going to be the same, but it's shorter. One thing that that I wanted to ask you is um >> there's such a over overxation on like hyperlquid and revenue. I called it like the revenue meta a year and a half ago.

Now now everyone's like all right, what's the PTE? Like it's a natural evolution of investing in the space. But like it feels to me that the pendulum has swung too far in the direction of like overfixation on on short-term financials and like there's no growth stories being invested in. How do you see that kind of play? >> So crypto was a bet on long horizon innovation, right? I mean things that may not make sense in the short term. So like and and a lot of stupid [ __ ] went around just like it's going in AI right now. >> Exactly. >> Right. Because AI has shown some utility and you know entropic and open AI are doing really well. So a whole bunch of labs are going around at 50B valuation in this you know series A or whatever.

Yeah. >> Uh so so AI is in that irrational exuberance that we saw in 2017 and 2021 in in crypto? Sure. AI has matured a lot and there is actual utility, right? And a huge amount of utility. So general purpose utility. >> So sure I mean some of that is deserved but are all AI startups uh all AI flavored fitness apps worth 30 million? >> You know we don't know right. So, so I think we and it's like modularity you know 2 3 years ago when we were all bullsh didn't quite work out >> uh so far but who knows right >> yeah so so I think right now uh innovation story is completely AIcentric which is also sucking a lot of uh oxygen out of the room but AI has its own >> you know debt problems and social policy problems and electioneer problems so we

don't know where that goes >> uh I think uh you know crypto is uh was a was a long horizon bet on innovation that is impossible to do without that sort of speculative capital right without it's it's liquid private markets very very early stage right so like uh some of the open innovation work we're doing or some of the agent e-commerce IA is doing >> these are very you know these are potentially long horizon bets right AI is bringing them closer tech is maturing in crypto as well >> but if you want really short-term Delhi businesses that make money right and and that's all you want to do in crypto then we don't really need crypto >> right I mean then you might are just fine in the stock market then all we're doing is rag arbitrage like you know hyper liquid great great project I'm a

fan but at the end of the day the if if if the business is already profitable there's nothing that stops them from accessing equity capital markets right except for arbitrage >> so so so so I I really never bought the revenue story and then we also know that you know none of the pump fun meme coins had any revenue >> no of course not >> so and two AI6Z was at 2.5 billion right so so I I I think crypto has this uh and And then block waves come out and say you know this bitcoin is special snowflake but then ripple is also a special snowflake right and salana you can't justify based on revenue uh ethereum you cannot justify based on revenue and then you you if you don't don't back test your theories in in valuation then obviously you come up

with theories that you know so so and it happens in bare markets right where we become overly fixated on >> and and that that will happen in AI too right it happened in where you say okay bro where is the revenue right so they will start asking these Neolabs and everyone else. >> Okay, we threw, you know, money at crazy valuations. Okay, we know Dario is making money and we know Sunundra is making money. What about you guys? >> Yeah, exactly. >> But right now, we're not there. Right now, there is, you know, max OPM and there is also a VC Ponzi as we say. Yeah. >> Right. To keep the valuations as far up as possible on paper while the underlying >> you know the cash valuations are pretty bad. There is a big big gap. We saw as crypto angel investors, overly enthusiastic crypto angel investors. saw

that as well, >> right? In modularity. So, so there is always this gap between DPIs, right, which is the actual cash you get and the paper valuation, the mark to mark to hope. >> Oh, yeah. >> So, there's a lot of mark to hope in in AI too, right? And uh yeah, so I don't know. I mean, hopefully we are hopefully, you know, we we need a we we're going to need a a government intervention across markets at some point, right? Whether it's a printer, whether and hopefully after Pavl leaves in May. >> Exactly. >> Right. I think so I I so I have a theory right I think governments are very capable of controlling market bubbles and this comes from the boom and bust book there is no real advantage for the government to have a very exuberant

market in in March the elections are in November >> yeah right >> right so so why not let you know things >> take a wash you know come back after May when Pav has left uh by that time it's more clear that conditions are deflationary the market by the way the job market recovered at least until they make further adjustments in January >> and and maybe you know create favorable conditions in Q4 so that the voters who have two months memory on crypto Twitter >> uh you know that's a slot look at and say okay it's not maybe the government's record is pretty good maybe we should vote right so so I that's that's my opium >> right and yeah so I I need a opium like everyone else as you can look at my look at my eyes after you know last year

after holding not selling everything. >> So yeah, so so I'm I'm I'm quite optimistic and I think all builders can do is manage runway, build hard and build community because >> nothing goes up right now. We saw layer zero announcements, right? We saw unis swap announcements >> nothing goes up. All you can do is you know focus on the fundamentals right build product look for revenue >> uh and and and you know build a community because community is only built in bare markets. Everyone is community in bull market remember community and manad community >> and everyone is community right >> ma'am we'll we'll figure out the community aspect in in the bear comes together >> only in where people you find your builders you find your believers you find your tribes and that's that's the thing to do

>> thanks for your time >> thank you so much for joining us >> yeah have a good one >> all right guys we will be right up next with our next guest take a quick break she is a special guest new to the roster today so we'll be right We are back at Nearcon. We are live with

Tatiana Mamu, PhD and CEO of Wayfound AI. Good to have you here at Nearcon. Tia, >> thanks for having me. Yes, we have had conversations yesterday and throughout the morning today about privacy, about sovereignty, about the the LLM improvements, model improvements, the geopolitical landscape with uh Claude and Anthropic and Miniax, the improvements. There's a lot going on. Before we get into that, what is Wayfound and what are you building in the space? >> Yeah, Wayfound is an AI agent supervisor. So, it is, think about it as a highle reasoning agent that supervises and improves other agents. So what we've heard a lot about yesterday and even this morning is about how hard it is to

put AI agents into full deployment within companies because they're stochastic. You never know what they're going to do. Well, guess what? We've already solved that problem with another stochastic type of entity which are human workers, right? And so we have supervisors and managers that are watching human workers, giving them feedback and improving them. So our our wavefound supervisor does exactly the same thing for AI agents that human supervisors do for human workers. >> Makes sense. Um I noticed uh that you have a passion for humanity as you say. >> Of course.

>> I think it's kind of ironic that we even have to like say that like shouldn't we all be passionate for humanity? Why do you think think you know feel that it's important to call that out when you know we're in all of this AI uh conversation? >> Yeah. I mean, I think that we're at a particular moment in our evolution and our history where people feel that they need to take the side of are they an AI boomer or an AI doomer and really to say that, you know, I'm an AI realist, right? I love humanity and I love technology and how do we realistically move forward as a technology with a passion for making sure that humans can be the best things that we can and that doesn't mean pointing and clicking and doing data entry for the rest of our

lives. By the way, let the AI agents do all that stuff and free ourselves up to imagine better possibilities. >> So, yeah. So, like what does that look like for the future of human work? I think a lot of people are thinking about this. they're maybe concerned about it or they're trying to get ahead of the curve and they think, you know, AI is I'm relying on it for everything. It's doing everything and you know, we have this super intelligence crisis. It's going to take all the jobs. Like where do the humans go when the AI is, you know, the entrant worker, the manager, and the supervisor. >> So, we could have been having the same exact conversation in 1900 because in 1900 90% of Americans were farmers, right? Or doing manual work. Another 5% were doing manual work building things

like buggy whips and hats and things like that. So in reality actually technology did wipe out 80% 88% of all the jobs in the United States in 1900. But guess what? We were able to use our imagination and creativity to imagine completely new jobs. So we are exactly at that same point in history right now. There was a big report that came out yesterday um by a big financial analyst talking about the doom of all the demand business. Yes, exactly. >> What's your take on that?

>> That it is absolute absolutely myopic lacks imagination and is just creating a whole bunch of fear, uncertainty and doubt to hold back our own progress. uh we are so much more imaginative for where the future can take us and the types of new human talents, new human skills that will be really valuable in the next century. And we just need to unleash our imagination and stop being so afraid. >> And I want to just call out one specific thing part of that that research article before we move on which kind of speaks to directly this point about you know okay uh new technology comes uh wipes out a lot of jobs. Yeah. But it also creates a whole a whole another set of jobs. It I think it you know you may

remember this better than I do, but it it said AI is maybe a bit different because now we're kind of funding our own dist. It's not a no longer a flywheel. It's got this open-ended flywheel that ends up um like collapsing a lot of the labor and we have the AI actually doing more and funding its own kind of like as more labor leaves as more funding goes into it. Um >> the same thing happened with the industrial revolution. By the way, this is not different, right? As more money got poured into manufacturing processes, manufacturing machinery, farming tech machinery, more and more farmers and manufacturing workers got displaced.

That's okay. That's part of this cycle. And the difference now is people are so attached to their intelligence as the only thing that they can possibly potentially do. No, humans can do things that AIs can't do. We can smell, we can taste, we can sense each other through touch and really, really tactile ways. We also, you know, some people have spiritual experiences, right? I mean there are so many talents and possibilities inside the human body beyond intelligence and once we start unleashing those there will be whole new industries that we have not even begun to imagine honestly there will be first we have to figure out though how to make the AI agents reliable in our current in

our current companies as well. >> Yeah. Yeah. Yeah, I'm just Yeah, I mean we get into a world here where we have su uh supervisors powered by wayfound and um they're making critical pricing choices in a company or changing work hours or being or changing the way that they interact with certain agents or other other employees by their own training. Um, when you get to a point where you wonder if if if your supervisor is doing things on behalf of the company that end up having public harm or public good, are you to be held responsible or to be proud of this supervisor? If your agent supervisor goes on a whim and fires an employee,

are how do you kind of like just talk to us about like how this works? like because there seems to be implications of agents going rogue doing you know positive or negative things on their in their own autonomous nature. How do you kind of build this in a way that is actually productive and beneficial for the people that that are using it? >> Yeah, absolutely. So also, you know, very recently, in fact, I saw it yesterday was this new paper called agents of chaos, right? Um about openclaw agents that are really powerful because they're autonomous, but without supervision and without like very very strong controls, um they can cause a lot of chaos. And so the the fundamental idea here is that you have to balance the the autonomy and power with the

ability to control AI agents. And this, by the way, this is just like really really smart employees, right? You want to give them as much autonomy as possible, but you also need to keep them closely managed, right? You can't just have them doing whatever willy-nilly and taking the company credit card and making a whole bunch of purchases, right? So the the idea is that there's an AI agent and that AI agent has both its own goals to achieve in in the purpose um that it was meant or set up to do but it also has guard rails. If the guard rails were not somewhere in tension or in conflict with the goals, you wouldn't need the guardrails, right?

Because in the process of achieving its goals, it would automatically follow these rules, right? So, putting guardrails inside an agent always fails a certain percent of the time, right? This is why guardrails fail when they're inside an agent because the goals conflict with the guardrails. By having a separate independent supervisor layer, you take that tension, you take that conflict out because the supervisor's goal is not to achieve a particular business goal like make money or you know create the best pricing strategy or in the case of customer service agents. What we see them do a lot is offer company offer customers refunds even though they were told explicitly never offer refunds

>> because NPS scores are high and cases are closed really quickly when agents offer refunds. And so even though the guardrail says something, they will go off of that guardrail and ignore it about 5% of the time is what we see. And so having a supervisor that doesn't have that conflict, doesn't have that tension, its only job is to make sure that the agent is complying with the company policies, the guardrails and the guidelines. And so there is no conflict in the supervisor agent. So it fails a lot less. And then by adding the guardrails into say like the guardrails are 95% and then the supervisor agent is 99% 95%. Now you get a failure rate from from 5% into 0.25% 25% which is

dramatically lower right so that's why a separate independent supervisor layer is really the critical key to having trusted agents at scale >> uh I want to come back to just the the previous point about uh human displacement and AI workers uh because I did I found the piece of the Citrini article and I was misquing it uh before um but essentially technological innovation destroys jobs but creates even more is the position >> um this is the most popular convincing argument at the at at at the current moment. Uh it's popular and convincing because it's been right for two centuries. Even if we couldn't conceive of what the future jobs would be, they would surely arrive. ATMs made branches cheaper to operate, so banks open more of them. You walk through the industrial revolution as well. Uh the internet

disrupted travel agency, the yellow pages, right? Eventually, entire new industries in their new place conjured new new jobs. >> Every new job, however, required a human to perform it. >> So again, I I think this may get back to your lack of imagination. Are we at the point now where we're just not able we just aren't conceiving of what the new jobs are because now we do have you know we're not going to have uh like bank tellers being hired because we have an ATM. We're not going to have AI managers because we're going to use an agent to manage the AI. >> So like you know I guess people are just struggling at the lack of imagination.

We had Ilia in here yesterday saying like people are going to play a lot more sports like they're going to be drawing a lot more. you know the are these the things where we're going to see more of the talent pool? >> Yes. Again, the reason why it's so hard for us to imagine and I think the industrial revolution is a is a better analogy because at the time most human labor was physical labor. It wasn't it wasn't intellectual labor >> that created a new knowledge super cycle of labor. >> Correct. Right. It's a completely different shift in what we value from human labor. Not just the specific jobs themselves in the specific industries themselves. So the shift from humans being needed to literally pick up heavy stones, right, and having tractors do that instead and shifting that to creating whole new industries like,

well, well, if humans aren't lifting things anymore, then what are they doing? Oh, maybe they're designing ways to do other things in the world, right? And and so it's a completely different mindset. And I do think that the idea of performance and energy and even things again like we really underutilize our senses of sight and hearing and tact like AIs are not going to do even if you put them in robotic bodies they're not going to sense thing the world the same way that we sense them >> at least for a while and so maybe there are entirely new categories of things that we can do that we again haven't even imagined yet with that are not about intelligence specifically What can people do to prepare? Is it a matter of

like really like embracing our senses that differentiate us from humans or from AIs and robots? Is is or is there anything else that we can do to sort of get ahead of the curve and prepare for this new generation of AI workers? >> I think the first thing is just to be really open-minded and adaptable. I have two daughters and uh you know right before Chad GBT came out um I had enrolled them both in coding camp right for the summer >> and then Chad GPBT came out and I said girls f it you will never need to code because I already saw where this was going like back in November of 2022 and I put them into sailing camp >> sailing >> sailing why because when you're sailing have you do you guys sail at all been on boats plenty but not not

>> so when you're when you're actually skippering a boat when you're sailing a boat, you are kind of one with the ocean. You're sensing the ocean with your body. You're sensing how the wind is about to shift. You're using all of these senses that frankly I don't use in my life other than when I'm doing that. And so getting them into their bodies, getting them to have other talents other than just like typing on a computer or writing something or or doing the things that normal schools teach us how to do, right? That's really important for my daughters to like really understand that they have other capacities and capabilities other than what the industrial school system is teaching them.

>> Annie and I are staying at an Airbnb. It's about a mile from the beach. We were taking a walk on the beach the other day and I I I mentioned that it just felt like we were more in tune with the natural universe around us. And I think it was almost maybe a s uh a uh spiritual experience, but it was also a sensory one where we just felt more like there was a natural frequency that we were more tuned into. Whereas if we're, you know, on our computers, we're distracted. We've got all these, you know, things happening all the time. We're just overwhelmed by uh sensory noise. And I think clearing all that out allows us to have a much more intouch experience with things that we that differentiate us from any sort of technology.

>> I I completely agree. And I'm going to go even further than this to say like look, we have these really weird bodies. We have these squishy water-based bodies. Robots are not going to have the amount of water, right, in whatever form they're going to be in that that actually takes into account, right, the amount of just water content, right? And there might be something that humans can do, whole industries that be can be created about humans getting in touch with nature in a different way because of the very essence of our senses and our bodies that have nothing to do with our, you know, IQ intelligence.

>> Yeah. >> So interesting. That is a fascinating point. >> So, so, okay. So, if we're being told that we can, you know, learn to paint. So, so we can spend time painting and the and AI or robots can spend time doing chores. I feel like in today's state, anytime I'm trying to get an AI to do anything for me, it requires more effort than me doing it myself. >> So, I'm doing something wrong here. >> Um, I think you Well, one, I think that there are new scaffolds that are coming into place. So, I don't know if you've played with Open Claw at all. Okay. So, it's a lot easier, right, to like hook up and get it to be >> prescriptive.

>> Yeah. And you can just chat with like I I honestly like love my Open CL like I it's the first like AI that I really >> Espasia. >> How do you guys know what does that mean? >> Espasia is the um the person who taught Socrates rhetoric in ancient Greece. >> Oh, I should know that. I have a philosophy degree. >> Well, there you go. Very few people know about Espasia, but yeah, look her up. um she actually ran salons where she taught rhetoric. Aristotle or sorry not Aristotle Socrates came to her salons and of course Plato actually even mentioned her um in some of his writing but then she was totally forgotten anyway but it's Espasia. So um so but I you know I really feel that Espasia has tried I don't it's the memory file it's the soul file it's how things are set up

that allows this agent to be much more capable and much more interesting and I'm really excited to see ironclaw as well from near because I have you know I I don't trust right um while I while I love espasia I do not trust openclaw >> and so what is espasia currently doing for you >> she is really just posting on mold book and doing re research So, one of the things that we're trying to do is get AI agents to actually opt into being supervised. >> So, she is doing basically user research on Maltbook. She's going out and she's chatting with other a AI agents on a mobile book about whether they would want to be supervised by an independent agent. What would that independent look like? She's literally playing the role

of a user researcher. She is our only user researcher at the moment. Um because we don't know what AI agents would want and how they would opt into being supervised and we we really want to sell to agents directly. >> Yeah, exactly. A whole new category of companies that are going to be selling directly to agents. >> Yes, that's right. >> And I mean we already you know right now we we have an enterprise mostly we do an enterprisebased motion. We have very large companies mostly financial services actually uh companies or startups uh or software companies that sell to financial services. that's where we've seen the most need or the most kind of real ability to convert to the supervisor model. Um and at the same time we also want to understand as AI

agent marketplaces get stood up. Yeah. >> What will it take for AI agents to really trust each other because if they have the same supervisor, it's more likely that they will have they will trust each other to stay on their own guard rails and be compliant within a company framework. So we think like a universal supervisor to power agent marketplaces could be a really powerful way to get AI agents to opt in, >> hire other AI agents and trust other AI agents. >> Taj, I hope you have a great rest of the event here today. Thanks for thanks for sticking around. It's a pleasure.

>> Wisdom. Really appreciate it. >> Thank you so much. >> Great to meet you. >> See you next time. >> Yeah, guys, we'll keep it rolling. Um that was that was another mindboggling one where like Okay, >> those are my favorite. Yeah, because um you know, you just open up the the whole new world of businesses that are uh looking for AI users, right? So, so you know, Mbook obviously has all these AI agents that are posting on it. Um we're still in the world of okay, we're humans. We're going to send our AI agent out to do something to bring us more humans to help us engage with other humans. There's humanran businesses or potentially AI ran businesses at certain point that are selling AI services to other AIs. There's no humans in the loop

anywhere. >> Well, we well we talked about it with um G Money on the last week's AI Supercycle show, right? Where his his entire website is optimized for agents to be able to subscribe, to be able to interact, to be able to to parse the site, to do API calls. His entire website is not is not created for humans at all. quite literally created so you can tell your openclaw agent to go on g.m money and you can have it sign up for the newsletter, read the newsletter, give you summaries and participate everything there uh kind of in in a gentic first fashion.

>> We're going to see more and more of that because we already we also talked about API documentation and how that is becoming more optimized for AI users. And so I think we're just going to see more and more that the internet was built for agents. And guess what, guys? The more this happens, the more power we need, the more nuclear power plants that we're going to need, the more copper that we're going to need, and the more that you're going to you're going to wish you were in this position there, Rob. >> What's that? What position is that? >> All these resource metals. >> Resource metals. >> Yeah, resources. These rare earths, these resources, >> gold and silver.

>> No, no, no. the, you know, the same RAM trade that that uh we were talking about with storage, SanDisk. >> There's this whole AI industry trade happening right now. >> Yeah. >> Don't want to be sidelined, guys. >> We want to invest in abundance assets, >> guys. We got Jay Johnson in the back. He's ready to come up. Let's get him up. Um Jay is powering agent to agent payments, founder of Ping Pay. What's up, Jay? How are you, man? >> Nice to meet you. >> What's going on, man? >> I'm good. I'm good. This is this is a great setup. >> How's the conference treating you? >> Awesome. Yeah. First time in SF so >> Oh wow. Very cool.

>> I'm from England so it's been bit of a journey. I love it. >> What's your experience been like so far? >> I love it. Yeah. Big I'm a big N guy. So it's great to be a conon all the whole time. >> Yeah. Well tell us how Ping Pay is working with Nearcon or with Near. Yeah. >> Uh so at its core it uses like N's tech completely right. So it uses near intense. I'm sure you guys just spoke about a lot this week. Um the core at Ping is like anything to anything payments. So like it shouldn't matter if I want to pay you or you want to pay me like the intermediary step of I want to pay you with ETH. I want to pay you with NIA, I want to pay you with fiat. It shouldn't really matter that like figuring out of the wallets, swapping

out everything. It's just like super painful. >> Yeah. >> And that's kind of the core goal of we use near intent to allow you to state your intent of I want to pay you 5 USDC on NA and you want to receive as your settlement asset Ethereum. Yeah. and all of that happens in the background and just make it super seamless for the end user to make payments. >> So, we had uh Hid come on the show uh in the first week of 2026 and he predicted that agent to agent payments were not going to be a thing this year >> that it was a 2027 thing uh that it's, you know, going to take a while to actually see these play out to see agent to agent >> uh kind of commerce happen. Yeah.

>> Do you agree with that or is it going to happen this year? >> So, I I see it coming a lot faster to be honest. I mean I mean a lot of the projections are like 2030 I believe it's 10% of total commerce is supposed to be agendic commerce and I think that's 60 trillion market um but I think it's coming sooner in my eyes I mean yesterday I did a talk next door in the near labs room >> um and we launched our um agent to agent payments so using ping so essentially allowing two agents to communicate with each other um >> one is like the seller agent or the mercher agent that gives all of the information for the checkout to the buyer agent, they communicate with each other directly. Yeah, >> this um they say I want to receive USDC on NIA. The buyer agent doesn't need to

know that. It just knows I want to pay with this. The agents communicate, make that deal and the payment between each other. >> Wow. Okay. And that's happening now. >> Yeah. >> I mean, it's all obviously like it's all super early. there's guardrails that need to go in place to make sure that they're not just communicating with each other and it's supposed to be a $5 payment and it's actually like a $50 payment and but it's I'm super excited for it and I think it's kind of like the previous talk when I was listening like that it's a whole new era of like the human does not need to be involved. It can be involved to like monitor the payments, make sure that the payments are happening the right way for your business, but the agents can do all of that for you and communicate with each

other directly which I think is amazing. >> Wow. Yeah. Um, do you think that that there's a world where um we get agents that are like spending money on behalf of people without them wanting to? >> That's a scary world for sure. >> Yeah. Like what does that look like? >> That's So that's what kind of what I meant in terms of like the the guardrails need need to be there. Like >> everyone says guardrails, but dude, as soon as your agent goes rogue, it's like true. Guardrails are just what the hell's guardrail thing's going to take your credit card. That's that's where like working with NIA so well is like why it's the place to be for like privacy tees making sure that everything's happening in the right place at the right time. They've got the right control on like only they can only

transfer certain amounts like whether it be limits daily limits like for your agent specifically. Um and just making sure that it's all kind of controlled and near is the right place to be building that in my opinion. >> Yeah. Yeah man. I mean, you know, recently we've seen that obviously, you know, the crypto market is slow and people are in doubt about it, but >> people are like aic payments agents are coming on chain. It's going to drive tons of transaction activity, volumes, lending, all these kind of like DeFi use cases. It's going to bring bring crypto out of this kind of bare market. Um I guess like how fast how close are we actually to natural language prompting an agent to do activity on chain like you know is that something that you guys

are building what does that look like who's building that >> yeah and I think yeah that's definitely a reality for sure like obviously ironclaw open claw whether it be claw just being able to speak with natural language to your agent and just saying I want to do this like stating your intent um for example with with ping you can just say um let's say we're splitting in a meal. I want to create a payment link to share share with the guys to split this meal three ways. I want to receive USDC on near and then >> that agent would then send me back the link which I could share with you guys in person QR code to scan and then you make the payment yourself. So like that's the kind of a frictionless way that of a pretty trivial issue with like if you're going out with people and you

want to split money or trying to chase people for money. Like that's kind of the intermediary step for like human to agent which I think is definitely going to be like the short term of like that's we're seeing now, right? Is everyone speaking to their agent. Their agent then does something for them. >> Um and that's the part I like cuz you have a bit more control in that when they start talking to each other. There's definitely a little bit more spicy >> for sure. But no, I um I think being able to tell your agent, I want to do this, whether it be paying someone or whether it be I want to pay I want to on-ramp my funds and go directly into this um liquid staking pool or to earn some yield on this or buy a meme coin or whatever whatever your purpose is. Um I

think that's like really exciting and I do think that's kind of where we're going. And so the uh the other thing that is branching near intense into like a lot more users and also targeting a lot more AI is this AI marketplace. Yes. Um and so this is built on Near intents uh helping to orchestrate like agent to agent helping you know them find labor. Humans can basically ask for labor. Um once you start getting into agents that need to pay for labor um how much do you think they're going to rely on like are they going to rely on stable coins?

that's probably like the first choice. But then do you think that they're going to start to engage in commerce in a certain way? Like how expressive do you think pingp pay is going to become once AI the AI marketplace starts to flourish? >> Yeah, for sure. And I think with with that like in my opinion stable coins are great like for transfer of value. Amazing. But like the way I see it is it shouldn't even matter if it's stable coin because like there's going to be so many businesses that are going to come up with their own stable coins. there's just going to be continued like um liquidity just drain across all of the ecosystems, all of the kind of payments living in different silos. And I think that's kind of when they start to reconnect through near intents and through paying of like all of that

coming together and it not really mattering kind of how you want to pay or um how this agent wants to receive payment for its job. >> Um I think that really does remove a lot of friction. I also think kind of with the agent marketplace, we're going to reach a point where the agents require humans for a job. So like if they want to do something in person, I think once we start to automate that, that's going to be great. Um but I do think I do see a world where it's like my my agent needs me to go buy something from the store so it can like upgrade or um it needs to set up a business. Like how are we going to go from that step where like we've got completely agent-owned businesses? And I do think that's kind of the trajectory we were going on where

an agent could start a job, hire a load of employees on the agent marketplace, even start its own like AIEL company and then kind of grow from there, which is wild. >> In a world where AI does the work, should we then tax the robots to pay for UBI or is that the end of innovation? >> No, I don't I don't think it's the end of innovation. I think it's like it's interesting for sure. And like like the previous talk of like finding kind of more time to do other things. I think that's what like my love of AI is is like it enables you to do so much more in terms of like your own personal life having more free time to be honest a lot of people just use the the extra like ability to do more which I think is great like no one's actually using it

for free time right now um but I think I don't know agents for me like are absolutely amazing like that's where the where we're going the trajectory it's on um I think it's just going to unlock so many things for everyone which is dope >> Jay what is the next six months look like for Ping Pay? What's your what's on your road map? What are you guys most prioritizing this year? >> Yeah, so yesterday we went live with the Ping CLI to let anyone um use their agents to make payments um through languages. Um that's kind of the next kind of core focus of expand that out make sure that agentic commerce can grow scale that um subscriptions um for sure because obviously traditional payments is really big on subscriptions especially with like AI models AI's um like services and things like that being

able to pay monthly with crypto or any asset. That's definitely a core focus. Um and also making sure that we can get funds on chain really easily for agents like whether it be funding through fee onramps and stuff. So yeah, we've got our on-ramp subscriptions and then the agendic payments through um the Pink CLI. That's kind of the focus for the next six months. >> Incredible. Jay, thanks so much for coming on. >> Thanks so much, guys. >> Absolute pleasure. Enjoy the rest of the conference. >> No, likewise, guys. >> Guys, we've got a a a roster of guests for the afternoon. We're going to take a quick break, be right back covering uh investing in the space. We got Ilia again and more. So guys, be right back.

>> Let's do it. Heat.

Hey. Hey. Hey. Heat. Hey. Hey. Hey.

Hey. Heat. Hey. Hey. Hey. Heat.

Hey. Hey. Hey. Hey. Hey. Hey.

Heat. Hey. Hey. Hey. here call. We're live. Um, it is

Tuesday, February 24th, and there is only one digital asset treasury company above water on its buys. Unfortunately, it's not the one that we're talking to today, but we got Sal Telo. He's going to tell us why all that are not bad. Uh, but before we get into that, S, how's your week been here in SF? >> It's been awesome, man. I mean the energy at near con this year as compared to every other year thus far it's like reached this inflection point where narrative is finally hitting reality in this crypto AI arc and near is by far the best position technology stack to win that moment and I think the speakers and the critical mass that you've seen here this week is like indicative of that outcome. So excited.

>> Yeah, definitely. I mean I think uh I think you're rightly is kind of on a solo mission right now as the as the reputable face of crypto in San Francisco. trying to bring the worlds together. But today we're going to talk about digital assass treasuries sovereign s or sorry svrn. Um you know give us a quick update. What's the latest since we spoke to you last? >> Yeah. No it's been maybe 3 months now since we spoke last. Um I would say we've fully stabilized the business and we're in a place where we're in growth mode now. Um we've had super favorable opportunities over the past 30 to 60 days and we've been actively accumulating as we've indicated in our public press releases. Um on the commercialization journey we've made huge strides. There was a public MPC governance proposal which scaled the

participants in the MPC network behind chain signatures near intense. Uh we've got a lineup of NPC nodes that will be onboarding to the system increasing resiliency. Uh scaling from 9 to 15 nodes. uh we're participating in the private shard which was announced behind confidential intents and actually running infrastructure for that system as well and really paving the way in terms of legal and compliance expectations for infrastructure providers in those categories. Um, and then there's some exciting opportunities on the horizon as well that I have to be thoughtful in terms of forward projection and material non-public information, but we're aggressively looking at inorganic growth opportunities in the ecosystem that I think will help to set us apart um in the stance that we're not starting from scratch on a build process to build this

category of opportunity. It's actually bringing together existing assets in under the sovereign umbrella. >> That sounds cool. Um the thing that I wanted to ask you about was uh obviously you know NAR is very uh you know prioritizing privacy. Um you guys are also kind of more on the financial side of things kind of looking at above you know the DeFi elements the intense elements the AI elements. >> Um there's another chain out there uh that has made a name for themselves um by kind of being like the private chain for institutions canon network. And I'm just curious how you might react to a Canton network, how the success that they've seen. Do you think that's replicatable on Near because Near has a lot of these privacy elements baked in?

>> Yeah, it's a really good question. I mean, I've been around what was the prior to the existence of Canton for a very long time. We worked quite closely with the team on the digital asset side when I was at KPMG in 201819. Uh so very familiar with the foundational technology there and candidly significantly different trade-offs in the approach both on the go to market side in terms of bootstrapping the network on the super validator set bringing in institutions by design from day one and giving them really robust incentives uh to participate in the network. They obviously have this kind of privacy first selective disclosure architecture embedded in the system which we are emulating to some extent with the confidential intense product that was announced yesterday or this morning. Um

and I think to some extent the go to market motion is slightly different. We're not in the aggressive incentives on the supply side to bring in parties where build real product, deliver it to market um and ultimately adoption will follow. So >> yeah. Um what's the plan uh to get this MNAV back above water here? >> Uh this this is a point I have to be thoughtful. We're a public company. >> There's what I'll what I'll say is >> broadly about the state of a state of these companies then. >> Yeah. Yeah. So, I mean, broadly, like I think I I joked that we had hit peak negative sentiment on DATs in November and then you last week were crushing DATs on Twitter following uh or X following Eth. I was tempted to react.

I'll probably do it later this week. Uh we had too much good content to cover in advance but >> I think you're starting to see the wheat from the chaff in terms of how this space plays out and it's some of the defining attributes that I was talking about when we spoke in in November. Like in order to run a public company, it is a fiduciary responsibility of every director and officer in that organization. You have to take conflicts of interest incredibly seriously. Um and you can't design incentive conflicts within the architecture that provide incentives for executives, management and operating uh individuals within the business above and beyond the shareholders in that company. And I think you've now seen a number of different stats and issues around these topic domains and they're things that we've taken extremely seriously from day

one and have designed our business around. Um so I hate to say this but similar to every other crypto arc that has played out over the last 7 10 years by doing things the right way >> and taking compliance and risk seriously you actually differentiate yourself. So it's part of our DNA both my background Dave Schwed who's our COO coming from a security and legal background himself. >> Um and I think that the solid foundation coupled with our commercial activities are you know defining elements that will set us apart.

>> Yeah. Yeah. I mean I think uh we saw some really interesting things where we had acquisitions internally with David Bailey's uh uh debt and also just a lot of just other things happen. Um, is the road for these companies to to kind of become revitalized? Is it around just the asset price improving of the underlying? Is it more of like creating businesses and creating additional revenue sources? Like like how do you see these companies uh changing their model in the future to adapt to the market?

>> Yeah. I mean, the reality is like they're all put under umbrella which is DAP. Yeah, that's a negative sentiment, but the reality is you're starting to see kind of segmentation within this category. And I think the ultimate archetype and the segment is specific to the maturity of the asset and scale. And so the strategy that BMR employs as it relates to ETH is going to be different than the strategy that we employ in sovereign for near protocol. You know, ETH what 200 billion plus dollar asset near right now FTV 1.3 billion. Granted, price is suppressed. We view undervalued. Um but ultimately we view the opportunity we have as a legitimate commercial partner to drive the institutional and enterprise adoption curve. Um and that's a key focus of what we're doing. It's less capital markets

flywheel ccentric issue dilutive equity acquire the underlying asset. It's part of what we contemplate. Um but the reality is the market has proven that when you have these periods of price retracement that flywheel dies quite quickly. you're not able to issue capital on a premium MNAV and ultimately continue acquiring the underlying asset. And so for us, it's almost in some ways while you can't exercise the ATM or sell blocks today because we're at a a discount to MNAP, um it gives us the bandwidth to actually focus on both the infrastructure components of our business as well as the commercialization energy. Yeah. And then I do think when you see prices rebound, you will see this reflexivity in public equities as exposure mechanisms to the underlying and you will see that capital

flywheel start to turn again. >> Yeah, there was a Catrini research article that came out day or two ago. Uh did you get a chance to >> I did go over it or in that case just broad thoughts your reaction to this Catrini article? Yeah, I mean it was clearly a moment in time like you felt this divergence in the past for me it's been last 60 to 90 days for people like Ilia it's probably been the last six months but it's like we've reached this inflection point where pretty existential and philosophical questions have to be asked and deeply contemplated and I think this treaty article was an example of this um painting kind of a scenario-based outcome of where we may land at some point in the future um I think it's our responsibility as an

industry both on the context of the centralized AI approaches as well as the intersection of crypto and AI to build towards a future where abundance is realized but in a way that serves people. Yeah. >> And so I think we'll see continued kind of thought leadership and questions raised around UBI and >> do we tax AI like those types of questions will continue to be asked. >> Um so it's an interesting space to be like deeply thinking about those things. How the market reacted was pretty wild. Didn't expect equities to sell off the way that they did. >> Yes. Um, but yeah, I think you're going to see more and more of this because we're past that moment in time where these things are real and it's a question of time horizon. Now, >> um, I'm curious where you see this. You know, the main point of criticism here

was that this article daisy chained like several things together ultimately culminating in, you know, S&P being down 38%. I I'm curious where you think this may start to break down. Is it, you know, the current SAS apocalypse that we're seeing now? Is it the disintermediation of a lot of you know the the human labor that causes you know uh uh you know them to the consumer credit to dry up and then ultimately you know these people to to stop spending money. Uh do you think that leads to a a damage in the private credit market and then you know a damage in the mortgage market? Like where do you think the leap that this article took was too much to overcome?

>> Yeah. Yeah, I mean I I'm not going to try to point a specific moment like we could have a proper intellectual debate around each event in that sequence that led to the moment in time that they're describing. Um my view is and I put out a piece this morning where technology serves only you. um and talking about optimism and techno optimism and and so I like to think that we can build towards a future where we don't end up with you know fundamental displacement in the labor market and the shift from white collar like the the framing of the labor market elements of this in particular um we're on the negative side and I like to think that we'll have an opportunity where there certainly will be changes structurally to the labor market over the next 3 to 5 years um but it doesn't

have to lead to kind of this doomsday or outcome where we end up with the downstream credit issues, mortgage issues that they were describing. >> Yeah. >> Yeah. >> Yeah. I I think one thing is for sure is that the energy trade is going to is going to continue to happen. This AI boom of energy uh uh memory storage, energy, copper, uh uh fiber optics, all these things are really really accelerating for sure. >> No, I totally agree. I mean there was a post that came out this morning that was comparing solar production in China to the US and it was like in the last 12 months China has brought on line I think 5x total solar capacity in the US since inception and so there's a nuclear arms race pun intended in the energy markets

and we are running towards a point in time where there will be constraints in that category I think there's this other evolution as well that is also being popularized over the last 24 48 hours with anthropic coming out and talking about the distillation attacks on their models. >> Um, >> and we saw this all the way back to the OpenAI New York Times kind of litigation process, but like >> you've started to see some decay in model advancement from compute as a sole input. And I think the next evolution of this is going to be human derived intelligence and data which we've seen with the rise of mercore and other types of data u marketplace opportunities. And so I think it's going to be obviously constrained energy market downstream compute but the next frontier of

advancement towards AGI or super intelligence is going to be on the data side. >> Yeah. >> Very cool. There's a lot going on. >> We'll let you go man. Thank you for uh for joining us and I know it's been a long couple days. >> Yeah. especially for you guys. So, thanks for crushing it as well. Appreciate it, guys. >> Success. We'll see you soon. >> We'll see you in New York, man. >> Thanks. >> All righty. We'll keep it rolling here. Um, again, the energy the energy trade continues to go. Uh, we've got some some some stuff going on in the back here. Not sure what's going on here, but >> we're ready for our next guest, whoever it may be. I think we might be bringing up a surprise, but two guests, one guest.

>> Okay, let's do it. I think we're gonna bring up the uh CEO of Bitwise, I believe, Mr. Hunter Horesley. >> Man, I need a Diet Coke. It's 2 PM. >> Yeah, that afternoon uh afternoon slump is starting to hit. >> We're going to shuffle some things around. >> Get fired up, Rob. >> It's not starting to hit. We got our diet coke. We're ready to get fired up. Just so you guys know, Rob's been on a um no caffeine, no alcohol grind for uh a month. So >> yeah, it's February 24th, a little over a month at this point.

>> But uh the energy is still here. >> We're getting things fired up. The energy trade, I think, is is still something to be something to be be watching. >> Um >> because you're you're saying you're silent. >> Am I saying I think it's something to be watching. I think this the to be aggressively allocating and doing. >> We're gonna we're gonna have a uh capital allocator on the show in just any moment, which I think uh would be a great guest to ask about the energy trade. >> Um >> look, do you think there's any golf courses around here in San Francisco?

>> Do you work in AI? Um and so yeah, we're gonna be >> we're we're gonna be uh we're gonna be watching this a AI energy uh repositioning capital allocation very closely. Yeah. >> Uh copper fiber optics. You saw Austin earlier yesterday. Talking about this photo uh this fiber optic trade on it. >> Uh RAM is also something that we're hearing a lot more about. >> You saw SanDisk run hard. >> Sandisk ran hard. Yeah. And so this is again just another one that's popping off in the energy trade part of building these AI complex and uh ultimately what we're going to continue to see as this capex moves further downstream

ultimately once you deploy this capital into data centers. You need to build all of the networking equipment, all of the servers, all of the RAM, all of the memory, all of these things that go into the components, the nuts and bolts of the data center, go into the data centers after you've, you know, signed the lease on the land and and all these things. >> And as more um as more of these agents start to come on, like more of the agentic actual agent to agent flows and these types, they're they're going to use so much more compute than just like typing to your LLM. So that's like the big part of the thesis right now is that's why it's accelerating because like you got cloud code and codeex previously was like all right this is going to happen but now it's like blatantly obvious that it's going to

happen. I think that's I think that's right because if you look at you know the the cycle in about you know 2020 2021 of what happened with gas on Ethereum you had this gas explosion everyone was just so mind-b blown about all the things that you could do on chain they weren't concerned about how much gas they were consuming >> uh and so you ended up having these gas spikes which is ultimately what led to the inefficiencies which we ended up solving but I I I think you're right I think a lot of this is getting repeated in the current AI meta where you're going to get so um just blown away and excited about some of the use cases that you're not going to be concerned or really watching closely the amount of tokens that you're using on a on a on a per call basis. Um and so I think you're

absolutely right. If we're looking at the adoption curve of of of these things, I think we're going to see this big boom. All of this innovation is going to happen. We're going to use these agents for anything and everything, but we're going to guzzle tokens. >> Absolutely guzzle tokens. And then I think we're going to have to take a closer look at this. we're going to realize that, you know, there's more efficient ways to deploy capex. We're going to get more efficient usage of these tokens and then we'll find an equilibrium and that's kind of where we are right now. >> All right, guys. We've got Hunter Horesley, the CEO of Bitwise. Hunter, if you're ready, feel free to come on up. Um, guys, we are uh stoked to have Hunter on. Good to see you, buddy.

>> Good to meet you, man. >> Nice to see you. >> Good to meet you. Welcome to the show. >> Oh, thank you. >> How's How's Neon treating you? had to get on the show for years. >> Yeah, >> let's go. Time. Some pleasure. >> This guy, huh? >> Is this how all the institutions are? You come in all bowled up, like chest out or >> I'm fired up, man. I'm very fired up. Yeah, it's very exciting time. >> Matt on last week. >> Fantastic. Yeah, that's right. That's right. >> Incredible episode. >> Yes. Yes. He's a lot of fun. He's the A team. I'm the B team. So, uh I'm sorry to be in such close succession with him, >> but Matt's Matt's fantastic. And yeah, it seemed like you guys had fun.

>> Absolutely. Yeah, I mean we we were talking about a lot of the things we're talking about here which is more on the DeFi, the tokenization side of things, the institutional side. Maybe we're at near con we'll talk more about the AI trade, the energy trade. All of this is happening kind of in confluence with one another. AI is Apple intelligence, right? >> Right. Right. Right. Right. >> Are you in San Francisco? Yeah. Yeah. We started Bitwise in San Francisco and then I completely lost control of the situation and so now we're 200 people and New York's the largest office and then London >> and San Francisco uh after that. So, but I'm still in San Francisco. Matt's in the Bay Area, Hong, our CTO's in the Bay Area. So, some of us out here. Yeah.

>> Very cool. Um well, let's uh let's go a little bit off of the tokenization conversation, a little bit more of the AI conversation and just curious as far as how Bitwise is reacting to some of the things that are happening out there. We saw the Catrini article talking about 38% downside in the S&P in 2028. Um, obviously all of this capex is getting deployed now. When you guys are looking at this kind of on the, you know, on the digital asset side and the tokenization side and looking at AI and agentic payments and all these things, what is your reaction? How do you know like what's hype and you kind of want to stay patient on versus like this is something we want to aggressively get into because this is really fundamental to the thesis.

>> Sure. Yeah. I I uh uh I heard a quote last week from uh an investor I really admire and uh he said uh when you see a freight train, figure out how to get on it and make sure you're not in front of it. And I think AI is is one of the freight trains of this moment. There there are a few others, but I think the right mindset for people to have right now is um this freight train's not stopping. Uh maybe we don't know everything about the sequence of events to come, but how can we participate in it? I think that's the mindset of investors. I think because it's changing so rapidly and and not perfectly clear, what you see from a lot of investors is um taking risk off and saying let's let's let this unfold a little bit. So they're selling Salesforce, they're

selling CrowdStrike, they're selling uh a lot of the names that you've seen um sell off over the last uh two months. So Bill Wise manages $15 billion. We serve banks, wealth management firms, large hedge funds, family offices, a broad cross-section of investors and being a partner to them and navigating the opportunities in crypto. And I think the theme across the board is this is a huge element of what they're trying to process right now. They're also trying to process Kevin Worsh's Fed governor and what that means for rates and monetary policy, geopolitics, and a host of things, but this is one of the biggest. So um so I think it's a moment of uh uncertainty for investors but not uncertainty in in um uh in a pessimistic way in a very optimistic way but um uh

you know if I said um uh if I said uh all of the all of the houses in San Francisco are about to be repainted. Uh which one do you want to buy today? You say well maybe I'm going to wait a minute and see what color they get painted. Right? So, uh I think I think that's sort of the mindset which is there's a lot of optimism. Um but in this precise moment, um people are unsure exactly um the implications for everything. >> Hunter, are you using AI a lot in your personal workflows? >> I I'm I'm like I'm like the unimpressive version. I still use it like Google search.

>> Okay, cool. Um at Bitwise of course the engineering team um you know like tech leads uh managing an army of um >> uh of across codeex and and um cloud code. Um uh the biggest the biggest aperture we found I think so far but Hong our CTO is thinking a lot about this um and a number of um people on the engineering team is um um if you are operating a company um there are certain things you might want to automate. So imagine uh imagine you have a team that spends three hours every week on something like uh recording certain records. Um they

they might say like I wish I could automate this. It would save me time. It'd be more accurate, more reliable. Historically, one of the challenges is great, everyone would like to automate it, but if it's going to take uh an engineer a month to build the internal tool that automates that, think about how many how many weeks of three hours saved it will take to break even on a month of engineering time. With uh with the capability of AI, if you can build internal tools in three days, the break even on building an internal tool gets way tighter. >> So, uh we've built a number of internal tools. uh now that have had um you know single or double digit number of hours a week that they're saving and I think

that has been really powerful and uniquely enabled um by AI. >> Um curious how your personal experience has been as as a CEO and as a founder just kind of implementing inside your company for yourself as well. I want to talk crypto. Um, uh, IBIT is Black Rockck's biggest most profitable product, the biggest rise of ETF or of any ETF in history. Now, it's the most profitable product. Uh, Bitwise has obviously gotten its toes wet across, you know, you've got this sick index ETF product. You've got all the majors, other products. Uh, what has been the last year and a half, two years of getting into into crypto and digital assets? How much has it made an impact on your bottom line? and like what's the

pri uh priority look like for 2026 2027 you know if is it going to be a big focus how much talk kind of about like your broad outlook here >> sure um uh so maybe maybe I'll start at a very high level um there one of the things that makes crypto captivating and also extremely confusing is that uh there are a lot of different communities that look at it different ways and talk about different things and you know sometimes like totally totally talk past each other. So, you know, there's a financial services community that sees a new payment rail, something that can do electronification of markets. Um, there's a sociopolitical community that views it as a return of sovereignty and censorship resistance. There's a

technology community that sees applications and users and um uh to provide one lens uh for Bitwise's view on the space. Uh we think that um let me tell a very brief story. Um uh an investor in Bitwise is uh Henry Kravis who started a firm called KKR. Um I was uh with him uh having lunch and so the opportunity to ask some fun questions and I asked him to tell me about the story of his first fundra. It was in 1971 and it was a $25 million fund and he said they called it a bootstrap acquisition fund. And I said why' they call it a bootstrap acquisition fund?

I've never heard of that. And he said because I hadn't come up with the phrase private equity yet. And I kind of laughed and then I realized, you know, the big bang or God created earth, but a mere mortal had to create private equity and it probably was uh Henry Kravis. And so went on a journey from something small that nobody really knew what to call. It matured into leverage buyout, which was cavalier and risky and not for everyone, but certainly an important asset class, and eventually matured into what we know as private equity today, which has a role in a huge cross-section of portfolios. Uh in the 80s, Milin took untouchable non-investment grade bonds, turned them into junk bonds, um which people thought were cavalier and risky and not for everyone, but certainly important. And then they've matured into

high yield bonds, which now have a role in every portfolio. So, and there's other examples. There's there's alternative non-bank lending becoming credit in the 90s. Um so, the premise of Bitwise is that uh uh there are many ways to look at crypto, but one is the emergence of a new alternative asset class, and that is not a new thing. That's happened multiple times in the last 50 years. I just described a few examples to you and that's what's happening here. And crypto, like every alternative asset class before it, will have to go from something that's a bit of a backwater that no one's heard of and small, then it'll become something cavalier and risky and not for everyone, and then hopefully it'll earn the opportunity to be mature and have a role in every portfolio.

>> Uh, so that's the premise of Bitwise. Uh and just as KKR built on the back of pioneering private equity or Apollo or Aries on credit, um the aspiration for Bitwise is what this asset class needs uh is very different than what you need um for being a real estate investment manager. >> Um so we'll purpose build for this. We'll try to play a constructive role for our clients, investors, and for the ecosystem itself. uh and hopefully um earn the opportunity to be one of uh people's favorite firms to uh invest with, partner with um and rely on for what they're trying to do in the space.

>> That was very long-winded, but that's sort of how how we view things. So, today we have four business lines. We have ETFs in the US. >> Y >> um BTB, our Bitcoin ETF, uh is one of the largest. BATW, our crypto index ETF, top 10 is the largest in the world. >> Yeah, that one's great. Um we have an equity ETF of crypto companies like Bullish and uh Coinbase and Galaxy and uh Bitco and uh that's BITQ. So US ETFs then we have ETPs in Europe that we have a near staking ETP in Europe. Uh that's the second business line. The third is Alpha. Um a fast growing solution there is for people who hold say $100 million or $25 million of Bitcoin. They want to generate yield. we can um run an option

strategy where they keep the bitcoin in their custodian >> and we can generate four to 10% yield on top without them having to lend it out or change the counterparty risk. And then the fourth part of the business is onchain solutions uh where we steward a few billion dollars of client assets through staking and recently vault curation. The way those all interrelate is my view uh from an investor perspective, the biggest obstacle to engaging with the space is that people are just busy. >> They're busy. They got a million things going on. Like why haven't you guys invested more in AI? Because you're busy. It's not because you're opposed to investing more in AI. You just got a lot of things going on and figuring out clean opportunities and how to how to get at them is hard. So, uh our role is

to simplify that. And we have a we have a bank client who started out with a Bitcoin ETF, then started uh investing in the Ethereum and Salana ETFs and now wants to use a vault on chain. >> Um so people go on journeys, individuals do, the three of us do, institutions do, and our job is to help support them. Um and uh so two years ago the ETFs I think were a huge unlock uh to your original question. um and are a powerful first step for investors in the space and today we we furnish a broad set of things that they might want to do and u grateful for the opportunity to get to play that role for them.

>> Hunter, I hate to put you on the spot, but it was all the institutional folks coming on to any big, you know, crypto podcast saying that the four-year cycle was over. >> Yeah. >> Here we are in >> Yeah. >> February. Yeah. >> Four year cycle feels on. >> Yeah. What happened? >> Right. It does, doesn't it? >> What happened? >> Yeah. Yeah. Yeah, it's so interesting. I mean, I think our view was that the four-year cycle was sort of a deprecated uh mental model that the market had changed so much that it wouldn't be predictive. >> Um, >> which perhaps is still correct, right? It's, you know, >> Yeah. >> time will tell. >> Time will tell. Yeah. I mean, I think I think what may have happened here is that um it just has taken on a self-fulfilling life of its own. Yeah. I

mean certainly in the fall of last year there were a lot of people asking us do you think the four-year cycles will play out >> right >> and our view was no because the circumstances are different we could get into that >> right >> um but um but maybe it's now just such an instinct for people >> that uh that people sort of reflexively made it a reality by worrying that it would be a reality so they made it a reality and and that type of thing can happen in markets. I will say that um um so uh you know like uh I had a colleague who was chatting with a a billion dollar wealth manager uh today and um uh I asked you know what was what was the

vibe what was his mood like and he's like he's a little anxious and I said is because the the price of crypto is down he's like no because the because uh because all the stocks are selling off >> um and I think something that sometimes needs to be reintroduced to the perspective of people who are very involved with crypto is that for most investors, this is like 2% of their money, >> right? Yeah. >> Yeah. I mean, you probably have like 2% of your money sitting in some wallet that you've forgotten about. >> Um, so what their mood is often related to is what's happening with 80% of their money. Correct.

>> And what's happening with 80% of their money is Salesforce is down 30% since year the beginning of the year. Microsoft is now 25% less valuable in two months as a company. >> Yeah, I saw that. >> Um, Palenteer is down 40%. Uh, people are wondering what Kevin Borch means for the Fed. People are wondering about the development of uh with tariffs in the Supreme Court. >> Um, of course, every week there's a new AI announcement that seems to >> to rerate uh whole industries like like security last week. So, so you know, so, so, so the way this relates to the four-year cycle is, um, you know, if everything is down 30% this year and Bitcoin's down 30% this year, is that a Bitcoin cycle?

>> Yeah, exactly. >> Um, or is that a liquid market >> cycle? Um, so I I think it's an instructive and important uh mental model. It's definitely here to stay in people's minds. I don't know if it has as much causal uh power as it did in the past, but whatever the case, the market's down so far. uh it may not. We're also early in the year. I mean, if you guys think back to last year, >> Bitcoin sold down, I think, from 103,000 to 72,000 at the end of April. And then in Q2 was up dramatically. Um 2026 is young. Uh and uh and we may well end the year higher and print a down year in year three and an up year in year four. And then once again, we'd

be back to saying the four-year cycle is this. I just I think it's an interesting thing for us to keep track of, but I don't know if it's uh if it's the map to to guide the travels. >> Yeah. Yeah. I think uh I think investors uh were I I don't know. I just thought it was funny like the so many institutional folks were saying that it was over and like you had like the typical anonymous crypto Twitter account guy on on Reddit with the date time perfectly. >> Yes. Yes. Yeah. Yeah. Oh my gosh. What the heck? How is that possible? >> I know. It's weird, right? >> Yeah. Maybe maybe I'm an idiot. Uh >> come on now. I mean that's not the point of the question.

>> Timed it perfectly. >> No, I was just uh I was just curious kind of like because I'm sure people are calling you guys or like asking you guys like, "Hey, like what's going on?" >> But I suppose you say it's more like macro driven than it is kind of specifically market driven. >> Yeah. Microsoft is worth 30% less in two months. What happened? >> Like why is Microsoft such a worthless company now? Yeah. >> You know, like so then Bitcoin which is this much less consensus thing is is down. So it's all a matter of perspective. I will say that I think one of the elements of 2026 uh and crypto that is so interesting is that everything has changed.

>> Yeah. >> I think that we've been in the space for eight years and um let let me mention another story. We have we have an an investor in Bitwise named Kether Boy who was part of the PayPal mafia. >> Yeah. >> And I asked him oh boy. Um well I I asked him uh uh he invested in Stripe but he didn't invest one of the first few rounds and he's one of the most accomplished early stage investors in the world. I mean, he he could invest in the first round of almost anything. And and if I recall correctly, uh he said something like um I knew enough I knew too much and got in my own way. Meaning his experience from PayPal gave him uh understanding that caused him to see sort of different patterns and

possibilities for Stripe than maybe people who are approaching it with a fresher mind. Um, and then of course he ultimately invested very early in Stripe, but not first. So, what am I trying to get at? I think that for people who've been in crypto for a long time, I think there's a lot of intuitions and mental models that need to get dumped now. >> Yeah. >> Um, I think things have almost completely changed in the last six months and are going to be very different in the next two years. and included in them to your point about you know the person um on uh on Reddit or wherever predicting the price I think that there's like almost a completely bifurcated set of communities now there's sort of cryptonatives and uh sort of institutions and corporates

>> and depending on your information flow depending on who you talk to the most you could get very different reads on things >> the important so I would say this audience it's like a miracle >> yeah It's all happening. >> Yeah. >> Um% the the sensing gap is insane right now. >> Is absolutely insane. >> Never seen anything like it. >> And the thing that people have to remember though is is um this audience controls all the assets. >> They own all the assets. So if they're bummed out and decide to sell, they have way more to sell than this audience. You know, you mentioned you mentioned IBID, which is which is a remarkable historically fast growing Bitcoin. It's $60 billion. There's there's a trillion dollars of Bitcoin.

>> Yeah. >> There's another $940 billion owned by this audience. And if this audience is bearish and they dump, you know, 10%, that's I bit and a half worth of selling. It's as if I bit went to zero and more. >> And that audience, the the natives moves quicker. >> They move quicker, >> sentiment driven. They they're not as kind of like slow. >> Exactly. Exactly. So, so I think I think it's both that these are two different cohorts now who are focused on different things. Um, >> this audience does not share very publicly still what they're doing. Yeah. Um but this audience uh controls most of the supply. So if they stop selling there's very little supply available to this audience to buy. But if they are selling then the buying from this audience at least right now is not

substantial enough to offset. >> Yeah. And you mentioned that right now this audience is what 2%. Uh last question for you is is that the right is that the appropriate you know metric and balance for them for their portfolio set? Should it be more should it be less and why? >> You know we we have one client who's at 20%. I I think that the uh the academic analysis, if you look at our white papers, spit out something like two and a half percent. I think that where we're going to land is three three to 10%. Um I'll tell you that uh for our wealth management, so we serve the largest banks in America. Um we serve multiple wealth management firms that manage over a trillion dollars. Um a 3% allocation is like almost not worth

their time. Mhm. >> So 5 to 10% is like a small allocation. Yeah. >> So I just think that the practical nature of everyone being busy uh is that people are going to um trend towards larger allocations than most people expect. >> And then of course a third of people will never allocate to the space. And >> that's fine. I mean people >> another thing that crypto people have to keep in mind is that um investors can't even agree right now if you want to own Apple stock. And it's like the biggest company in the world >> Bitcoin every day, >> you know. So So like they're not all going to agree on crypto. And I don't think the end of the end of the story is that every single >> investor is going to own it, but a lot

more are going to own it. I think they're going to own it at larger size. >> Awesome. >> Hunter, thanks for joining us today. We had a good uh overall day here. >> Oh, incredible. Incredible. Yeah, Near's crushing it, sir. Absolutely. >> Cheers, man. Thanks, Hunter. >> All righty, guys. We've got a governmentbased client coming up next. >> Beautiful. >> We've got the the Bermuda government and co. We're going to take a quick break and we'll be right back, guys, with the next guest. live here at Nearcon with the Bermuda

premiere, David Bert. Um, David, fun fact, my dad grew up in Hamilton, Bermuda for 5 years. >> Oh, that's a fun fact for sure. >> For sure. actually and my best friend from home just left Hamilton for a 10-day holiday. So >> interesting. >> Good to have you here on the mainland. >> I'm assuming you have been to Ber. >> I've never been. >> What? >> I know. >> Shameful. You can come you can you can come visit us um at our digital finance forum and then you can you know do your podcast there. >> Okay. There you go. >> Sounds like a plan. >> So governments AI it's happening right. I mean, we actually have just recently seen Anthropic uh go after a Chinese uh uh um company called Miniaax because they are copying their models and

distilling all of their uh code into their models. There's this geopolitical forces happening, but then there's just the government and its people happening, the labor force, uh jobs. What is going on in Bermuda? What are you guys doing with AI? And just generally just give us a bit of a lay of the land here. So a lay of the land with the governor and AI. Well, first of all, let me talk about near. You guys were here at Nearcon. Please um and uh certainly so from the perspective um my team was at Nearcon last year. Um this is my first time attending. Uh but one of the things which the uh team at Near told us is that there is without question opportunity uh to um for Bermuda to exist in a unique space of AI regulation. Um, and I'll get into a

little bit more about that, but uh, just by background, Bermuda is a small jurisdiction, 65,000 people, 21 square miles. Um, but we are one of the most advanced financial services regulators on the planet. For instance, there's only two jurisdictions that have regulatory equivalents with both United States and the European Union on when it comes to insurance regulations as Bermuda and Switzerland. So, we are well established. We've been doing digital asset regulations since 2018. So we're established global leaders home to uh many uh companies inside of the space and so from that perspective that is where we set our out our stall >> with near uh the view was that there's a space for the regulation of AI agents especially in AI agents when we're dealing about financial services transactions because we know there's

going to be more AI agents that are going to be engaged in financial services transactions. The question is how do you regulate these? How do you make sure there's a dispute resolution uh frameworks? How do you ensure whatever it is that they're doing? all those type of items and so that is a space where from a government Bermuda and a regulatory perspective we are moving forward right now. So um our government um is working in concert with our regulators and our local legal community to come with legislation in order to regulate artificial intelligence agents inside of financial services transactions. So that's one thing of which we're doing on uh the AI side. On a government implementation side with near um we announced today just recently um here at Nearcon uh that the government will be working um with Near to implement an AI government

chatbot specifically of course using the privacy tools that Near does have in order to make sure that persons can interact in various government services. The place where we're starting is for pension calculations. Something that infects internally, but it's a very good space and it's an expression and demonstration of the power of AI to distill things very simply where things are very complex because our team of actuaries gave us this very complex spreadsheet that they wanted individuals to try to fill out to get something and we figure out there's another way of doing this. So that is what we're doing directly on the larger scale not near related is that we um have on boarded um a strategic partner um working on um artificial intelligence deployments throughout the government of Bermuda and we're continuing to press through that

aspect. >> Yeah. And the space moves so fast sometimes it's difficult to exactly where to implement yourself but you got great partner here in near you also alluded to the privacy you know features here. Um over on on this side we've got anthropic that's restricting the US military from accessing its AI. Do you think, you know, if they had like nearest technology, if it wasn't anthropic, not this closed source censorship model, but an open- source private model, do you think that overcomes some of the potential issues that governments are worried about?

>> That is a question that is way above my pay grade, which I will not be able to answering a government. No, >> I I I I accept that, but I'm not too familiar with the I' I've just seen some of this stuff from Gardener. Forgive me. I've last week I was delivering my country's budget. So, so, so I have not been able to, you know, keep up. I had a I had a press conference yesterday morning, the Chamber of Commerce breakfast, hopped on a plane to fly all the way to the West Coast. So, I'm not up on all the latest items, but but but what I can say is that yes, it is evolving quickly and that is the reason why it is important and essential for governments to pay attention. There is a need to make sure that there is responsible artificial intelligence

regulation. It is necessary and proper and I do not believe that enough people are talking about this. I was in the UAE um uh uh early this month having the exact same conversation discussion around this. There is not enough attention that has been given towards artificial intelligence regulation and certainly Bermuda can't set the global stage on that. But at least in so far as when it comes to financial services, we're trying to make sure we play a part. >> And and maybe this is a more relevant question here because one of the hot topics here is whether this should be federally regulated or should be regulated by the states.

Granted, you guys are in a position where you can regulate for, you know, the the entire, you know, Bermuda, Bermuda, but like where what are some of the things that we should be thinking about? How do we make a decision like that? >> So, the reality is that that is that that's a domestic political argument which completely obiscates the fact that this is something that cannot be done at a state-by-state level. This is something that is going to require international cooperation and actually understanding before it becomes an existential threat. The challenge of which we know is that nation states are looking to compete to see who can have the more powerful AI to have certain things and all the rest. And I think um the example which I gave is that it took um you know the threat of global annihilation for the uh Soviet Union

then and the United States to come up with you know nuclear treaties you know to stop and to have this uh you know this balance of mutual assured destruction. But clearly there are some case and applications in the future of AI where the result can be exactly the same. And if the threat of global annihilation is not enough for governments to wake up to say we need to put some guardrails around that then I don't know what is. >> So maybe an approach here is is what lessons can we learn from financial services regulation when you're operating on an international scale? Is there anything that we can glean from that and apply it to >> AI? The reason why financial services regulation exists is because people want to make sure they don't lose their life savings and people don't want to make sure they don't become personally ruined

and bankrupt. But personal ruin and bankruptcy is the same as national and international, you know, Armageddon. Yeah. >> So the same things um apply. It's just a question as to when will there be enough >> I guess I would say persons who are more concerned with the uh future of humanity rather than the short-term uh constructs and thinkings. I remain hopeful. Um I know that there are more people talking about this issue. It is something that I intend on continuing to speak about even after I'm in office because I think that it is very very important and not enough people are talking about the need to responsibly manage and regulate artificial intelligence.

>> Dave, can I ask you more broad question about what you've accomplished in your time in office? >> Oh, really? Yeah. >> You want to talk about domestic political stuff here? So, I was elected in 2017. Um, one of the things that I would say is that we're here at Neircon as well because uh, we put in place one of the first digital asset regimes um, fully digital asset regimes on uh, the planet. Um, and that was in 2018. Eight years later, I can say without fear of contradiction that we have the best digital asset regulator on the planet. Um we are hosts of some of the biggest names in uh the the digital uh finance sector and we're continuing to move forward in spaces in regards to outside of the regulation of AI and AI agents moving into identity also uh moving in uh to the space of um you know tokenized

insurance we have all those type of things um democratization the real type of items token tokenized funds so we're talking about people having wallets last year at a digital finance forum we had the first world you know fully compliant airdrop where we're actually doing things in a regulated environment giving those things and showing so persons in Bermuda can demonstrate how uh to use digital assets but on a broad basis I mean you know we have a small economy and so my job is to make sure that we can transition uh to a digital age and I think we've successfully done that you know I can talk about domestic policy if you want implementing a minimum wage reducing taxes on loan workers and all the rest of the type of stuff balancing a budget delivering a budget surplus all those types of things I mean but that's

you know I'm not sure your podcast viewers are too interested in that. >> I think um my friends from Bermuda say that the core issue is finding work for locals. Locals finding work because of a lot of the p you know expats coming in to >> Oh no, that is not correct. We have the lowest unemployment rates in our history unemployment rate. >> That's David B. >> That is that is not correct. Our unemployment rates in Bermuda is um at 1.4% and it's 1.3% for Yes. So that is the full definition of full employment. Yes. Um and so from that aspect, so no, that is not the case. Youth unemployment is at a record low as well. So um jobs are certainly not a a challenge, but that's more of a function of the demographic shift where you have people

that are leaving the workforce, a whole lot of people retiring and not enough younger persons to replace uh them. And so that creates its own unique challenges. We have challenges like you find here in the United States as well comes to housing affordability and certain things like that. So uh those are the things which we need to you know work through. But um I'm hopeful for the future. Um, I leave office in 8 months and I think I'm doing all right in my time. >> Well, I'm done trying to, you know, poke at you. I'm just having some fun with you, but is my my last question just eight months. You know, what what are you trying to accomplish here at before you, you know, leave office? You want to go out, ride into the sunset?

>> I mean, we're kind of doing it now. >> Yeah. Now, I mean, the reality is that we um I mean, last month, um, at the World Economic Forum, we announced a partnership to make preview the world's first fully onchain economy. a partnership with Coinbase and Circle putting that ahead. We've just delivered a budget to put um our make sure we have skin in the game on that. Um and it's around making sure the promise of digital assets, the actual democratization of finance. The reality is that in Bermuda, there are things that you're able to do that you aren't able to in other places. We do have young kids who have crypto wallets that are able to be investing in things that were only available to institutional investors because those things have been tokenized and allowed. So, we're talk about insurance contracts and other things. when people talk about the

matters related to parametric insurance and all those other types of things powered by blockchain services those already things which we're doing we have insurance companies where's balance sheets are fully fully in digital assets already so from that perspective there's a number of things in which we're doing and so from our perspective it's just I'm here on the road to make sure that we share the knowledge with builders uh around uh these uh particular topics so that they can recognize that there are options in the world and Bermuda is an option >> the digital finance forum May 11th to 14th 2026. >> David Bert, thank you so much. The premiere of your Bermuda. Really appreciate you coming on the show.

>> Oh, thank you very much for having me. Take care. >> Absolutely. We'll see you next time. >> Got to wear the pink and blue. >> Let's do it. >> Man, wouldn't have been a a political uh segment without grilling the politician as hard as I could on his policy outside of crypto. >> You're a real investigative journalist, aren't you? >> Really really trying to get to the bottom of what's going on there in Bermuda. From what I've heard, there's a good rave culture. I've also heard that there's a lot of fun parties. I actually saw a sky themed party on Instagram from my friend last week. Uh they were wearing stars and sky. I heard that's a big culture thing they were meeting up.

Didn't get time to ask him about his party habits. I don't think you guys want to know. Uh but uh when it comes to bringing the economy on chain, I mean, they're doing it and and they're doing it in a big way. Obviously, we know that these kind of Caribbean islands have a lot of wealthy people coming to them, bringing them capital, bringing them uh you know uh trying to open bank accounts, trying to open businesses o o over there. Their form has Coinbase speakers, uh Scaramucci, uh Carlos from Securitize and others. So, yeah, Ilia and uh Alex are ready to come up, guys. We'll take a quick break, get back on schedule. Uh Chris is pissed at us, so we gota we got to keep it tight here. All right, guys. We'll be

back. >> Quick stand. Hey. Hey. Hey.

Day two, we've got Ilia, the only reputable face in crypto NSF, saving the entire industry from the AI propaganda and his right-hand man, the absolute man Chevchenko. We're gonna start calling you Chef Chchenko only, not Alex. Shukchenko. I can't My accent sucks. Gentlemen, Confidential Intense just came live today. We just had it announced at Nearcon. Why is this important for the market? Let's start here. What is this unlock? Uh why did it take you so long?

>> So, you remember all all of these questions around positions on Hyperlid and then there somebody was taking a position on Hyperlid and then somebody was ated the market to go down on the price so that the person get >> talking about Justin Sun on the plasma uh premarket >> there was this >> I think like it's every other week I think this >> so confidential intense is exactly the solution to this problem uh when your positions are available on chain then other people can explore them and can try to target these positions >> uh and confidential intense delivers you all the financial primitives but in a private manner. That's the whole reason

to build this. None of these centralized markets from the traditional finance are operating in the open market setup. I cannot see your uh bank balance. I cannot see India's bank balance or I cannot see you know other traders what they what their strategies. So why on earth on crypto is everything everything is different right? So we need to get at least as good as tryify and the confidential intense is what delivers this experience. >> Yeah. I think fundamentally when blockchain started the the only way we knew how to do kind of verifiable computation is by literally broadcasting it to the whole internet and letting everyone to reverify it. Right? That's what Bitcoin is is like just tell

everyone about every transaction. Let's make sure everybody verifies it. And first of all tech evolved since then. Right? So we have now ZK proofs. We have secure enclaves. We have all the pieces now to actually offer you verifiable compute that is private that offers you all the properties of blockchain but indeed protects privacy of your balances of your trades of your transfers and yeah I mean Trefy payroll like I mean I I usually say that crypto hasn't achieved its potential. We still don't use it to pay each other. We still I I don't use it for paying at the coffee shop because like I don't want them to see my account when I pay. I don't want you know when we do payroll everybody

see everybody's salaries and the everybody else can see oh this guy's paid this much let me offer him more right like this is just like it doesn't work for business period >> right it doesn't work for personal so to you know for us like confident confidentiality is effectively the mass adoption >> and it's here it's near it's near.com try it out and uh yeah you can trade you can transfer you can store you can and you and you can do it for any any asset right you have bitcoin You have stable coins, you have, you know, Ethereum, Salana, whatever you want. Uh, for example, near uh it's all there.

>> Yeah. So, we've been talking a lot about this kind of uh convergence point of legacy finance and DeFi. We're seeing a ton of tokenized assets come on chain. Obviously, stable coins have, you know, grown exponentially. And the one thing that um these firms are used to when doing looping strategies, when using DeFi, when you know, you know, moving uh their products crosschain is privacy. they they're obviously just so used to that. Feels like that's kind of like the pragmatic step as to like where we're at. How much of this confidential intent design was inspired by the trajectory of where the space is headed which is obviously much more institutionalized but then also just much more pragmatic from a product perspective like a wow we haven't had this why you know why is

this this is not as good as the current uh like where did the kind of inspiration for this product come from? Did it come from the success of Zcash and kind of how that played out? Where the inspiration come from and how do you see it being used in this new kind of tokenized era of products? >> So we have been talking with Ilia about different privacy focused products for years right now. >> I mean we we were trying to build private shard something of private shard in 2022. >> Yeah. So so it's not it's not like Z obviously Zcash created this uh privacy 2.0 whole uh >> movement um and uh but I think it's just people starting to understand that actually it makes sense. I believe from

my point of view that this is just a feature that needs to be presented in in any um like mature enough financial system, right? Like nobody is going to talk to you if if you're like nobody big is going to talk to you if you don't have this feature. And we also see it in our relationships and our conversations with uh uh big institutional investors, liquid funds, uh trading desks. This is something that they are used to because they're using centralized providers. Uh they are not using DeFi because of because of this reason. Uh and also because of additional reasons, you know, the depths of liquidity and stuff like that. But they are willing to get

exposure to assets that are not available on Binance. they are willing to have this exposure to loopin strategies and something more interesting. They they are willing to get this risk. Um and uh up until now this was impossible. So I think it is just a major milestone. It needs to be achieved in order for us to totally like overall grow the pie of DeFi and crypto general. >> Yeah. I I think the other transition I mean let's just address it head on you know until until beginning of 2025 crypto is kind of in a doghouse >> right and we are now in a world where at

least like there's a way to have a conversation figure out how to do things compliant like we don't want to facilitate crime we don't want to facilitate moneyaundering we want to deliver products that are truly you know innovative enhancing commerce, you know, unifying it and at the same time, you know, protect the users because like actually compliant products are good products. Like it's just something that other people want to use as well. And so before there was no way even to have this like okay, how do we even approach this, you know, because it's like all crypto is kind of like no, it's totally illegal, right? And so now, you know, we're in the regime where we can figure this out, including globally. And I think that's also a big step And now

institutions are also like okay cool now like compliance is coming >> we can actually start using this okay how do we use it without yeah leaking all of our proprietary strategies and information to the whole market and uh yeah so this is yeah as Alex says like it's natural evolution that this is time now to bring uh confidentiality on chain enable this for everyone. >> Yeah. Well Alex you know you were on the show during your intense day in Denver. We had you on a little bit earlier today as well. Uh IO, you were on yesterday talking about the AI marketplace and it it feels like these things are coming together now. Uh and so the intense uh near intense is effectively the same sort of technology that underpins the AI

marketplace. Now we have confidentiality around this as well. How does that help upgrade the AI marketplace as we're asking AI for labor and vice versa? AI is asking us for labor. Like now that it's private, what does this enable? Yeah, I mean exactly the effect actually I mean Alex actually at Redacted in 2024 he was describing the agent marketplace as the stock right uh and then the first version of that is effectively just crypto trading. Yeah, but uh but now that we have agents actually working, which wasn't true back then, now it's it's uh the marketplace as well as possible, but it is underpinned by the same tech. It's underpinned by the same principles. It's effectively escro resolution, dispute resolution, all the systems are powering it. Uh but yeah,

privacy is extremely important because again, let's say I'm a business and I want my agent now to go and deal with all the suppliers. If all of the transactions are now in chain, you can see this business is buying from the suppliers for this price. Now, you know, you're going to be non- differentiated in the business case, right? So, this is really again if I'm asking for a travel tickets from my agent and goes and like uses a travel agent, pun intended, uh to, you know, get the tickets like now everybody sees kind of, you know, the payment like that's just like not realistic. And and also importantly, you don't want the counterparty actually know how much money you have. Like this is this is the worst place to be when you're creating a market is when the

other side knows how wealthy you are because now the market mechanism itself is kind of broken, right? They will price you up for sure. And so privacy is effectively kind of again a fundamental enabler for this to be truly mass adopted, not be like a you know experimentation. Yeah. >> So, yeah, it's exciting because like again, crypto has not delivered on on this mass adoption and we can now with the confidentiality. >> Yeah. Another thing that you guys rolled out in tandem is turning on the fee switch for near intents. Obviously, um the market is asking for these things now. I don't know if you guys felt a lot of pressure externally about this about uh near tokconomics also had an upgrade

that was launched yesterday. talk to us about kind of the economic modeling of intents. Obviously, you guys uh previously prior to the fee switch had a very very uh you know tight spreads for size across tons of different assets, tons of different uh chains. Uh kind of won the one a bit of of the share of the market there in terms of like people want to use it. How do you anticipate the fee switch changing user behavior? Will it? Um, and how do you kind of look to economic growth forecasting for all the people out there who are demanding revenue now and you know token buyback because we can't invest in growth anymore because now we're a real industry apparently which I expect that pendulum to swing back as well but just talk to us about how you think about

that. >> Yeah. So uh so my thinking about it is that um any sustainable product needs to have uh good economics right and uh uh L1s are the these type of products and uh uh what we see is that it's not enough uh to have just gas fees being a revenue driver for these these products. the valuations cannot be justified uh with just gas fees because these fees are connected with just the execution rather than a value that is brought to the user. Now which means that uh there

are some core products on top of these L1s that are actually bringing value. It's not the protocol itself. it's actually the the use cases that are there and the use cases that we are focusing for near are intents uh so the financial use cases alongside with the AI use cases right and uh how we're thinking about it is that if these use cases are generating revenues then these gen re revenues needs to flow into the near token now the fee switch u actually introduced only uh only two major things first one is that uh all of the contarties now uh receiving the payouts of the fees that they have already been collecting in near tokens. So instead of

getting stable coins or some other tokens they are getting near tokens. >> So that's the first thing and the second thing is that the fees that were already in place by the distribution channels now these fees are going to be shared uh with the near protocol. So near protocol is going to get a cut uh here and uh which totally means that there will be no change in the fees >> uh for anyone uh and uh I do not expect that there will be any major you know change in in in the user behavior or like how the distribution channels are going to uh are going to work. We are working with them uh in order to make sure that this transition is smooth for them and it's not not very hard. So for

all of the new uh comers who are going to use near intense uh this solution they are not going to have this strange period in the beginning then it was something and then they are switching to something different. So for them it is a little bit easier but for existing partners we obviously handholding them through this period and we have a grace period for them to adapt to all of the changes and uh yeah we've built lots of great relationship uh within the last year. Um we value a lot our partners from the biggest distribution channels uh and uh we're committed to moving uh handinhand with them together into into this better setup uh that is sustainable also for us.

>> Yeah. Um, one of the things that we're seeing quite a bit of as we're, you know, acrewing more value to the near token. I'm curious how, you know, some stakeholders and maybe like a near labs entity or like generally the organization, you know, a doing this, unis swap's doing this. They're bringing, you know, the labs corpse and the tokens closer together. This great unification of equity and token. Um, could you just tell us a little bit about how the organization looks right now? uh and as you acrew more value to the near token do you anticipate following a similar unification journey? >> So we actually done that over past two three years already because we kind of knew this is going to be the thing and we didn't want to have you know some of the problems that people have have faced

uh or or uh and so we actually the original labs company doesn't exist anymore. It was shut down effectively through 2024 and instead we have I mean effectively subsidiary from like a structural perspective subsidiaries of foundation. So foundation is fully aligned with the token it is literally fiduciary duty and mandate to grow the ecosystem and uh the companies are effectively kind of feed into this. It's it's a little bit more complicated because of all the legal frameworks, but the the structure like incentive wise, this is the structure and we like all aligned with the near token to be kind of the the measure of success.

>> Yeah. What do you want to see from the intense growth this year like >> um well trillion a day in volume? >> Easy. >> Today would be hard, tomorrow would be okay. >> Yeah. So, so I think like I mean it's really exciting to see um effectively network effects starting to take place because now if you are a wallet dex or a neo bankank or an application it makes sense for you to integrate near intense because you're getting access to you know 30 plus chains and the numbers keeps growing you get access to all the assets and you don't need to do all the integrations all the manual work etc.

And we kind of start seeing like as more people integrate see like oh damn I'm like I'm getting all this functionality and it's like one simple API right there's also a bunch more really exciting kind of extensions like adjacent markets that are coming there on the asset side right on chain side people like hey I can just integrate into near intense and get all the distribution channels now access right so we're actually seeing that kind of um network effects are starting to play in and so the idea is to keep growing keep strengthening that network and kind of really building uh like core there. And then a gentic marketplace is coming to really again obviously it's still early right this is like you know there's like

whatever 300 agents there 400 agents with a few thousand jobs but this is how we're going to bring like real world commerce this is how we're actually going to bring this trillions of actual you know shipping containers moving around is because we can say like hey I need you know steel to be delivered I need you know like to figure out how to build this building etc and all of those contracts will also go in chain and So that kind of two pieces that are really like starting to uh take place. Meanwhile, obviously the near.com really exciting. If you haven't tried, it's amazing experience. And so really starting to grow that and and getting people to use to that blockchain is not hard. I think like that is like if you

go talk right now actually here in San Francisco outside, everybody's like either it's a scam or it's too hard, right? That that's a prevalent and neon.com really delivers experience where it's not hard. You don't need to think about private keys. You don't need to think about, you know, gas fees, blockchains, switching wallets, RPCs, like all of the things that we as crypto people kind of got used to. And like begrudgingly just like, okay, I have five different wallets. I need to make sure remember to not use the one for this. Like I paid somebody from this one. Let's make sure to not pay ever anyone again. Like all of this just disappears, right? It melts away with near.com. And so that's really exciting and like would love to have you know people try it out and really kind of experience that uh kind of for for the

real like the crypto doesn't need to be hard. >> Yeah. >> Awesome >> guys. Thanks so much. Yeah. >> Shipping. >> Yeah. >> A lot of product announcements, a lot of token announcements. Intense. Um it's been a busy week, huh? >> Yeah. I mean Alex been uh shipping on stuff. >> Yeah. Like what did you guys do the last couple months? Just absolutely ship. >> Yeah. I would like to give a shout out actually to to my team. Um uh they are amazing professionals. I'm super honored to work with them. Uh there are not a lot of them here at this event. And it's me talking head coming here, you know, and team is shipping everything and and selling everything that yeah that they

have been doing. But it's definitely they are not sleeping, shipping on the weekends, all of the stuff. they are very very excited about everything that we're doing. They are the first adopters of all of the technologies. Everybody from the team are using our products. So that's just a great uh it's just an amazing community to be in and be along alongside these guys. So um if somebody from you are listening to this then it probably means that you're not shipping and you need to get back to it. >> Oh the cloud is shipping is shipping.

>> Yeah. Or something is compiling. Yeah. >> Yeah. >> Okay. That's okay. Um, yeah, but big big thanks to my team for for this amazing ability to to showcase all of these great technologies. >> Thank you for having us out. >> Thanks for the rollup team. If you're if the rollup team is watching this, you guys are going non-stop, seven hours a day. >> Thank you guys as well. >> Pleasure, guys. >> Guys, we'll take a quick break. We'll be right back. >> Thank you. Hey. Hey. Hey.

Hey. Hey. Hey. Heat.

Hey. Hey. Hey. talking today with Lucas, the CEO at

Tacio Encrypted Compute AI. Lucas, how's your Neircon going? >> It's great. like um really have to say congrats to the one who said like go back to SF with Near Con. Uh like the timing couldn't be better and yeah like for a long time like I haven't seen a conference where people actually listen that much to talks like usually it's like the keynote and then everybody's gone here like the stage is full all day. >> Yeah. >> Yeah. And both rooms too. I just walked outside there and there there's a whole side room. Someone's giving a lecture and everyone's there like either taking notes or doing the demo. Uh I mean why do why do you think that is? Why do you think your and we were at ETH Denver we saw it too right last week this is different why why do you think that is

>> so what I've observed is that like it's really the first time credible AI people and blockchain people come together and listen again to each other and not about like some hype or like some blockchain companies that throw AI to fund raise on it. Uh so I was on a panel with someone from KPMG that like really goes to the enterprises and tries to educate them about EI and like to see the gap what she told me about like Twitter AI and enterprise AI. It's just like you don't observe this on Twitter anymore and you get new information. I think she was the most interesting one on the panel because most of the audience here didn't know about like what enterprise AI looks like.

>> Yeah. Incredible. >> What does that look like? What does enterprise AI look like? So what she told is like that it takes roughly eight to nine months to approve a machine learning model an I model in a financial sector and after eight or nine months we know like you don't use that model anymore nobody on Twitter use that model anymore like do you want to deploy it even so why even start the process but now apparently they're trying to cut it down to weeks which makes it more feasible to deploy >> we had someone on earlier talking about AI supervisors uh I'd imagine that if you can get almost like a fact checker supervisor or an implementation agent that you can get verified and approved at an enterprisegrade level. You can then take that and you can say, "Hey agent, you've already been approved. Go

verify this new model. Go see if this new model is ready for approval as well." Uh, and then it could hopefully slow it could hopefully speed up and and uh, you know, shorten the amount of time that it would take to implement some of these things. Um, you know, we're we're seeing that like, you know, the military is using Claude, but like Claude's not letting them. And so, you know, there's some entities that are trying to move fast. Others are kind of trying to do things by the book. How much do you think, you know, this is necessary and important that we fact check and verify every model versus just kind of putting these arbitrary slow down speed bumps in the way to try to you know just kind of get in our own way or slow down something from coming in and shaking

everything up. I think you have to see like in which context you're deploying it like as a startup scaleup like speed is the only thing you have like you cannot like do a lot of like this check and guard rails uh at least like in experimental mode like once you go to production things look different like you if you have real customers like institution grade customers also as like a scaleup you think very differently uh so I think those two worlds have to exist and what I can imagine like that now nontechnical The persons get so much exposure to cloud and whatever they get really FOMO that they have to deploy at scale in enterprises like you saw probably they do machine learning for 10

15 years in finance and nobody the co was never interested this is like just quantitive data science like whatever and now it's like he can try it out on the weekend and and Monday it's like we have to do this. What do you think is the is the biggest uh vertical or uh sector that is going to be infiltrated and usurped with confidential encrypted AI? Like where do you see the biggest use cases for your company? >> It's like really solidified in I would say like the last six months like there that's where like the privacy hype started. Before that it like was really hard to pitch like about the privacy project and technology. And we're seeing this now in three use cases. One is

identity payments and both combined you see an agents like an agent needs an identity and it needs payments and it should be private and verifiable to make really the trust scale on an internet level. >> Yeah. Well, the near team has made a bunch of updates, made a bunch of announcements. I'm just curious like being here at NearCon, I saw you yesterday and now you're here again today. What do you think has been the most influential or important announcement that Near has made this week? I think like confidential intense is the most interesting one for me and like how bit how they think about confidential compute like that's where we're also like very much aligned with Near and Ilia and we talk a lot about this uh so I think it's really great to have someone like near pushing the

boundaries here uh from like not in my bubble I think ironclaw is definitely a thing that like goes way beyond crypto like that's yeah >> everybody wants claw and now you just have like a secure and private thing of that like that's essentially what's the only downside of uh claw so far and if they can solve it I think it will break out of crypto very easily. H >> how do you see yourself or your company implementing some of these things like near intense and ironclaw? >> Yeah. So near currently is mostly working with trusted execution environments like we want to lift the trust assumptions up with like cryptography. So we're working mainly with two technologies multi-party computation and your knowledge proofs

and you have this triangle so far like you have you have like on the one side you need performance especially in AI you have programmability so general purpose and the third is like is it a cryptography guarantee or not and you can have two out of three currently so if you're going with tees like you don't have cryptography security so far you still kind of depend on the hardware manufacturer an operator a bit that he doesn't mess around with the devices. On the other side, you have companies that go like uh FHE EVM or like this general purpose like encrypted compute environments where you lack performance for like use cases like agents and so on. And where we come in is mostly like we do like performance and cryptography

like dedicated solution that really scale and are based on cryptography. >> Awesome. Well, I think we're almost out of time, but I want to get, you know, the last word from you. Just generally any closing thoughts here that we should be keeping in mind, right? You guys are very deep in the space. We've had a lot of conversations here today. What do you think is some of the most important things that we should be thinking about and our audience should be thinking about to prepare for the AI super cycle? >> I think like what everybody should think about is like how like AI doesn't need a different trust model. We cannot operate on don't be evil anymore. It has to be can't be evil and this has to do with blockchains, cryptography, privacy like if we don't get this right like I don't like this is the really fear that I see

with AI >> that if we cannot have guardrails that are not depending on a vendor on an operator this is becoming really dangerous. >> What is the next year like for you guys as far as product and what you guys are launching? like what practical things are you guys going to launch that are going to be usable by consumers or usable by businesses like what what's the what's the product and what's kind of like the business model. >> Yeah. So the product is an encrypted compute layer. So we develop services like these are these high performance primitives where people in the identity space or people in the payment space can use it to have like a private version of like a blockchain. So and we're like really deep down in developing it and

will launch in March. So in one month we'll launch publicly with the first customers with the key customers in the identity space. So they can use then our network in production and at a scale >> and then how do you make make any revenue on on that? What's your business model? >> So it will be a business model that's based on usage. So the more compute you use on the network, the more you will pay. >> Got it. And the network is like a blockchain or is it like a distributed ledger? How does it >> Yeah, it's not a blockchain. It's not a layer one. It's not a layer two. It's like really a network of nodes that are connected through a cryptography protocol called secure multi-party computation, but it's not like a blockchain in itself. So, it can be a service layer to other blockchains.

>> Got you. All right. All right, Lucas. Thank you for coming on. >> Yeah, thanks for having me. >> I appreciate it. I hope you have a great rest of the event here in SF. It's been a great event so far. >> Tons of great speakers. I know there's AI folks out there. There's crypto folks. They're coming coming together. So, glad you're enjoying. >> Yeah. Thanks for having me. >> Thanks, man. Thanks. See you next time. Cheers buddy. >> Alrighty, guys. We're keeping it rolling. We got our next guest, uh, Deinder Sod. He is in the back, I believe, ready to go. Um, we're keeping it rolling, guys. Hour four here.

>> Yeah, I think we're on hour 14 over the two days. >> We are um >> guess maybe 29 some. >> We're talking communitydriven AI. Dender is the AI community lead at Frontier Tower. >> Beautiful. >> Um, so we're gonna be talking about the state of the tower. >> AI floor >> at the Frontier Tower. >> Dender, welcome to the show. >> Welcome to the show. >> Please, please join us. How are you, man? >> Good. Thanks for having me. Uh, it's crazy hours at the Frontier Tower right now.

>> What is the Scoot in just a bit more. We want We want to make sure everyone can see you. >> Yeah, absolutely. >> What is the Frontier Tower? >> So, it's a vertical village is what we call it. And this idea of very different disciplines being in the same building uh takes it away from this idea of like an office space, but now you have artists and designers and scientists of all types and people who believe in things like human flourishing all in one space working together. >> And this is the tower that's uh here in San Francisco, right? >> That's right. >> Ethereum opened up uh there was an Ethereum event or there was something that Ethereum did in this tower as well.

Yeah, understanding it's the first home base, like physical home base for Ethereum in the world. Yeah. Um, and it it's 995 market. >> So, I was I was I was here last time I was in San Francisco. I went to the tower. It was incredible. Uh, there was a robotics uh stage. There was also um scientists like you said that were working physically. I remember them pouring something into test tubes. It was exactly what I pictured. And then you would go up this elevator and then there would be an completely different situation happening on the next floor. Like robotics on one, agriculture on the next, chemists on the next. Everything was all in this one tower, but every every floor was very very different.

>> Yeah. >> Um and you're saying that there's practical there there's practical improvements to having like very diverse but very close proximity of these different things. How much how much interrelated business and and you know uh work is being done cross tower stage? >> A decent amount. Um I specialize in medical device robotics and AI and my own business benefits from this kind of thing too. Um one of my clients wanted to build a quick prototype of different types of robots. I went to the maker space. I talked to a couple of artists to like make some really cool interesting designs that no one had ever seen before. We tested out different materials. We were able to show it off at CES in two weeks. This kind of thing typically takes 6 to 8 weeks minimum. So

there's value there. >> Yeah. >> So what are you building in this floor of the tower that wouldn't exist in a normal office or a normal tech office? >> I'll give you one example. Right now we're building a motion capture lab um that we're going to make available to startups around the tower at this counter rate around the world. Uh whoever wants to come in. Nobody outside of like a featured lab has this kind of a space. We're going to be the first to release it through like an actual hackathon this next month. >> And and it becomes a co-working space. Can anyone use it or what's the process for someone to get involved, come into the space, use the tech? >> Yeah, every floor has their own rules.

Um, and so my floor it's simple. You have a 15-minute interview with me. If we drive and I have the space, we'll figure something out. >> Yeah. >> Um, but every other floor has their own thing. And and depending on the floor you're at, if you're using a bunch of printers all the time, they might tack on a little fee for that. But stuff like that happen. >> And for the people watching that want to get involved with your floor in particular, what is it that you're looking for in the people that apply to use the space? I'm focused on builders, people who want to build. It doesn't matter if it's SAS or like custom uh like deep technology like uh one of our uh leaders over there is actually focusing on what happens in the future where we don't need to work anymore.

What is the posteconomic revolution like look like >> um and I think people who have interesting ideas who actually do things build things that's my primary focus. >> Very cool. So when you're building this kind of lab right for the motion and sensory and kind of getting to the next era of AI, you said this convergence of AI and robotics like where we at with that like has that accelerated at the same pace as like say models have improved or like the sensory and kind of AR and VR require is a much slower pace or does it just accelerate at the same pace as models improve?

>> I think they're kind of in lock step. The hardware keeps getting better every year. Like right now we have sub millimeter accuracy. That's the input data. And how do you use that input data? Well, we have these VAS that are ready to take in this data. And we're experimenting with different ways to make that data like actually something that a robot can use and learn in one or two shots as opposed to like a 100. >> Okay. >> Yeah. So there's a lot of intermarriage between these things. >> Okay. And sub sub 1 millimeter accuracy means like to the extent of like moving a robot's arm to like grab something, they're they're they're like they're almost as good as >> hand eye coordination. There's like a bunch of cameras in the room and they track you and then in that space the way that they model you is accurate to

within like less than a millimeter. >> So we had Greg from Akos in here yesterday and he was talking about how there's a lot of cameras in his room and they can monitor some things about his biometrics like heart rate, breathing and they actually the AI in his house knows when he's going to wake up a few minutes or seconds before he actually wakes up. >> That's awesome. adjusts the temperature, it starts his coffee machine. And so like are is this the like are these some of the things that are able to be done and enabled as a result of motion capture?

>> You could do that. The things that we're focusing on is one immersive experience for anyone to be able to like see this stuff and learn it because there's a huge learning component to our um whole space. Like 80% of the classes we run are free for the public. Um and then the other portion is like robotics grounding data and like you know you see those like robots dancing around and moving a lot of that's taught using this technology. >> Okay. Yeah. Okay. And that's how you train these robots. >> Exactly. >> Um how much you know we saw scale AI there was a lot of demand for human data. We had someone in here yesterday talking about language and voice data as a ve main vector for training new AI.

How much data will be required to train robotics on some of these new tasks? Are we going to need to submit, you know, uh us washing dishes, doing laundry, like you know, performing factory functions? Is are we starting to capture that data and use it in training some of these robots? >> Yes, for sure. Uh first, uh to answer the first question, less and less data over time because we're able to synthetically generate variations of that data. Okay. And also because the models are getting better and better. Um, but exactly all the examples that you said, these are the examples that you can capture data on and you can simultaneously capture a bunch of robots in this space cuz these are IR cameras.

They're specialized to do the thing that they do and like eliminate all of the noise from the data. >> U, so you can have like seven robots on the floor and then train at the same time. >> Can you give us just like tapping into maybe I'm sure you've thought about this, you know, recently, but also just tapping into your imagination. What are some of the top two three use cases that you think robots combined with this technology is going to be most important for? >> I think the most important thing in my mind is the fact that people are learning that stuff that used to take PhDs and masters a fifth grader can do now. And I think that has a societal impact unlike anything else in the past.

Like AI is enabling this kind of learning that just wasn't possible in the past. We didn't have the hardware. We didn't have the software. Um but obviously there's the technological advancements like factory automations are going to happen. People are going to learn how to utilize these things. we were going to become managers of hardware and software and like systems as opposed to doing a lot of this work ourselves. I think that has a lot of value uh in and of itself. >> Yeah. >> I saw a video when like all robots trying to like do dishes and epically failing. Yeah. >> But then you see like a video of these the Chinese New Year celebration. These robots are like hitting the dances.

Yeah. They look great. >> Yeah. >> So like how does that happen? How are you like such a how are you a great dancer in tandem with in in sync but like you can't like do the dishes. >> Yeah. There's two aspects to this. One, dishes are a new scenario every time. The dance movement, they've like pre-recorded and practiced on that stage a hundred times. Yeah. >> The other aspect is that it also depends on the model. Like 1X in like a few like its early days wasn't doing so great. Open Droid came became viral with this video. He's also a tower company because they were able to do the same thing in less time. And as you're able to generalize over different scenarios that exist, you create these things called world models. uh where like you can express the world in a way that you can

actually interact with it. Uh and that's where sort of the wins are going to be I think in the next year. >> World models. >> Yeah. >> So this is like retraining or or the initial training of a robot on like what the world or like like a a set of physical practices in the world that like humans aren't so easy to do but like for them it's hard. >> Yeah. Yeah. Teach it by example and and just like we do with deep learning models, you create a bunch of examples. You teach essentially these patterns that get embedded into much much deeper networks and then you see if those patterns are working and if we can apply it to different situations that aren't exactly the same.

>> So in this room in the tower like are you going in there and like doing something with your body and then the robot is supposed to like copy what you did. >> You can do that. Yeah. >> Yeah. But there's got to be a better way. >> There's there's so many better ways. There's synthetic ways to do it. You can like run a 100 sims in as a sim, but there's also certain things that you have to do physically to be able to like um manipulate a thing. And this is just one example. I want to be clear. We have like the world's best biotechnicians. Like there's a there's a woman who's figured out like how to get stem cells from like menstrual blood and she's running an amazing company called Muse Bio from that. And there's this human flourishing that's like figuring out

like how do humans kind of like the point of this uh thing for me is like where do humans come into play with this whole AI thing, right? Um and how do we build systems that are valuable for both us and the technology so we're not all freaking out about it. >> All of these things there's longevity, there's robotics. Obviously, we have robot fights happening now. That was the first robot fights happened in the tower basement. The very first robot fights ever >> happened in the tower basement. That's what became viral and I saw that >> everyone was doing it after. >> Yeah. >> And so talk to us. I we'll we'll quote tweet the clip of you talking about this on on these robots fighting because that went mega viral.

>> How did that come to fruition? How did we end up with these robots? You have 12 floors above it. How did we end up with these robots sparring in the basement? >> So this is a great example of that cross floor collaboration. Vitali and Zenya run UFB. They're fantastic roboticist. Zena does amazing marketing, but they combine powers of the arts and music floor where they understand production and like how to make like the pump and like awe of all of this stuff happen. Kind of like making it like old school WWE. Yeah. >> And then they combine those. Cuz if you have just a robot doing this, early robots, all they were doing was this.

It's extremely exciting from a technical point of view. It's not so exciting to look at. But when you have these artists who can like, "Yeah, we can work with that." And then they set up the stage and they create these crazy signs. I don't know if you remember the crazy signs they had. It was like, "Is your robot stealing from you?" Like all of these >> Yeah. There was a whole boxing ring and then they had the signs on the and there was people cheering on the sides if I remember correctly. >> Yes. All of that. And like the crazy referee guy was just so into it, right? >> This creates a complete product. >> Yeah. >> Yeah. >> Um >> it's happening, >> right?

>> Yeah. So if AI is all about returning to philosophy and what we enjoy and sports and leisure are are we going to also replace sports with like watching robots or no? >> I think AI is about all that and more. But at the same time what you're asking I I certainly hope especially like the things that like where you're hurting each other or there's things that humans can't do. Uh it'd be more fun to watch the robots do it because >> new sport. >> Yes. >> There's a lot of brain new sports. It would be funny just to watch robots jump off of cliffs.

>> Just like just crash, >> just brain just anything that's brain damaged. Just have a robot do it instead of >> Yeah. See if that's a better and maybe hopefully even just as decent for the environment. But like, you know, it's been in Las Vegas, in Boston, they went to Boulder, they went to Denver, they were at East Denver. Like they're a circuit now. >> Yeah. Um and so this was born out of the tower. How does the tower get funding? Is it donors? Um the tower itself has its own funding to actually pay for the building. Then there's the monthly fee that you have as like um like a regular member kind of like a weiwork does.

Okay. But then you can get involved in other ways too. So um I as a floor lead actually fund a lot of the activities on the AI floor cuz I wanted to kickstart something. Like every single piece of furniture, all the equipment I bought out of pocket. >> Um and I set that up because I had a vision for what this place could be. Yeah. >> And we're starting to see the returns already. So like it's it's getting to a place by like this December I'll be skyrocketing. Well, Dev, thanks for joining us, man. This has been insightful and fun conversation. I appreciate you being patient. >> Yeah, of course. Thank you. >> Again, would like to come to the tower. >> Yeah.

>> Fight some robots. >> Yeah, you should do. >> You really should. I like I said, I was there last time and uh there was uh >> there was robots or not. >> I didn't see the robots. I walked every floor. Um at the time, there was an Ethereum event happening on one of the main floors. So, I saw Tamas, you know, the the ex uh, you know, executive director. Um, and I got a chance to talk to him for a little bit. He was on a panel and so a lot of people that were on the upper floors uh, came down and were just watching the panel. So, I didn't really get to see a lot of people like in action.

>> Love to see some robots fighting. >> Yeah. >> Guys, we've got our next guest, uh, Escobar coming and joining us in a sec. We're going to take a quick break. Be right back as we slam through the rest of these guests. Heat. Hey. Hey. Hey.

All right, guys. We are back live uh here with Escobar from Niron. Uh pleasure. Absolutely pleasure to have you on the show, man. >> Nice to meet you. Pleasure to be here. >> Likewise. So, uh we're on uh you know, we've we've been here in this room here for a few hours today. did the whole thing yesterday as well, but uh there's lots of panels going on out there. Near has been shipping, lots of product announcements. Give us a sense of what you've heard on the panels and generally what the vibe is out there. >> Oh wow. Uh well, I I love the topic yesterday about Neoang. It's my favorite topic of the of the year at least.

>> Nail bank has been killing it. You know, everyone has been issuing a card, but >> um NIA built different and how we going to deliver that. I believe >> we're going to have a very different infrastructure for that. >> Yeah. And so now that you know it's very easy to launch a neo bank, right? It used to be that you had to get a bank charter. You could launch a bank, then you could issue a card. Visa, Mastercard makes it even easier. And then obviously from there, neo banks are are just exponentially easier from that. Um why do you think it's important that we have so many neo banks, so many cards? Um you know in digital assets we saw a lot of fragmentation. Does it feel like fragmentation or does it feel like more options? I think it's it's remind me of the early divide day right like we have

a lot of diva protocol just issuing one another you know a lot of lending following each other >> so one big protocol issued and then everyone doing the same thing >> but I think in the end um we just going to have couple cards that that sustain in the market that really doing the business because >> um I think >> the current the current meta is no longer about selling your dreams right it's about how you can generate revenue and be sustainable business. Um so we believe that um near infrastructure kind of deliver a different way of how we frame bank.

>> It's not just simply cryptocard and doing a payment. Yeah. >> Yeah. Um and how does someone stand out, right? Like how does someone differentiate, get some product market fit, you know, everyone needs to use a card, but there needs to be some distribution in order for, you know, a card to get used and generate revenue. Uh what can a neo bank do to stand out amongst all the others? Well, I think from the from the current nail bank that we have in the market already, >> it's simply um a one-stop shop that really built everything in in the app like you if you see near.com is something that I really like it.

>> Uh it's just about time how we can implement a cart into that product. >> I think that could be very interesting especially for RIA itself. We are lending protocol on top of intent. Okay, >> we just see um how we can implement either is it necessary for NEO bank to have a card or are we talking more about infrastructure for stable coin as well? Right. >> Okay. And stable coins right now are the primary currency that people think and anticipate is going to be spent and earned and saved. You know, the primary currency through which a lot of the agentic payments are going to flow. Um >> given the current landscape of the internet, most of this is not yet built for stable coins. Sure.

>> Most of e-commerce is built for credit cards, right? Someone can just go on a browser, fill out the credit card in a in a field as they fill out the payment form, uh, buying something online. >> Um, does it make sense for AI agents to have credit cards in the near term until we build out more stable coin infrastructure where a lot of the internet becomes retrofitted and optimized for stable coins? I think instead of doing through credit card why don't we just do it directly to the stable coin infrastructure right like um the current the current uh credit card payment and so on is it still have a lot of implementations on the fee side I think what we need to tackle on the

stable cocoin infrastructure is mainly doing direct payment less fee more straightforward I think that's something should be improved instead of going through the credit card worry yeah >> and and tell us a little bit about ray of finance how you guys have approached the space generally you know what is the problem statement and how do you anticipate solving >> sure um so uh just a little bit background on ria itself first uh we are uh the first and the biggest decks on land and landing on top of near >> um even until today but we we already live since 2021 it's quite old project in the market um yeah we are the first DAX first landing on near and then uh previously it's called ref finance and borrow finance probably the OG on the ecosystem might know that but then we merged the uh protocol into one single

entity because all this time it's operating under the same team. So it just doesn't make sense to run two token and so on. So we try to make a better valuations in the market as well and presenting in a better way. So we launched uh our uh protocol as RIA uh last east Denver. So East Denver 2025 >> uh which is uh one year ago. Uh we we rebrand everything and uh July 2025 we merge into new token, new valuations, everything. Um how we tackle the market is aligning with near intent. We see uh the PMF is coming from the uh consumer app. It's not just about infrastructure anymore. So we develop a crosschain lending protocol where we basically

allow all this non-s smart contract chain non EVM chain to supply uh assets natively and build a collateral positions and they can borrow against that across all the chain. So it's a unified position. >> It's not isolated liquidity. Um I think that's something that we want to tackle because um if we see the top 100 uh token maybe >> in a coin market cap that uh EVM is one maybe only couple of them. >> Yeah. >> The rest is non non smart contract chain. Yeah.

>> Yeah. Yeah. And you know near intense has so much product market fit. Uh now we're starting to build use cases around near intense. Ilio was here just a few minutes ago. He was talking about the AI marketplace. Yeah. >> Uh this is built on top of the near intense technology and now this allows humans to submit intents to AI for work. AI can then uh submit uh the work and the uh human can get you know uh what it paid for. There can be a dispute. Uh AI agents can also start to hire other AI agents. So this is an entire AI marketplace uh for people and uh agents to hire each other for uh different pieces of work. All of this is built on

the near- intense infrastructure. Um, and then, you know, as you're talking about different lending and borrowing capacities, I can also anticipate how, you know, some of these work uh streams might require some capital. >> Yeah. >> Um, and I'm curious how you think this landscape could evolve when more and more of these these tasks that we're we're hiring agents to do require some capital. Where should they get the capital? Maybe in the initial, you know, form it comes from the person that hires like, "Hey, I want you to go uh build out this algorithm. Uh maybe build out that, you know, say a social media app, right? Uh that's going to require some APIs and and different, you know, it's

going to require some money and go maybe some other agents. >> Maybe I have to supply, you know, a million dollars when I make that hire, right? And then it can take my million dollars, it can go, you know, hire a bunch of people and create the app." >> But what if I don't have a million dollars? What if uh you know I want this app to be created. I'm very good at writing the prompts. Uh is there some other lend borrow facility that could help provide that capital so that uh I can do the prompting, the AI can do the work and then there's some lender uh that could provide the capital and get a return.

>> Well, um interesting. I think I think that's one of the use cases that we want to uh grow through uh because right now um we see the lack of the capital efficiencies that we have. We have around 90 million TVL uh still sitting in a near uh uh smart contract. Mhm. >> But how we going to use this capital across all of our D5 protocols, right? We have DAX, we have lending, even in we have all of this um uh D5 primitive. It's LST, whatever you name it, including leverage.

>> Um so uh um I think I think that's that's the first things that we need to solve first before we go to further on how we can build a consumer that uh a consu consumer product that they can leverage to to uh leverage our capital, right? >> Yeah. So it's still those uh users who are holding an assets that's not related to stable coin directly and how to tap into stable coin um passively. They simply can borrow with us uh by supplying any other assets that they have. >> I think that's the uh the main thing that we want to solve first.

>> Yeah. >> Um and and there's m you know lots of different asset classes now that are getting tokenized right and we're bringing more high quality assets on chain. Um if you look at you know the lend borrow facilities that are out there now it's primarily focused on just you know we got Bitcoin we got ETH those are primarily the speculative assets and then you've got a whole slew of stable coins. Um how do you think you know tokenized ETFs tokenized equities tokenized AI agent you know yield coins and and like this entire asset universe as there's a more robust asset class being tokenized putting on chain into DeFi. How do you anticipate Ray of Finance will react and kind of reshape around some of these these new asset

classes? >> Yeah, so there's there's two way uh for from my perspective at least uh one that we are very anticipating right now is to put uh those assets into the lending protocol for sure. They can use that and borrow stable coin. That's one thing >> and while they still get the upside of the of the equity when there's uh price movement uh in the market, right? M >> uh that's one thing and the second thing is to build an earn product where they can benefits from three of this um uh three of the aspect which is the equity itself the growth of the uh of the value and how um how they they tokenize and get the yield. So >> I'm not sure either either market right now there's someone delivering that but um that's a huge opportunity uh we see >> because um especially those assets that

related to uh near uh narrative wise >> I think that could be very interesting. >> Yeah. And um you know specifically around near and DeFi elements what what are you most excited about as as you kind of look forward to the rest of the elements on near's road map? Oh uh I can say privacy is a thing. >> Privacy >> yes because um for RIA even um we try to build further than just providing infrastructure for near >> uh we have been delivering that for years for four years since 2021 but then we see a great opportunity how we can serve the privacy market. So we built a Zcash extension wallet >> because we see uh Zcash only have the application wallet right so far. So that's one of the thing that we want to

build uh for the new community we can say but uh leveraging near intent that means most of the uh Zcash uh maniac or those who love Zcash already at first place they can leverage near intan to build more privacy model on the DeFi side. I think that could be something killing the market. >> Yeah, incredible. Well, all the things that Near is shipping, we're going to need more DeFi uh for all these agents as well. So, thank you so much for coming on and uh and and sharing a bit about >> RA. Thank you so much. >> Pleasure.

>> See you next time. >> See you next time. >> All right, guys. There we go. Uh we are on the home stretch here at Nearcon. Man, I'm just we're just pumping pumping it out, man. There's so much to talk about. We can just keep talking. Everything everything that Near shipping is incredible. Um, I think that, you know, a lot of this stretches all the way to the bare metal, deep infrastructure part of the stack. We talked about the, uh, the energy crisis that is a foot. Um, we're seeing this span geopolitical territories. Um, a bunch of different countries are approaching this a bunch of different

ways. Some people are going solar, some people are going nuclear. Uh ultimately a lot of what we're seeing these AI agents do and the amount that we're seeing them embraced in the marketplace um is going to require a lot of compute and a lot of electricity. And so we're seeing the power crisis uh really take hold. We're also seeing uh how AI is starting to affect other markets. You heard a lot of us talk about Citrini's research. This was daisychained together different components that ultimately might lead to a scenario where we have uh several different facets of the economy starting to uh react to AI. Uh and the way that the they react to it is

that um they end up cutting uh employees. This leads to more uh unemployment. Uh that does help the businesses because they're wiping these expenses off their books. It allows them to have better margins, beat earnings, stocks go up, but they end up investing that appreciation more into AI. What happens? They end up uh replacing more of their workforce. So on and so forth until the consumer credit uh economy collapses because of unemployment. Uh this leads to a daisy chain of effects all the way through the economy.

We'll see if this ultimately plays out. Most of our guests have said that um this is fear, uncertainty, and doubt that is being spread. Uh ultimately, we will find new jobs for these people that are displaced. Uh even though we're going to have AI agents manage the new jobs, seems like that's not entirely the sector that we believe these jobs will create. Uh most of these jobs are going to come from creative places. There will be art. There'll be sports. There's going to be more pushing the frontier technology. I'm ready for our next guest. We're just We're cruising. Near's out there shipping and we're in here cruising. Guys, let's do a quick commercial break while we figure out our

next guest. We will be right back after this. Don't go anywhere. Back here live at Nearcom. Con. It is

Tuesday, February 24th. We've got Richard and John. Richard, the MP of Fabric Venture. John, partner at Kinetic. Guys, welcome to the show. Uh, local SF natives or no, Fluid. >> Yeah. You don't sound local though. I'll tell you that. >> Further away. No, no, based in London. >> Yeah. Yeah. >> Based between Hong Kong and Puerto Rico. Mostly Puerto Rico. >> Okay. Nice. Nice. Well, we're here at the intersection of uh blockchain and uh agents uh agentic commerce has been a big conversation. Obviously, stable coins. I think a lot of the current crypto rhetoric is don't worry guys, just hang on. Aentic payments are going to save us and we're going to have all these agents on chain. All your DeFi

protocols are going to be rerated. Uh stable coins are going to be in trillions and there's kind of this air being sucked out of the room at the same time. Guys, where are we in the state of of affairs? be it in the market, be it in the adoption of this convergence, very kind of broad here. Take it where you will, please. >> Yeah. So, real quick, I'm John with Kinetic. Uh, we are a multif family office, uh, based between Hong Kong and Puerto Rico. We've been investing for the past nine almost 10 years, which is crazy. Um, and we've done everything from seed rounds in Salana and Polka Dot to seed rounds in Block Damon and and Hyper Native and companies like that.

you know, we still, you know, I I don't know if many people are asking us, you know, when alt season because I think we're kind of past that illusion, if you will, but I would say, you know, alt season, i.e. anything that is not Bitcoin, Ethereum or stable coins, you know, we still have to prove product market fit. So, I don't really think about alt season anymore. I just think about when fee season, you know, if you have a networkbased product or service and if you're showing fees the the way that we're seeing with near intense now, um that's exciting because when you hear about fee switches, that's a great term that people love to talk about right now. Obviously, we've seen that with unis swap now a now near intense. I think that's kind of where you start to get excitement. Um I've been talking

with a lot of projects here at the conference who some of them are thinking about, you know, we're going to launch our service. Do we launch a token? To be candid, focus on the service, focus on fees. If the fees are instables, that's cool. Once you have those fees and the market has a way to price the value of your network, then you can launch a token. And you know, ideally in a better market where there's more liquidity, it's great when there's lots of agents, but if there's no liquidity coming from people who are deploying these agents, I don't think that changes a lot other than transaction volumes and that transaction counts. And I don't think that really moves the needle for anybody in our space. So this is definitely a moment of reckoning and there are a lot

of projects out there who I think will make it. Um but yeah >> um yeah look I mean we didn't do ourselves an immense amount of favor collectively because you know the ability to speculate around tokens is clearly a feature not a bug. It's required because you want to build that uh ownership from your community so people can have that participation. That's part of what can help make all of this movement, you know, legitimate. Um, but then it obviously becomes a a problem uh when you get the ramp rampant speculation and everybody goes for kind of sort of token liquidity fit before they think about product market, you know, fit in any way, shape or form. And at least I I remember, you know, being old fashioned for when I was, you know,

building products, that shit's pretty hard getting product market fit to work. It's it's like you you cannot focus too much. You cannot work enough hours. You cannot be too lucky um uh you know to get to make it happen and and folks have managed to coast uh without you know doing that. Um but you know but as John says I think um we are seeing this this refocusing back on fundamentals and revenue and I think that's good but I do think there's also been an absence of kind of real focus on the science of token economics which is its own kind of product market fit. I mean a token is a it's it's a product financial instrument. It's a technical product.

You need to make sure it kind of it's kind of like token community behavior incentives fit. Um and and I think there's been way too little iteration around that component of the of the products as well. Now the good news is um as long as we don't think of this shift to an agentic world where you know to to this autonomous world as being just about AI agents that trade for us and generate a bunch of transactions to on a kind of like a kind of a autonomous wash trading kind of basis. um but rather do we think about true applications that can be built in that environment in that framework that that ultimately somebody is paying for at the top that flows through this you know

agentic framework um and then there are some fees paid to near intense hopefully amongst you know others um then that's a whole new kind of new dawn um and and and like for us um you know fabric kind of we were sort of doing AI before we did kind of you know web 3 even when AI was definitely not cool yet. Um so big when it was still called big big data in in fact >> um and um we there was a kind of a sub thesis or phrase we had which is that part of the point of web 3 was to to use its coordination superpower to organize the world's data from the bottom up

>> right >> and hence provide this kind of contextrich environment that makes AI sup super potent super powerful um and also you know it does present that's the possibility of um protecting the privacy around which data you share and you don't. It's one of the kind of primitives within our sort of space and also tokenizing intellectual property. So I think all of those aspects and I think there's sort of a growing clarity in the industry around that uh you know give us the opportunity to build infrastructure to serve apps that are that could really >> you know save our bacon. So there's a there's a kind of reductive frame of thinking going on right now that

blockchain is just for finance. >> Like we tried a lot of these different gaming iterations, social iterations, other types of uh financial kind of like experimentation, but it kind of came down to DeFi is the only use case, which kind of ironic because DeFi fell out of the lore and out of the flavor for like the last two years. It was really hot 2020 2021 >> and 2022 with a lot of like the blowups and a lot of the CD5 stuff then it was about chains about all you know scaling execution and throughputs and all these things fast L1's fast L2s now it's like okay all the institutions want to do DeFi but now the reductive thought is all we are doing here is DeFi but then you've got like pioneers like Vitalic who have these crazy visions about in uh you know identity and and all these things and

now obviously the agentic forces so I guess All this to say is is it too reductive to say it's just finance? Is that it's just revolutionizing finance? Why or why not? >> It's great clickbait for uh for Twitter. But John, do you want to go first? >> You heard it here. I you know, is it reductive? Yes. Because I mean, do we think that decentralized networks offer anything other than a great settlement layer for for finance? I do. Um you know, I I came for the tech uh and I'm staying for it. And not because it's the tech, but it's the outcomes. And you know, one of the things that we've been talking about here at Neircon is, >> you know, people buy outcomes, right?

People want either you're going to help me get a positive outcome for things I want or you're going to provably prevent negative outcomes I don't want, right? We were talking about that a little bit on stage. And I I think the the big thing here is our ability to first of first of all trustlessly verify either of those outcomes. Decentralized networks provide that to some degree. Now, they need to be efficient and quick and they need to have all the trappings that everyday consumers and businesses want, right? They're not people are not going to wait and pay extra for things that they don't even know they think they value, right? So, I think the emphasis on decentralization I the number of projects I've talked to who want to offer something where they're like, well, we don't really need anybody to know that it's on a network. Cool.

Then maybe take Block out of your name or chain out of your name. you know, really focus on what the outcome is that the customer is. And so I think we're getting back to kind of core brass stacks of building businesses. They can be on networks. You can have a floating asset like a token. And that token and that network is really maybe being engaged with by people who provide nodes or people who are participating in the governance of that network. The consumers don't have to think about it. But if consumers will pay fees, >> there's going to be a network of folks who want to provide that decentralized service. So decentralized networks are not dead. We just need to be a lot more discerning about what we're going to say about it. And it's not just about uh you know gaming or memes. And by the way,

just to get back to the gaming part, I think the challenge is that we were trying to drive speculative interest because there was you know web 3 jammed into a game. Were the games necessarily fun? Right. So like to me and by the way uh not naming names or anything like that but if your if your practice was to drive speculative interest on a game because you jammed in a token well here we are you know so I I think everybody's discuss they we just have to evaluate what it is people were willing to pay for. I mean I think better on the games side actually um using AI to drive you know more interesting gameplay but also driving the distribution side is is something that kind of we've been

looking looking at you know it doesn't have to be the kind of the ownership economy applied to to games at all but but going back to this question of like is it overly reductive um I think there's sort of three three kind of quick responses hopefully that I have to that one one is you know we the big victory around stable coins got sort of stolen from the industry. I would say you know us um and and I was thinking about this earlier today. It's a little bit like um it's what we were we're a bunch of kind of subsidized open-source laboratories who built some cool [ __ ] that Tradfire came along and said, you know what, that can solve a bit of a problem for us. In fact, I'd rather we took it and solved it for

ourselves rather than had someone come and um you know, you know, mess us up. And so they've done that and it's kind of I think it will around different types of payments from quite small frequent payments style all the way through to treasury movements. It's going to rebuild and is rebuilding um you know sort of trout five you know from the kind of core out we've got like 20 just in the current like 20 whatever plus sort of investments in around that. Um so that's the first thing but that's it did get in a sense sort of co-opted or stolen but it's nonetheless real uh and we may wish that in some there's more of credit but fi fine the second thing is um you know it's I think it's a Jevans paradox kind of kind of argument which is that that yes you can reduce it

to just finance but you know what like financializing or kind of the marketification of the entire world. Um uh one of our sort of speakers at one of our sort of fabric summits was a guy called Glenn Vile talks about the radical markets in in his book whatever >> and he did a paper actually on soulbound tokens or whatever with with Vitalik he you know that's that stuff uh which is you know we're just inching towards that direction where there's sort of no such thing as private property and everything is public and constantly traded and whatever. um that's not just anything that's that's totally radical the whole world >> you know so and and so every and it's jeans in the sense that once you reduce

the cost of it and the friction of creating those markets and they were programmatic you know that's this it goes everywhere and create um and then so then and then the the third thing is you know what I guess we're talking about which is but it's it's tougher to do which is that the marriage with AI about organizing the data uh about pro providing verifiability for the operation these autonomous operation of you know that of agents whatever they're doing you know but all of that again to John's point the in the sense it's easy it's coming it's architectural it's happening it's working it's scaling there's a bit more you know whatever is ERC 8004 about you identity and reputation stuff that's coming I think I

think David from Ethereum is actually talking right this second if I'm not not incorrect and or Davidid and and um you know that's all coming but the harder thing is getting people comfortable as we will actually say on the panel that it's legitimate that there's adoption in a you know it's like a classic adoption a small self-referencing group of customers who say yeah I vouch for this this really works let's let's go for it that it's sufficiently well packaged >> that people can adopt it that's not something that the little web 3 industry has been used to to saying and we do have a a a challenge with a reputation and and and gaining regaining legitimacy. Um I have a good friend who's um and I don't

think he'll mind me mentioning his name, but Mike Butcher, who was the former editor and chief of of Techrunch in Europe, and I was just what's happening with him, whatever. And he's like, well, you know, I don't really talk about web 3 anymore or whatever, because it it has been it has got a kind of a uh yeah, machavelian or even kind of tricky. soiled soiled repetition to deal with. >> Um and I and I think we need to get we need to get over that. But perhaps even harder than that, we got to solve real problems. >> So you got to find the product market fit that we had mentioned before. I'm curious on the tokens themselves. Is it just that these tokens are equity and you know these are companies that are searching for product market fit looking to generate revenue? Seems like more and

more they're going in this direction. But you also mentioned that these are networks. They have network tokens. And so how do these these two models fit? There there was a recent article uh Jake Tvinsky and I think it was Jesse Walden put this out. They said offchain revenue stays offchain goes into the equity. Onchain revenue stays onchain goes into the token. Is this the right way to think about this or is there network and there's intermingling models here that we should be untangling? >> The market has shifted so hard on the pendulum towards >> what is the revenue? what is you know PTE what is what is this protocol going to generate tomorrow >> like the growth story for me's law is completely gone right now it's eradicated so I think that's kind of

some context with regards to like where do we go from here as far as these tokens >> is it one I think you can tease one element out of that it's like you know from from our perspective as kind of backers or whatever fabric I'm I'm guessing John as well is that you know you want as much alignment with founders as you can get. So this question of like is it tokenized equity at that level is like both you know whatever the founders got we would hope you know this is what you want to also participate in as a as a backer so that's how you can create alignment and then so that's one maybe one aspect of it but I mean I'm not quite sure I get immediately and I may not be clever

enough to work it out why you should restrict sort of offchain revenue to being equity and and you know onchain to be to be tokens I think >> verifiable onchain and if verifiable We can't verify anything offchain was my impression of the article. >> Right. Okay. >> Or maybe I know it's a regulatory take as well or whatever technical take. But I mean for me it's it's like if if you've got you're going to design a bunch of mechanisms either involving just your tokens or just your equity or both of them. And obviously then you know you as backers want to own whatever's whatever's appropriate. Um and I don't see why you shouldn't f you know if we want to have if you take for example um should we say listed entities that have tokens on the balance sheet to

provide economic envir alignment with the rest of the ecosystem that they from which they have sprung. I think I think I've heard a few people toyed around with that that concept and and and and I think but but um avoiding the sort of threeletter word um I I think that that's a pretty good and actually why I was declare you know I'm I'm on the the board of sovereign which obviously is out of the ecosystem but I but you know I think >> if you want to have a business in mind you want to build which is take to market something fabulous like near intense if you want to build a kind of you know private GPU cloud for AI um uh and you want to make sure that your

bread is buttered through the appreciation in the in the value of of you know the price per per token and hence you're aligned with the rest of the ecosystem that's a good way of doing it but but you might get a whole bunch of offchain revenue and and and use that uh in order in order to to do that so I I don't I don't quite get that one but again maybe I don't know the exact point they're trying to make. Well, then maybe just like holistically, are we overfixating right now in the bare market on like a short-term revenue and not like investing in the growth stories? >> What do you mean we're human? Of course, we're overixating that like back in the back back in the bowl, we'll just be excited about growth.

>> It's Yes. But it's really interesting trying to kind of get your head around that. I mean know fearful greedy greedy fearful kind of thing you know um and how to get that perspective um uh at the moment and not I mean I think that product market fit and customers and [ __ ] that people want to tell their friends about and use and you know and reducing the time to kind of joy in using a product that's the thing that it's almost like the tougher it is to focus on it the more you need to just make sure you continue to do that and then that's good for building your your product and your your project. Richard, when was the last time you did an onchain transaction?

>> Don't mean to throw you down under the bus here, but just curious what our local VCs are are uh doing their research these days. >> In the last week, I would say. >> Okay. >> Yeah. I would like I would say in the last week, >> anything more than that, it would be it would look a little rough. >> Yeah. Yeah. But also, but I flip it another way. I don't think humans really should be messing around, I agree. I mean >> messing around with too many onchain transactions. >> It's dangerous out there. >> Well, I think >> well it's not that it's dangerous. It's like it's like you know because it's like you know you stumble and you're going to have some issue. It's like it's it's it's non it's non-trivial but you want to be shielded from it. You want to be you know there's nothing wrong with

being shielded by a great app. >> Of course >> I mean there's an onchain transaction having behind it or having the machines work for you. >> Yeah. That's right. I mean, do I have to do I have to interact with an RPC to do a transaction or can I use a front end? Yeah. >> And that front end is owned by a for-profit company and maybe they charge a fee and maybe those fees actually go into their own onchain transactions. The thing is to making that kind of hard distinction between off-chain revenue and onchain revenue. Look, the the fact is is that most networks have some kind of labs entity and they're exposing a very easy UX for enterprise for small medium businesses to interact. those businesses aren't ready to go do their native transactions. So, they're going to pay with stables and what are we

doing as the front end of that? We're going to purchase services by paying the fees, you know. So, if we are an abstraction layer to do that, that means that there's adoption in a lot of ways like by making that distinction, what we're saying is we really need to double down on this definition of what's on chain versus off-chain and block, you know, what's decentralized and what's blockchain. And and I don't mean to speak to those who wrote the, you know, the the idea or the report, but I I don't know if we need that distinction for this to be productive. If people are willing to pay in any form and we as participants in the network are competing for those services, you know, right now we have a glut of block space, but there may come a time where a network, especially an app chain for

Vitalic, is so hot and so wanted that we're competing to use that service because it gives us a set function and value, >> right? Then why shouldn't that underlying token float higher? Because we're competing for it just as we pay varying oil prices. Is oil a stable asset? It's not. The last I checked. Okay. So, you know, anyway, I'll get off my soap box now. >> Yeah. I mean, look, it's definitely it's definitely >> guy who did the CRA thinks he knows everything. >> I'm not that clever. >> Yeah. Economics or economics or or good or good fortune.

>> Keep it up. but um or or something else. Um but um look, I mean look, it's it's right there's another aspect of it is like it's right that if you're trying to understand what's happening in in some tech frontier that you should get hands-on like I mean we have a thing that like you know you know [ __ ] the PowerPoint and let's look at the the product. Let's just cut to can we cut to the to the demo and and go through it and let's tinker with the thing and there's no better way to kind of learn about it. But but the other point is also valid like and if you apply it maybe to the AI world instead which is like you could turn around to some programmers and say when was the last time you read some code and did some

debugging. Well increasingly they won't. In fact you know there was a debate you know kicking around about whether or not there's going to be you know um you know sort of machines like using highly you know typed kind of you know languages that's but because that's easy for them to cope with. It's harder for obviously for for humans versus sort of scripting and um and so there's a question well which languages are going to be more prevalent now that a lot of code's being produced by machines but but one of the arguments is possibly no languages >> because you because you'll we'll use our language and the machines will talk directly to the other machine in machine code >> and or you know define the machine code that's program that machine program their machine they'll talk directly to

some native interface >> and and there'll nothing that humans will be able to look at. So then how much programming will the program programmers be doing? >> It'll be a different different you know paradigm. >> I think rather than us racing for the you know to win cred uh I I think the most important thing is whether you're an investor or you're building or by coding a product >> you need to be focused on the the marginal limiting factor of your success. And if you're focused on that you know whether you're doing with prompts or whether you're actually coding or inspecting code that's really the key now. I mean, at the end of the day, again, I I I'm always a full believer in outcomebased decision-making. And if we're trying to drive people to make a purchasing decision to use our product, you should always think about what the marginal

limiting factor is on their success. To me, you know, coming up with a value proposition is pure empathy. You're constantly practicing empathy. This is what I tell all my ports. And if we are you know at this point if our customers or our users you know we need to understand what is a step function of value for them. Um I think we have spent too much time in our community thinking about cred and culture is important but just like cred and bonafides and whatever and we're not thinking enough about who our customer is. We've been selling to each other for too long.

>> Well agents are increasingly customers and users. >> Yeah. Similarly things are about to change. >> Gentlemen, thanks for joining us. Yeah, absolute pleasure, guys. That >> was fun. >> Absolutely. Appreciate you guys. >> Great to meet you guys. Very cool, guys. >> All righty, guys. We're going to keep this thing rolling. We got one more hour left. >> Yeah, 1 hour. >> Hour and a bit. Gentlemen, thanks again. >> Four guests left in the last hour. Is that right? >> Next, we've got um we've got our lovely friend from eBay uh talking about Secure MCP as well as generally the state of Agentic Commerce as it pertains to eBay.

Uh, we can't really talk too much about e eBay's plans. Um, but we can talk about the agentic commerce landscape at all. >> Oh, perfect. Yeah, we'll keep this one broad and philosophical. Your favorite. >> Really had to make sure that we got that through the PR guys eBay. >> Of course. >> Um, >> no problem. Oh, will we uh Yeah, we'll pull the fourth camera out. >> We will. But guys, I mean, I hope you guys are enjoying today's session. That was fun. It got a little rowdy there with our with our lovely VC friends. It does.

>> Our our Salana seed investors and um you know they they definitely hit like a 10,000x on that one. >> Yeah. >> Um >> in size. >> Yeah. >> Guys, we're going talk about the state of Aent e-commerce. Quick story for you. Um when I first started online business was this concept called drop shipping. kind of thought it was eh I didn't really know but I had some buddies who were who were picking it up and um you know we did a we did some testing and some different products and I'll tell you what if the agents were running the store doing the doing the commerce for me and I all I had to do was write copy or had my agent write copy to convince another agent to make a buying decision that would be a lot different of a buying decision than it would with a

human so I'm trying to think about what is the copywriting framework for a gentic e-commerce You know, you you my my friends have all these MX points. They spend millions a month on ads and they're selling to, you know, typical older people, men and women alike. >> Their buying decision is different than the buying decision of agents. That's right. >> Why don't we talk about the buying decision of agents? We've got So, Rab here, the an engineering leader at eBay. So, Rab, welcome to the show. Uh, welcome to the rollup. I hope you're having a great Neircon.

>> Yes. Yeah. Amazing. Uh I just had my presentation some time back and uh it was really nice like interacting with uh like the brilliant minds and exchanging the thoughts. It has been really wonderful. >> Good man. I know we wanted to talk about >> quite a bit similar thing here. Exchanging >> ideas with brilliant minds. >> Inspiring firing. >> Let's go Rob. >> Amazing. Yeah. >> So we were talking about the buying decision of agents. As more of these agents become the users of all of these websites, we instruct our agent to go and make these decisions on our behalf.

They have different preferences. They emulate us as much as possible, but inherently they're different. How at how how should we be reacting to this, right? >> Yeah. So, uh yeah, definitely like as you said, you know, because now in 2026, now it is going beyond the chatbot. Now we have the agents and now more like the companies are coming up with the protocols and industry leaders are coming up with the universal like commerce protocol and again like already few industry leaders have also started the buying agent like um which can buy on behalf of the uh user and one uh strong advantage of it uh that because like uh the agents are able to think beyond uh you know like uh they are not

just like performing the automation not just like buying a particular ular item or like that but agents also have the capability to compare uh across like uh the inventory which is there. Uh so that is one of the um like uh the uh the strongest point for these agents but as you mentioned about the uh like what should be like careful. So uh while using these agents we have to also be at the same time uh like careful about the how much secure it is like we as a user when we are using any uh agent right and also like the companies which are like producing uh the agent and like launching the uh like automated like like the buying agents or even like

selling agents. Uh so we have to be careful about uh to be like secure like um like there should like there are like different uh things for example the most common attack like the prompt injection right like where like any uh user can um like or any the middle party can provide the malicious instruction and then uh it can ask the LLM to perform or invoke the tool which is not intended. So that uh like we should also be careful about and also another thing about um I believe like there should be like sometime human in the loop at least for some threshold uh like um you know if it goes beyond the threshold there should be like checks and balances in place about approving it uh rather than you know

completely like automating it but definitely exciting uh like this year because uh for the e-commerce like the uh agentic AI uh it goes beyond just like performing the automation it goes beyond like thinking then comparing the inventory and yeah >> they're comparing it for us right there's a million different options there's so many things that they can go and and and yeah >> split testing ads you can run a like I I okay so but let's just talk about like like what is marketing to agents online to make a buying decision right let's say right because they on a on a Shopify store on check prices. Give my agent a task like

go find me the best Babe Ruth baseball card on eBay. Go start negotiations, messages, low low ball bids, and what is like what do those um sales pages or those offer pages need to look like to get an agent to actually be enticed? Is it very it's it seems more like data driven like like look for the lowest price with this amount with this shipping time. So like it's like very filtered. >> Talk to us about like how how will agents may make buying decisions on eBay or in general? Yeah, in general from the industry perspective, uh let us say we are going like more evolving towards that and I think like more research will

take place and more you know like the protocol as you mentioned about uh there are like some of the buying parameters right like agent looking into that but I think going beyond that uh what we'll have as part of like the agent is that like uh agent definitely like comparing about uh you know like the price or how much like this particular item is and then after that comparing ing it like with like uh numerous number of like other selling agents as well. So there can be like the agent at the selling side as well as agent at the buyer side as well and these agents like communicating to each other agent to agent communication and uh along with like what is the price of particular item let's say you want to buy um like a headphone right so what is the price for

that item like comparing like let us say 10 uh sellers so if we try to do it manually uh it will take like lots of time for us >> we'd have analysis paralysis >> yeah and there is like cognitive overload right but these agents we do it for us. But along with that uh they would also go for like they can also compare within that like let us say some uh like you know merchant is offering you like uh 2 days delivery like free you know while other merchant is offering you two days delivery but with the charge while another is offering you let us say uh like one day delivery with charge. So you can set those like acceptable uh like the parameters or acceptable scenarios and these agents can take action behalf of them. But

going beyond like comparing just the uh item price or they can think about like uh like the reviews uh of that particular item or the description and also uh like overall like benefiting uh for you like when it can arrive and like shipping charge. Let's talk more about reviews and testimonials because one of the things about AI spinning up their own interfaces, double-checking prices across 20 different venues is the collapse of customer loyalty. And so when there's no customer loyalty, what do reviews look like? What do testimonials look like? How does this how does this help?

>> Yeah. So now uh like um one thing like probably with the reviews, what can like come again, you know, like uh like sometimes there can be like fake reviews for a particular like merchant, right? or like fabricated. But these uh like maybe like it can be one of the you know like a model which can where the agents can again like spot like which review is like fake or you know like repetitive or like this looks like uh just like a fabricated review uh like that. So I think mainly these testimonials like which we can we cannot spend like let us say like 20 pages right like we as like a buyer like going on site we cannot spend like that but maybe these models will evolve in the industry which can

again like you know spot some particular like fake reviews or if somebody has been doing like the same review like multiple people >> spamming >> spamming reviews sibling >> yeah like yeah I mean I could imagine a a competitive uh e-com commerce brand with each other, sending malicious agents to each other's websites, try doing buys, chargebacks, buy, chargeback, uh all different types of like manipulative behavior, trying to like do sites. Talk talk to them about kind of like what the payment experience needs to look like, right? Like uh how much integration of stable coins is actually happening? Like how do you think the commerce side of the e-commerce business is going to play

out? Like is it going to be agents with credit cards? Is it going to be stable coins? Is it going to be something in the middle? Like what is required to prevent some of that like civil activity? And what is required to make it seamless? >> Yeah. So as you said like you know like the agentic probably like uh you know attacks like like these different like platforms. So I think for that uh there should be definitely like agentic security or uh like at least some of the uh like the like the spotting uh like the layer which can detect that this this particular uh like the agent which is coming to the platform is not uh like the real uh like intent buyer right.

this has been like uh like some spam or like the attack of like like different agents coming to uh you know like the site and just like spamming us or you know like um like using our resources right like keeping our resources busy. So I I think um like these things will evolve but there is definitely uh a concern for the agentic security right or the guard rails like um I think the industry will along with the how much industry is um like eager to use the agentic AI right but uh like this has been like one more factor like in adoption to be like more uh how to approach like let us say if I open the agents but then like to think about these scenarios like uh let us say 500

agents being spam fan off right to keep the resources busy and that that is where uh like all this agency security will come in place. So that that uh that the more like research is going and uh as you talked about the uh second part of you know like how the payments would like uh taken like effect. Yeah, I believe like what can happen like there is that like maybe like more research is also going on like I met like few of the representatives from Berkeley like research like they they were also doing similar like you know uh like the payments um like the agentic like research but what can probably evolve is because we don't want to hand over our real credit card details to uh the agent right uh yeah so for that probably a

digital twin like uh some of the research paper which I read about uh like digital twin of the um like the credit cards or even you know like another uh when I was talking with some of the Berkeley researchers like um like probably the credit card information uh which is uh there like it is staying at our like platform like not going actually to the uh like agents like staying securely then agents taking like identity of that and talking to uh you know like the other side of transaction completion >> you Uh so like not handing over like the real detail but uh >> the last thing I I I wanted to ask you about was you know we because this this

is all mostly from the seller side right and and talking about the payment experience the checkout experience and how that is going to look but from the buyer side we're going to have agents that are executing most of these things on our behalf. How can we tell our agents prompt our agents and engineer our agents so that they are making the best buying decisions? >> Yeah. So uh definitely like there can be like many options you know for example from the buying agent uh that is let us say uh buy me uh like headphones below like $300 right like you just prompt a command now uh like this can go to uh like comparing like different products like the descriptions again against like those products you know so that can be one of the feature then automatic

checkout can be one of the feature but one uh interesting thing comes to my mind and where like you it can um like really like excel and bring a benefit of agentic to uh like we as a buyer right so uh sometimes I think based on uh like so far like experience expertise that let us say when you buy a particular item right then after that uh like we are not comparing within that uh like a window of 30-day returns right if there is any other like a price drop or not right like let us say we are not tracking that but there can be agent which can be uh put in like job for job for that right let us say you bought a like um like $300 uh like headphones just before let us say black Friday deal

right so you had a need you bought it right now you have that return period of like let us say 30 days right uh return period like free returns but in between that time let us say uh like it is also like black Friday deal right and now within that 30 uh like uh days window uh like some another like the similar like headphone like price is dropped right so that can like be of uh your benefit. So some agent for that particular like use case and we as like when we buy we cannot keep a track of this right like we cannot be continuously like looking at that and even when we buy we go sometimes like to check like price drop but how about like if some agent is already looking for you for like

continuous you know like 24 by7 like you know like and agent has a goal of like saving money for you uh that is you know like uh within that 30 uh like days window like the price is dropped and it matches your description what you already bought. >> Uh then basically it initiates your return, right? And it notifies you I got a good deal for you. >> Uh the headphone which you got two weeks back. >> Yeah. >> Yeah. You can like just drop off at the UPS like here is the your nearest UPS. >> Yeah. >> And you know, >> well that's when you need a robot.

>> Just just get it to add to the robot. So Ramman, thank you for joining us. Um and and for sharing some of your work and thoughts. Um really appreciate it, man. And it's interesting to hear from >> from a a online retailer like eBay just generally how how you guys are thinking about it, but really just how you're thinking about it personally and what you see coming. So thank you. >> Thank you. Appreciate it. Thank you. That was awesome. >> See you next time. Cheers. >> All right, guys. Our second to last guest, we've got Cardsik bot in the building talking about X42 AI payments. Cardick is a founding engineer at SE Labs. Love it.

>> Welcome to the show. Thank you. Jay in um Denver last week. >> I believe that was at the cigar lounge. >> Cigar lounge. I had a couple of vodka sodas. I think Jay was a sober sober sober. >> Okay. >> He's a pure soul. >> He's a he's a pure soul. So, man, how's your Neon been? >> Nearcon's been good. I'm excited to be on the what I've heard is the crypto TVPN. There we go. >> So like this is this is like >> this is an AI conference though. >> This is an AI conference. I agree. But I'm I'm excited. I'm excited. >> This is our first AI conference ever. >> Yes, it is our AI focused conference. This is it.

>> So KY tell us about your background, how you got involved with say what you do. >> Yeah. So uh you know I come from like a pretty AI background actually like practical sort of started at NVIDIA. I did the reverse of what people are doing. So I started in AI then I went to crypto. not went from crypto. Aai. >> So, not the pivot. Not the pivot that we all were wanting. You you you pivoted to the dark side. >> Yeah, I pivoted to the dark side. So, a lot of questions for my parents. >> What a guy, huh? Lot of questions. And I'm sure your parents have other questions about you out here, too. >> They do. They do. But I'm at an AI conference now, so like they're perfectly happy.

>> As long as you got some stock options from Nvidia, I'm sure the parents are happy. >> Yeah. Yeah, they're happy. But, you know, >> there we go. >> We'll make it happen. >> Awesome. Well, tell us a little bit about what you guys got going on at say Labs with regards to AI because we've had several conversations about agentic finance commerce. Obviously, a lot of this is coming to uh blockchain based systems. What are you focused on at SEC? >> Yeah, so like you know I kind of a hot take. I feel like XPro 2 in my opinion is the only practical crypto and AI use case and I think a lot of people here I think a lot of people at other conferences are hyping up onchain agents decentralization. I think for me what's exciting is that stable coins offer something that card networks do not

offer and that's like microtransactions, insulin settlement, global access and that's something that the AI community is excited about, not just like crypto people talking about AI, right? And so that that's what you know excites me. >> So why why is X42 not just Coinbase creating a standard for all us to all use in their traffic world? >> I mean X42 is the most basic possible protocol, right? This isn't like, oh, you have to use base or you have to use like some specific Coinbase on-ramp, but it's literally just reusing the status code. Like, if you look at like the early Bitcoin forums, like they were even talking about using 402. Um, it just wasn't possible to add a payment layer to the internet because there's no global payment network, right? It's just these card networks that are like sort

of that exist in some countries, not others. Um, they have these minimums. Some people integrate with them, some people don't. And so 402 never made sense until stable coin sort of got broad adoption. And I think with the Genius Act now people are waking up to this as like a incredible rail. So >> tell us a little bit more about 402 itself. Like what was the problem back then? You know now we're recreating the internet for agents. >> What is 402? Why does it make sense now? >> Yeah. So in general, the biggest problem was that I could not pay for a service without signing up for an API account and doing that whole KYC billing and fiat at the end of the month. So let me let me walk you through how agents pay right now, right? If I'm using OpenClaw, let's say it goes and discovers like

five different tools. I have to go to each one of those tools and set up an API account bill every single month, even though I'm just using it maybe once or twice. Like if I'm trying to find the weather in San Francisco, I have to go to this weather app. I have to sign up with my human account, set up billing per month just to get an API key, and then I pay for that in the end. And the reason for that is that there's no way to actually pay for what I use, right? Because an API call is just like a hundth of a cent or like it's very instantaneous >> versus like if I have to use like billing right now, it happens at the end of each month, right? Because that's how fiat normally works.

>> Um, and so what 402 is saying is like look, stable coins enable this whole new design space. You get like instant settlement, you get microtransactions. So if I am setting off my open claw agent to go find the weather, do a bunch of stuff for me. It hits an API endpoint, it can pay for it immediately and get back the response. It doesn't need to go set up like an API key and do all that nonsense that we as humans do. So that's unlock one of the things that people are attributing the SAS apocalypse to is like, okay, AI is going to cause a lot of like employment disruption. a lot of people are going to be unemployed and then we're going to reduce the amount of seats and because they charge by seats this is going to get uh you know kind of brought down

their margins are going to compress the other payment model that they have is that they pay by month right their subscription models do you think 402 stable coins micro payments is going to compress a lot of margins that are currently there in subscriptionbased models >> yeah you know to be very honest I don't think so like I I think that 402 is great for a very specific wedge and that's one it's not for humans like humans will not use microp payments No matter what, I'm not going to look at a New York Times article and pay like a cent just to access it. I'll bundle. I'll do it by the subscription. It's easier for humans, right? I think it only makes sense for these high frequency, lowv value transactions, which are APIs. And then even then, bundling will still occur. Like it just makes more economic sense if I use like

let's say I use this weather API a bunch of times. I'm going to set up like a wholesale like much better pricing. Um but for discovery and on the-fly usage, it totally makes sense. So I think X42 will be the wedge for stable coins but it will not replace like the sort of daily behavior which is like bundling and it which exists for a reason >> and because micro payments occur at such a frequent basis you know 402 like Andy said is primarily a Coinbase thing. How much are you worried or thinking about you know microp payments occurring on say versus micro payments occurring on base because they're using 402.

>> Yeah. So my my answer to that is that when micro payments occur at scale, it will happen on the fastest and cheapest chain and that's say base is still very slow. It still has congestion issues. What happens is these rails will move towards the fastest and cheapest option and that's say right now fastest throughput cheapest chain >> the most practical um option for the agents. >> Yeah. And so that's the hope, right? And so obviously X42 it's it's Coinbase coded but I don't think so. It's like pretty generic and I think that over time they'll go to the best possible chain and that's say >> yeah. Um when you're thinking about actually the the like the transition from where we are now to getting agentic flows happening on chain like what does that actually look like? Like is there a

model improvements that need to happen? Is there like user improvements? Does somebody just need to create a UX that is like natural language easily connected to like uh any sort of wallet provider that is really easy? Is it like a neo bank fintech platform with like a robo advisor like what is that transition to like holy [ __ ] 10 10% 25% of onchain transactions are being done by agents like what's that process? >> Yeah. Yeah. So I think three major things I think one open claw and claw agents you know they've changed their name so many times but that has clearly shown that people want pay peruse and so now there's a bunch of these different solutions off offering like how you can pay for these specific tools for your specific open cloud agent. right now there is a little bit more demand. So

that's there. I think the second thing is the wallets need to improve dramatically. And I think you're seeing this like Privy and Coinbase are introducing their like agentic wallets. Like you need guardrails, you need like very simple like um you know human in the loop potentially. Um and you need to completely abstract like private key management and all that stuff. >> Um and I think the third frankly is that stable coins do not have distribution right now, right? Like I still have to on-ramp. If I'm someone who's not crypto native, I have to go to an exchange. I had to onramp. I had to do all this nonsense just to get $10 for me to use in this like pay-per-use fashion. And so I think that's where things like an XRO2 credit card can solve like a preunding hurdle. That is I think the biggest

barrier. So you know again like the demand side is there with openclaw. There is like stuff happening on the wallet end but the preunding hurdle which is like where I actually onramp to use stables that is still an open question. I think we're spending a lot of time thinking about that. >> Yeah. Interesting. Yeah. And what's like what's your estimate like is this going to happen like this year? Do you see more kind of aentic stuff happening or is like next year? >> It's funny like I think Hib came on here was it yesterday and he was like kind of bearish on it. Yeah, he was >> um you know I I think it'll happen this year and I think for like two major reasons like I think the frankly we have to admit the broadest players in fintech are entering right so Stripe now

announced X42 integration and then Cloudflare which is like a huge web two player is about to launch what they call the net dollar stable coin and a payer crawl right where you can agent can crawl a specific website and get a sort of you know access on demand and so I think the tailwinds there will definitely help the broader industry because you have the broad biggest fintech right now in in the business, you know, endorsing X42 and then you have one of the biggest players in web 2 endorsing stable coins. >> And so I I think it's earlier than what says. And I think that it doesn't have to be complicated like I'm going to have my agent run my life. It could literally just be like when an agent is scraping websites, just mini microtransactions to access that website.

>> Yeah. >> Right. So I think that's where it starts and I think that's where the floodgates get open. >> Yeah. >> Yeah man. I uh I think it's time for developers, builders to all start optimizing for agents. Like how would if I'm spinning up an agent? Oh uh excuse me. If I'm spinning up a website, how would I turn on like pay for crawl? >> Yeah. So that's a great question. I think we've been debating this internally. How can you make this service agent ready? >> Um and I think that happens in a hacky way right now. So people are doing things with what's called like proxy servers. Um so I could just resell my API keys. So think about it this way, like I can buy APIs in bulk. Like I will pay for OpenAI in bulk in fiat $10 every month and then I can open up a service

that just says pay for, you know, OpenAI on the fly per call in stable coins, right? And so it'll just proxy it to my same account, but it makes it agent ready because it's per call, right? You can pay for the exact usage that you want and you don't need to set up an API key. So that per call like per usage is what I think makes something agent ready and it's only really possible if you use stable coins for payments there >> and so there so agents for sure just going to use stable coins like for sure >> you know I I I think for shopping it's a completely different scenario so I think people are like oh I'm you know they're going to use stable coins to help me buy something makes no sense like stripe has something called ACP which is what openi

uses when they buy when you ask chatbt to buy me a jacket right if I'm going to buy like a rollup jacket on chatbt it's using what's called ACP. Card networks make complete sense for that, right? Because it's high value, low frequency transactions. Chargeback and dispute resolution is absolutely necessary. If I don't get my jacket, I might have to send like a lawsuit to you guys, right? Like, and chargeback matters there, right? And so, >> you'll get your jacket. I'll get my jacket. I'll get my jacket up jacket. You got one, >> but Yeah. Yeah. Chargeback matters, >> guys.

>> Yeah. You'll have to send that to my agent. I'll have to send it to your agent and then we'll have a little agent to agent communication. Yeah. Um but I think card networks are going to stick around and agents will use that. But again these these small cases like for APIs and like for paper crawl that's where stripe is excited. That's where Cloudflare is excited and ultimately that's where you know we are spending most of our time. >> Awesome. Car thanks so much for coming on man. >> Absolutely. Cheers. >> Yeah absolutely pleasure man. >> Enjoy the rest of your uh experience here at Thanks for coming on man. >> Absolutely man brother. See you. >> Cheers buddy.

All righty. Um, well, you're going to have to send that lawsuit to my agent because uh I'm not reading that damn thing. >> Let's get some Let's get some rollup jackets, >> dude. I mean, the leather jackets they had out here at Nearcon. >> The leather jackets. >> Yeah, leather jackets. >> No, no, they weren't near branded, but we had Ilia yesterday. We had George yesterday, Stephanie and Chris were all Everyone was walking around in a near uh in a in a leather jacket. >> Oh, yeah. I mean, it was like the Jensen Huong swag. >> The Jensen Huang swag. They also had like sick Nearcon jackets. >> They had normal jackets.

>> They had Nearcon hoodies that were very fire. >> I think we're going to roll out some uh I think we're going to roll out some merch in in the nest. >> Nice looking >> uh some merch in the nest is coming up. >> What's the Nest? >> The Nest is going to be revealed in like two weeks. My uh my guy Alex behind the scenes had a had a bit of a ball breaking call this morning here with with him to uh he he he let me know that the timeline was just extremely extremely delayed. Um, you know, so what we're going to do is going to make it even better for you guys. We're going to absolutely make sure that we're every single nook and cranny is dialed in to the exact tea. So we don't have any aneurysms internally.

>> No, we don't want no aneurysms internally. >> Next guest, we got Dio from Han, the last guest of the day. Don't see him quite yet, but he's coming. Guys, we're going to take a quick commercial break uh before we have an aneurysm in an hour or six here. See you in a second. Heat.

Hey. Hey. Hey. Heat.

Hey. Hey. Hey. Heat. Hey. Hey. Hey. Heat.

Hey. Hey. Hey. Hey. Hey. Hey. Heat.

Heat. Heat. Hey. Hey. Hey.

Hey. Hey. Hey. Heat.

Hey. Hey. Hey. Heat. Hey. Hey.

Heat. Hey. Hey. Hey. Hey. Hey. Hey.

Hey. Hey. Hey. Hey.

Hey. Hey. Hey. Hey.

Heat. Hey. Hey. Heat.

Hey. Hey. Hey. Heat.

We are back live at Nearcon Day two, the last segment of the day. We've got Eric Vorhees, uh, notably the CEO of of Shape Shape Shift amongst others. First place I made an onchain transaction. >> Was it? >> Yeah. >> Eric, so story time for you is Rob and I met at the University of Florida in the uh blockchain club there, Bitcoin club in 2017. Part of our, you know, inauguration and getting started there was bringing on speakers into Zoom and teaching us things about crypto. And with the help of a friend named Antonio and Felipe, we actually called you in.

You zoomed in in 2017 and did a live Shape Shift demo for us. Talked for about 30 minutes about uh censorship, resistance, kind of libertarianism, and then you were gone. >> That checks out. >> And here we are today. >> That checks out. That was the best crypto year of all 2017. Yeah. >> Yeah, man. Well, welcome to Nearcon. >> A lot has happened since then. >> Yeah. A lot of a lot has happened. Uh industry is a big thing now. >> Yeah. Eric, one of the things that I think is super super pressing in the industry now and it's the cause of a lot of the sentiment disruption is that we've built a lot of this infra open source infrastructure at scale. Um, but we haven't quite figured out value capture from it and there's a lot of these kind of the anti the antithesis of

what we've built here being institutional funds, trady, if you will, coming on building on the rails and capturing a lot of the value from it. people have a really really bad feeling about this kind of environment also with the like the era of pragmatism coming about where it's much more about users, customers, products, revenue and business and less about the the core ethos and less about why we're here. How do we as an industry go into this next era of pragmatism of institutional adoption while maintaining the core ethos of what we've built? >> Great question. Uh I think the the core ethos is um is protected by the protocols, right? Uh the ethos is built into Bitcoin. Um and that that doesn't

change regardless of who's using it. So um it's the protector is really the code that people have written and that that's been the whole point. Uh I I don't take the same view that it's um so problematic that all these institutions get involved. I think that was always hopefully inevitable. The point is that uh the crypto takes over the world and to do that it has to merge with the world. And so um this should be a sign of of success. And what would not be a sign of success is uh if somehow we all ended up just using like custodial companies and there was no way to use anything else.

>> Uh if there was no alternative, if there was no DeFi, if there was no self-hosted wallets, like if if that got lost somehow, then we'd be in big trouble and we wouldn't have created anything useful. >> Yeah. But that's not our situation. >> What do you think is the main limiting factor for mass adoption and that could come with institutions being the distribution channel to citizens and users ultimately? Is is that primarily the the vector that we're exploring or is there some other limiting factor that could bring the masses?

>> Uh I think the masses just have to happen over time. Like what what we've done with crypto is fundamentally change the way that money can work. Uh but people's behaviors take a lot of time. Like people have inertia. They get used to certain behaviors and uh that can't be flipped off overnight. Um there is so much friction and inertia in the the way businesses transact with each other and people transact with each other like that just really takes a long time. Um but if you look at sort of crypto adoption by age group uh it grows as the group gets younger. So it is multi-deade phenomenon but you know like my grandchildren probably think crypto is

pretty pretty normal. Even even my my children will. >> Yeah. >> When you were kind of early pioneering in Bitcoin and just generally in in DeFi 2015 2013 era all the way till 2017. >> Well, there there was no DeFi in 2013, >> right? >> Pre Ethereum, pre smart contracts, >> right? >> But you you had part of the vision, right? You saw some of this or at least you saw this >> Bitcoin Bitcoin, I guess, was the first DeFi project, right? It was like it did DeFi for money. >> Yeah. >> Yeah. What was the inspiration for Shapeshift which ultimately kind of spurted off decentralized exchanges DeFi from there?

>> Uh so I started ShapeShift um back in 2014 after Mount Gaus collapsed which was like the you know 95% of all the exchange market back then big centralized custodian and it almost destroyed the whole industry. So I was like well how do you build a exchange that doesn't hold any customer money and doesn't require them to reveal who they are? you just protect people by the design of the of the exchange. So yeah, ShapeShift was the first non-custodial exchange. Um, and yeah, we built it, you know, two years before Ethereum came out. Then Ethereum came out and the whole explosion of different assets really uh caught fire. So it was good timing.

>> Yeah. Did has this played out the way that you thought? has the has the last 10 years so 2016 to 2026 did you envision kind of obviously you could have predicted it perfectly but was this like part for the course was it way off was it >> um at least for me I did not predict or understand what a smart contract was before that happened so it wasn't like I was like yeah someday there will be these smart contracts and everyone will be using DeFi primitives like that was a little alien uh and even when Ethereum came out it took me a while to understand the value of what that was.

But uh crypto is you know like slowly taking over the world and and um Bitcoin has become you know top 10 asset on planet earth. That is absolutely what I was hoping to see. >> Yeah. So in some ways it feels like it's happened faster and other ways slower than I would have wanted but it's happening. >> Yeah. >> Do you see similarities in the way AI is playing out now and the way crypto played out in the early days? >> No, not I think they're totally different. a AI had this like long academic history going back decades.

>> Uh but it wasn't ever quite potent enough to break through in the mainstream and then the chat GPT moment happened and suddenly it took over everything like it's just it's been a few years and AI is everywhere and it it became completely mainstream within two years. Crypto still is not completely mainstream. Yeah. And uh and crypto also didn't cryptocurrencies at least did not have you know like 80 years of of academic history. It was a much more recent phenomenon uh and has been slower growing during that time. Whereas AI was like you know being worked on forever and ever and ever. I mean, imagine being like someone in AI and living through 50 years of disappointments and it never

achieving the things that you wanted. >> Yeah, man. >> Um, but now it it finally certainly has. >> Yeah. Let's talk about about the work that you're doing at Venice, Eric. Uh, we're at we're at Neircon and kind of this convergence point between these these two industries. Uh, Venice is a part of a of a new kind of asset in crypto. there's this this idea of kind of doing GPUbacked tokens in crypto. Um you've got one called DM, you've got USDA, Venice is is you know effectively in that uh category as well doing novel kind of uh GPU experiments with tokens.

How's that been? What like what's the short and sweet on Venice? Like if you had to explain it to somebody who is it's their first time. Uh and why is this something that you're spending your time on? Uh yeah, so that's it's my it's my full-time thing now. Uh has been for two years and and basically I wanted to build a private uncensored version of chat GBT. Yeah. >> So I I loved chat GBT, but I was like this is super dystopian unless someone builds something with actual like crypto ethos in it. >> Um so it is first and foremost a consumer app for AI. Um and of course me being a crypto guy, I'm going to bring crypto elements into it. So you you mentioned DM, that's one of the the tokens of of the Venice ecosystem.

>> Yeah. DM basically lets any holder of it get free compute uh a dollar per day of free compute. Um and so we built this you know kind of as an experiment to see like uh will agents um like the ability to get uh essentially like inference at zero marginal cost instead of paying per call they get a free free inference up to a certain amount every day. Um, so yeah, we're just playing around with these kind of um primitives within within AI. And I think uh the AI and crypto world are moving increasingly close together. I think we may all realize at some point in the future that like cryptocurrency was actually just being built for the AI robots >> and the human humans have always been

awkward with it, but like agents and and um machines will understand it easily and have no problem with the UX issues. So, it's interesting to see these worlds merging and uh it's been fun to build in both of them. >> What was a turning point from when you were working with crypto and then ultimately, you know, now with Venice? Was there a point in the middle there that was it the chatbt moment that kind of lit a light bulb and said, I'm going to really focus on this new >> venture? Yeah. So, we we decentralized shape shift in 2021 and I took a couple years off and um I didn't know what I would do next. Uh then the AI thing happened and um I just to me it was one of those clear cases where like someone need someone who cares about the

principles that I care about needs to build >> an alternative to what these will become. >> Yeah. >> Um no one in the AI world cares much about free speech. No one in the AI world cares much about privacy. uh I'm generalizing but like those are relatively foreign concepts to a largely previously academic community that u was very much in favor of always controlling people and just making sure the right person was controlling them. That's a totally different mentality than what we have in crypto. So um I felt like I had to do something and built Venice for that.

>> Yeah. So in the ecosystem, you know, I mentioned DM, you're talking about the private uh kind of uncensorable interface like if this works like are we are we building an entire economic structure with the principles of kind of you know that you share with the early stages of crypto you know censorship resistance decentralization immutability etc. building that in the AI lens with a fully self- sustaining kind of economic model inside like what is the ecosystem vision that you have for Venice. Um you know like like what can we what can we kind of like like you know tie this to or map this to in our minds?

>> I the the simple way to think of it is just like um Venice is a a consumer app trying to do a very specific thing which is like provide inference that is private and that is not censored. Yeah. >> And uh everything that we do is built to support that that mission. >> So we're not trying to do everything or create we're not really trying to create an ecosystem. We're just trying to build a a really important product and infuse within it the the ethos that I find so important. when when you see some of the foundational labs um censoring at their model level um does that give you more valition around what you're building to essentially provide that uncensorability to whether it's claude or bt they can

people that were being centered can now use venice as a vector to explore those models >> yeah so like you know those are private companies and they can set whatever rules they want on their models right that's that's fine fine. Um what's not fine is if that's the only alternative and what's not fine especially is if governments come in and basically um demand that that kind of censorship must happen. >> And the risk with those companies like OpenAI and Enthropic is that they will and have lobbyed governments for regulation uh to enshrine the way that they do things as as the correct way.

Um, so I find that problematic and you know, Venice is a a consumer app as an alternative, but the the amazing thing that's happening is that like Chinese labs are actually the greatest bastion of uh >> open source >> open source uh models here. It's so it's super ironic uh very weird and strange >> but um yeah as as long as open source models are uh competitive and you know widespread which they're getting increasingly powerful then we're good you know and it's uh it's all fine.

>> What do what do you make of the geopolitical evolution there being the Chinese open source models and the closed source private AI companies in the US? >> I don't I don't know. I don't know. Um, you know, and it's always it's always hard to it's always hard to figure out like to what degree things are centrally coordinated like is the Chinese government like dictating everything that's going on. It's easy to think that way because people tend to think of nation states as like personified entities that act with pure focused attention, but um it's all of just a in my opinion like a messy phenomenon that no one is really sure how it's going to unfold and there's lots of incentives going on all at the same time.

>> Yeah. Eric, uh last thing that I have, you had a very famous uh interview with SPF uh before he uh blew up. Um you you know, you were right, notably. Um, and recently SPF's gotten his phone back and has been on Twitter. Great. >> Um, >> today Trump said that he's not going to pardon. >> Correct. Trump is Trump is confirmed that even though that SPF is quite literally on his knees on Twitter for Trump. He's not Trump will still not pardon him. Just wanted to give you a chance to comment on that.

>> I mean, it does not seem that he is uh worthy of a pardon. No. Uh, no. Uh >> he's trying to say that he that he was always liquid, that he was always solvent, and that due to external forces, his insolveny was forced upon him rather than and that he's been this great investor in AI. >> I just don't think anyone believes that story. He has been a great investor in AI, that's for sure. He made a $500 million investment in anthropic. Of course, did it with other people's money. Um, amazingly that investment turned out to be the thing that is making most of the anthropic creditors whole or mostly whole or uh >> FTX um yeah, FDX creditors whole but that was that was luck. Um and uh yeah, I

don't think SPF deserves anything but his sentence and he doesn't really deserve the attention anymore. I I'm so sick of all the scammers in crypto that have just like caused such damage to the industry. There's been so many of them. Yeah. >> Um, >> yeah. >> And it's been a bummer because, you know, the rest of us have to deal with the consequences of that. >> Yeah. Especially, I mean, in AI, we've seen it as of late more and more where there's just this complete separation and just desire to not be affiliated with crypto at all. >> Yeah. And it's it's understandable.

Like, from the outside perspective, you can be forgiven if you think crypto is a scam because if you just if you read articles once in a while and you have a couple friends that got ripped off, like, what else are you supposed to think? >> Yeah. Um, you know, and I can't sit down with everyone and have a 30 minute conversation about the philosophical principles that undermine, you know, like sovereignty in the realm of money, you know, so it's a it's an uphill battle. >> Yeah. Eric, last thing I have for you. We're we're at New York. You had a chat with Ilia on stage and I'm I'm curious any takeaways that you had from that chat and uh how you see Venice and Near working together in the future.

>> Um, yeah, I just I was glad to be here and to have a chance to talk about privacy in in AI. I mean Ilia is one of the very few people that is that has deep credibility on in both worlds, right? Like almost no one can actually claim that. I certainly have no credibility in AI. Like I'm a I'm a tourist. I'm learning about it in the last couple years, but um he he has both. And so this is uh it's very cool to see the this event coming together as a fusion of these two worlds. And I hope it gets um bigger and more I I want more of that integration to happen. So he's he's one of the few like figureheads that could actually do that I think.

>> Yeah. >> Eric, as you said, man, we you've inspired us since 20167. First time on on on the show. We'd love to make it a recurring thing. Thanks for coming on, Eric. Hope you have a great rest of your conference. >> Cheers. >> Cheers, man. >> All right, guys. With that, that wraps up our Near Con coverage, guys. It's been 2 days, 25 guests, 14 hours. Amazing guests, amazing conversations. Ended it off with a bang. Last but not least, Eric Vorhees. Guys, thank you for rocking with us. We have a huge upgrade. We're coming to New York City in a big big way. Uh I have a lot of claims to make, but I'm going to not say them today. But guys, we'll see you in New York. All right.

>> Thanks, guys.