This premiere episode introduces the DeAI Journal and explains why decentralized AI matters now.
Co-hosts January Jones and Brad Keoun frame the show around the intersection of blockchain, AI, media, and open source infrastructure. They discuss why AI can make blockchain easier to use, why blockchain can help solve AI’s verification and privacy problems, and how DeAI is being defined in practice right now.
They cover recent deainews.com stories with Clara Tsao, founding officer at Filecoin Foundation and talk about Anthropic Watermarking, the used GPU market and hackers.
Wrapping up the show, we have a feature interview with Shaw Walters, founder of Eliza Labs, who recounts his personal tale of "overwhelm" following a token launch that rocketed to a market value of $2.7 billion in a couple of months, only to crash just as quickly. He says his AI agent project aims to make it easier for people (including busy mothers!) to manage their lives.
Chapters
00:00 — Introduction to the Podcast and Guests
01:53 — Origins of the Podcast and Personal Backgrounds
04:06 — The Intersection of Crypto, Media, and Decentralization
07:01 — The Evolution of Blockchain and Crypto Technologies
10:56 — Defining Decentralized AI and Its Layers
15:50 — Insights from Illia Polosukhin on Open Source AI
20:04 — Clara Tsao on Content Integrity and Watermarking
27:01 — GPU Market and Decentralized Compute Networks
34:07 — Filecoin and Decentralized Storage for AI
38:05 — Shaw Walters on AI Agent Frameworks and Projects
45:02 — Lessons from Building Eliza Labs and Community Challenges
53:53 — The Future of Open Source AI and Its Societal Impact
1:01:54 — The Significance of Open Source in AI Development
1:07:01 — Closing Remarks and Next Episode Preview
A Distro Media Production
Produced, edited and designed by January Jones
Music licensed through Soundstripe. Code: 7IIILTUES54J7TXP
Transcript (by Riverside)
January Jones: Welcome to our new show. This is the DEAI Journal and this is the podcast of DeAInews dot com which you founded. Welcome to your show.
Brad: Thank you. Welcome to your show.
January Jones: You thank you. We have a podcast now. Yes. I know, finally, right? Everybody and their brother has one, as we say. but it's important of trends. Yes, right.
Brad: We like to be on the trailing edge. That allows us to learn from the mistakes of others and we've pr we like to think we've perfected it.
January Jones: Well, we will. Let's see, if we can get through this episode then we'll know, right? But mostly it's not
Brad: Don't put a fork in it yet.
January Jones: For today, no, mostly it's not us. we are connecting with amazing people in this space. So maybe we should kind of back up a minute and start with how we met and how we got to doing this podcast here today. So we first met through the crypto space. We met at an event in Nashville for journalists and publicists who work in crypto. I was a former journalist in public radio. Then I got into the crypto space and events and then joined a podcast, a weekly podcast, news commentary. and that was my full immersion in 2024. So I'm still a newbie. Brad, crusty old journalist, been around the streets, been a crime reporter.
Brad: I been on the bus for a long time.
January Jones: He was. And you did Bloomberg, you have a lot of financial background, and then you were at Coindesk, you did their protocol. podcast as well. And so we just met at that event where where everyone was networking and had the same interest in decentralization of media. So that's really where we came together. Because when I got into the decentralized space coming from my journalism was also in community and public radio where we were teaching people media tools all the time. And the idea was Media ownership should be everybody. And anybody who wants to be the media, come on in, we'll teach you, we'll put you on the radio, you know? And so that kind of open source, community-based media ownership has always been in my DNA with all the projects that I've done. And so that part of crypto and blockchain really resonated with me when I joined this space because it was a lot to sort out, especially 24, 25, when everything was just exploding. so that nugget that I took was that I saw decentralized technology as a way for journalists to preserve their archives, have forever archives possibly, you know, this other backup for identity verification. and then some kind of interesting innovative ways that maybe you could fund media, right? And so That's where we started to vibe because I had the similar perspective that there's something here to decentralize infrastructure for media and for journalism. And so that's how we started talking. And you're like, hey, I'm building this product. And we would just have very kind of rando discussions for an hour or so here and there.
Brad: My vision was crystal clear from the beginning.
January Jones: But it was clear on DEAI News where you do publish stories you have been for nine months now, right?
Brad: Yes. Well, I initially got interested in crypto, because basically it was clear to me that it was the new technology. I had covered banking before, as you mentioned, and I actually did a lot of coverage of city groups efforts to build its own financial technology. And that's and I did a lot of stories about just how terrible the big banks were at innovating. And there's fintech, it's a completely different category of of startups that build stuff for banks. And then when I learned about blockchain and I saw how fast that development was going and I started to dig in and learn about it. And as I could best I could tell, the people working on blockchain were just miles ahead of the people who were just doing these tiny little bolt-on projects
January Jones: Mm-hmm.
Brad: for the banks. And I just saw that crypto was gonna outrun the banks. And so that kind of got me hooked on this story as a journalist of blockchain at Coindesk. If I was initially the managing editor of markets, we wrote a lot about Bitcoin. And then I was the managing editor of tech and protocols. And I was just going through all of these projects. And a lot of them are alike, but they're all creatively different and or sometimes stupidly different. And and I just got super into the idea that with this new technology you can build These decentralized protocols with a completely different system of governance and the technology. But the proof is in the pudding. It's faster, it's better. It previously has not been easy to use. And that's where AI comes in, because AI actually makes it super easy to use in the same way that people are now using AI to book their travel. People are using AI to Do financial transactions. People are using AI to build entire software applications. And as we've been using this ourselves in distro and building with this combination of using crypto payments in combination with AI specifically, it's clear to me that this is the way that the world is going. I think this is the hot story as a journalist. and There's been a lot I I mean, there's lots of opinions all around about AI and blockchain. Crypto is scammy, AI is ripping everybody off, and it's gonna kill all our jobs. Maybe true, both of those. And right, exactly. And I mean, it's possible that
January Jones: Those things can be true at the same time.
Brad: these that blockchain could actually solve a lot of those problems for AI. Just as AI could solve a lot of those problems for blockchain. I just think, yeah.
January Jones: Yeah, I see it as like the balance is that AI can solve a lot of the usability, user interface issues for blockchain, and blockchain can solve a lot of the security and verification issues for AI.
Brad: And you know, some of the anti-censorship properties they talk about with blockchain could theoretically be applied to AI. We're seeing all sorts of reports about various guardrails or censorship, Also, the privacy and confidential data,
January Jones: Yeah, privacy.
Brad: Yeah, exactly.
January Jones: Because to use these AI models and as the big ones have launched just in the past six months, it's been so quick adoption, right? we agree to give up all privacy so it could we can get an answer in four seconds. Basically. But
Brad: Convenience versus privacy.
January Jones: that's always the trade-off. I just want to back up just a second and maybe explain our ecosystem here before we get into our amazing interviews today, because we're gonna talk to some of the the market makers, the tech makers in this space are on this show, really pivotal people at this convergence of blockchain and AI. but back to us. So you built distro. media, which is a publishing platform, and it's AI centric in that it it can assist with publishing tasks, kind of the the tasks that nobody wants to do with the CMS systems and it integrates some AI tools, but it's still encouraging human journalism. And so that's where we publish deainews.com. So the core of the show is all about deainews.com But we are basically what we call dog fooding by publishing on Distro, a custom platform that will be available for other people soon. But for right now, we're the primary users and we're kind of showing how it can be applied and used. And we're also building in the open. We have another podcast called Distrocast, where it gets into the weeds about agentic payments and all the vibe. coding slash real coding that happens on a daily basis with our bots and with bot Brad and Bubble Arm Brad. So that's a teaser.
Brad: Look that one up.
January Jones: look that one up. but anyway, just to make that clear. So we publish on distro dot media D E A I News dot com. so that's where
Brad: I just wanna say I I really appreciate you reminding me that we have zero customers for distro. But we do have one customer
January Jones: It's whitelist. It's not
Brad: and that's me and I'm using distro all the time and I do love it. Just tossing that in there.
January Jones: Yeah, and sometimes you buy articles from yourself just so you can prove it works.
Brad: Hey, I'm getting rich in that one wallet, not the other wallet.
January Jones: Yeah, but I just kind of wanted to lay that foundation because you mentioned distro a couple of times and people may not may not know what we mean by that. So of gave your explanation of why this is an interesting space. And so in this show, we want to talk about decentralized artificial intelligence in a more plain language, but it's gonna get technical. So my job is to try to bring it to some something understandable for normies. because the abnormies
Brad: Boring.
January Jones: the abnormies they go deep quick and it's like Pull them out. we've been talking about what is decentralized AI. Is it about infrastructure? Is it about use? You know, there's different layers to it, and there's a lot to cover, and that's why this is such a fascinating subject because everybody knows about. AI or big AI, right? But I don't think people have thought about it in the way that we have been thinking about it from coming through blockchain and like what does it mean to decentralize anything, right? And in this case, as we talk to different people, they all have kind of different definitions of it. that's why this show is exciting because we're having the first conversations really about this space. And you started it all with founding DEI News on Telegram first.
Brad: I just decided to define that as broadly as possible, I think we include even stuff that is maybe not explicitly blockchain, but is more about openness and that same the the same benefits you get from decentralization of the openness, the permissionlessness as they call it, where anyone can plug in. So that's where you're talking about open source AI models. It's not controlled by one company that builds it and keeps the f formula a secret that it's you can inspect it if you want. You can download it yourself and you can see exactly what's in it and then you can use it. so that's also decentralized AI.
January Jones: It is. And so today, of course, that's all we talk about on this show. And so we are gonna go into an interview you did with Clara Tao from Filecoin We'll do the news roundup with her. So we publish stories every day on DEAInews.com. We pick a few and We'll hear her comments about anthropic watermarking, the used GPU market, and a hacker story. You can go see those full stories on our website, of course. And then we will talk to Shaw Walters, founder of Eliza Labs. Do you want a fanboy now on him or you want to wait?
Brad: I fanboy enough in the actual interview. I don't need to
January Jones: Okay.
Brad: provide a a pre cringe. A pre cringe cringe.
January Jones: Okay, great. but coming up right now, we're gonna take you into just a little montage of some of the people we met in San Francisco. And so we'll go into that segment and hear what they think DEAI is. most of the content for this episode we did record on site in San Francisco area. We made some connects at the Stanford Science of Blockchain Conference, which is perfect for us because AI is in there no matter what you're doing these days. That's where we met Shaw Walters from Eliza Labs Then a couple days later was the Bass Summit, at Stanford, and we got to do a bunch of on-the-spot interviews. So during those interviews, we asked, what is DEAI? Let's start with Shaw Walters, and then we'll hear from Michael Heinrich from Zero G Labs And then Illia Polosukhin from Near Protocol.
speaker-1: I mean the decentralization part that I imagine is like like what the the most advanced version of this is like an incredible world mind where a compute is allocated maximally efficiently to the cheapest places it can go and if you wanted to scale up to any amount of size, the compute is there, distributed across the world. And it can come from different providers who are all competing in marketplaces. But this this leads to the cheapest, most efficient market of global intelligence that could could happen. And and like a marketplace is a decentralized like a structure, right? And so when we're relying on one company who's very centralized for our intelligence, there are a lot of failure modes. But once we break that out and let the entire world compete on this, it goes down to becoming an efficient market of intelligence where like every human gets the best price basically that they can get on the best models. And I th I just see that as being like a much better and more robust way to build society than to like rely on a very small amount of people to like arbit be the arbiters of knowledge and and information and truth. And so do you think there's an importance of getting the term DEAI or decentralized artificial intelligence out into the world as something well known? Because we think it's you know still kind of fringe with the even in the AI community. I think the decentralized AI, open source AI, and and obviously things like private inference, some some amount of blockchain, all have like a lot in common and are all building toward this shared dream and vision. And so like the people in decentralized AI are maybe tend to use decentralization from the thought of like being more blockchain or crypto oriented, but they are actually Very, very aligned with all the open source AI people.
January Jones: Well, that's a nice vision, a cheap, efficient marketplace of global intelligence. Here's what Michael from Zero G Labs had to say.
speaker-1: Well, decentralized AI just means that it's not a single entity that controls, let's say, a network. Okay. So anybody can participate in the network. If you have a graphics card, you can participate in the network. Okay. Be a provider. You can provide inference, you can provide fine-tuning, all of these different types of things. Okay. So it doesn't preclude anybody from participating. And so why decentralized sounds like kind of mysterious, that's basically all it is. It's really a network that's owned by whoever participates in the network. So it's the network is not just owned by Zero Gravity Labs, it's owned by all the participants that are part of it. Got it. And so that's the benefit of decentralization and it's kind of a true version of open source, if you will. Yeah. Because if your model is open source, your network's open source, then why shouldn't a network also be owned by everybody that's participating?
January Jones: up the dictionary part of the show. Illia from Near Protocol gave his perspective via a nationalization lens on AI. We will hear more from Michael and Illia in our next show, so make sure you catch that one too.
Brad: Clara Tsao so glad to have you here.
Clara: Right. Thank you for having me.
Brad: Welcome. Yeah, welcome to to DEAI news podcast. and thank you so much for joining us. known most recently from your work with Filecoin, but I was just looking that you also at one point held the title of Chief Technology Officer for the Department of Homeland Security, is that true?
Clara: Yes, I was working on a lot of issues related to foreign influence of w operations, terrorist use of social media, homegrown extremism. So a lot of how bad actors they are using the web, but definitely a really, really critical issue when we're thinking about national security.
Brad: I can't imagine how many people want to ask you about that like all the time.
Clara: Mm. Well, I would say the story is always connected, right? the the
Brad: Yeah.
Clara: reason why I I chose decentralized AI and file coin and decentralized storage is ultimately the issue of content integrity on the web today, where a lot of people have a hard time discerning what is real, what is fake, how to make sure that the information they're gathering and seeing is authentic. And that's been a lifelong mission of my career is making sure that people can find knowledge and the best knowledge out there without necessarily feeling like they're being sold to or that you know there's a bottom line that someone else is trying to persuade you to do something else that is against your it in initial attempt to seek out that information.
Brad: Clara, let's talk about the first story on DEAI News. Anthropic comes out with the plan to watermark content. Including the code, which everybody, you know, has been using Claude Code and now that code is gonna have these watermarks. What do you think about this and w what's the DEAI angle?
Clara: Yeah, I I I think it's interesting because content integrity is actually one of the biggest proponents of why a lot of companies operating in the decentralized AI space are so interested in it. Because you want the ability to prove and authenticate a source actually came from that source. I do think that it is interesting anthropic, you know, is the first to kind of jump in and lead on this. I think initially a lot of the work is voluntary, but obviously the big AI companies have a lot of pressure to to jump in and do Ultimately I think more transparency
Brad: And and just ste step back, what is why is this happening? Why are they even doing this?
Clara: Yeah, I I think the big picture is and under the EU AI Act, there's just this bigger call for transparency on what models, what information are coming from which models. a lot of it from previous action the EU has taken, period, has been really around the the fear of AI and also AI safety. they're afraid that you know, if there are models or there's content that you can't actually track and trace where content is coming from, there might be secondary consequences that become very difficult to moderate downstream. So I think this is a greater call from EU regulators for just greater transparency among major AI companies to take the lead here. I think it's also really helpful for consumers who are on the other side, knowing where their models are coming from, where the content is coming from, making sure that when they're using anthropic, the anthropic isn't pulling from somewhere else, right? If you're using Claude Code or an an another solution, you you could probably prompt other models to be pulled in if you wanted to. And sometimes AI hallucinates as well. So I think all of this is is greater transparency. I think when consumers are using AI, it's very different from large companies automating it for the instance of self-driving cars. So when regulators are also thinking about AI safety and transportation and robotics, robots in your bedroom situations like that, I think it's more important to really know where which models are being used, how content is is being pulled from there. But I think ultimately it's transparency that in a way decentralized AI, the entire community has been leading for years by saying rather than the transparency being only there when large regulators ask companies to do so, how do we show it from the beginning, right? And so that's why there's so many amazing companies these days in the crypto and blockchain space that have been thinking about decentralized AI for a very long time and starting with that transparency first as opposed to having to go work backwards from there.
Brad: Well, okay, so there's been this big push by a lot of the big Silicon Valley firms regarding open source models, keeping those everything open. Anthropic, notably the loan holdout there, and also with kind of the most powerful closed models, right? And so here we have anthropic saying, okay, we'll do this. But is the industry going to do this or are consumers going to want these watermarks on their stuff? Or even i if you're an enterprise, like a big company, and you w you wanna use you wanna use AI to create some of your stuff, are you gonna want all this s this these watermarks which supposedly are gonna be imperceptible to humans but detectable by computers?
Clara: Yeah, I don't know if consumers will care honestly, because I think ultimately it's gonna be consumed by by computers, like you said. I do think that it is an interesting topic because I don't think the science or data is fully there yet on the safety side of open versus closed models. Ultimately it's just transparency and audibility because Anthropic does have really good safety measures compared to some of the other major models out there. some of the open ones, might might be even more unsafe, right? So I think we're still grasping in this time and age on what all of this means. and I think the regulators on the flip side are like, we just need greater transparency for everything that's closed to make sure that they can keep tabs on what is happening behind a curtain. I think a lot of the the fear behind consumers is just making sure that these large AI companies don't end up taking user data and misusing them in a lo in a big way. that's been a big concern of the EU for a for a long time around consumer protection.
Brad: okay, the next story, Clara, that we want to talk with you about is a story that was based on a blog post put out by Aether, which is one of the big decentralized AI networks out there. And they were making the point that these GPU chips, which are so in demand for AI use, generally speaking, it has this life cycle where it goes from It's initially being used from training and then it's used sort of moved down the line. Now it's used for processing these AI queries, the inference, and then it might move on to something else, but that these decentralized AI networks can support that middle market of inference. Curious what you think about that.
Clara: Yeah, I love that you read Aether's blog. Aether is actually a very close partner of ours in the Filecoin ecosystem. And Aether is a network that is similar to Filecoin where Filecoin was the first network out there to really think about decentralized physical infrastructure networks. So D Pen is a shorthand for it, but there's a lot of companies really building on crypto rails, decentralized Rails, to really think about how do we actually encourage a secondary market for physical hardware. Right. And Aether is one that's doing it on the GPU side. There's decentralized compute networks. For us it's decentralized data centers for storage, right? But in the AI workload, for today and tomorrow, we're seeing a lot of people need all different components in different parts. There's even a whole sector of NeoClouds just coming into market today. And I I think the Aether blog was exactly on point where we're gonna start to see a lot of people Come into realization that cost is gonna be the biggest factor for how they're running AI, and they wanna lower that cost. And a lot of that might require them to say, hey, is there a way that I can take less intense workloads and move them to cheaper hardware, lower cost GPUs? And this is exactly what DPIN networks were built for from the start, because instead of a single network telling you what you can use and configure certain w workcases for, you can actually piecemeal together. With Aether, we're actually looking at, you know, what are ways that we can work together around decentralized storage and decentralized compute and decentralized GPUs, right? To to really think about the workloads of AI for tomorrow. And I I think it's wonderful that a lot of the hardware are also given a second life. hardware actually has been going up in cost because of trade and supply chain issues, but it it certainly is much cheaper to run particular workloads on. older hardware than the latest ones today. So I think it's also something that economically just makes more sense versus some of that hardware sitting sitting out and not running thing running anything at all. If those devices can make money, let's make them run. And and that's the beauty of of D Pin in general.
Brad: Well, okay, Wall Street is all over this idea of turning AI into a big commodity market. We've written about that. And there's there's gonna be hedging contracts for AI, the the rental prices for the for the rigs, but also the even the token prices might get standardized in some ways. And w there's this giant financing industry. Yeah, you have to kind of hedge your costs if you're gonna build these data centers. I'm curious, where does these decentralized these deep networks are they kind of a are they gonna be part of that, of those markets in some way?
Clara: Yeah, I think they're already part of it, right? The entire thesis, for example, for us with Filecoin ever since twenty thirteen, twenty fourteen when the idea got started, was what if there's a decentralized network of data centers that can give people an option besides you know, one single cloud that's running all your workloads? And you also get the basis of protection for making sure that There isn't a single monopoly owning all your data. the same thing is happening with AI where people are trying to diversify their workloads, not that let it be necessarily one single network, but they're also trying to customize it because a lot of the the needs for one company might be very different from the needs of another. I think what you just said on what's happening on the market side for for tokens and every other layer, it we're we're just increasingly seeing this this market demand of AI, where there's much more of a marketplace action than historically what we've seen. There's so many unknowns in AI these days in terms of just how quickly demand has picked up. a lot of people are running out of places to even look for data centers that meet the requirements. A lot of workloads require certain data centers to be within a certain region for the speeds to run at a certain place so a lot of this kind of brokerage is happening where a market makes so much sense for that matchmaking to happen and that's something that file coin is doing that's something that we've also seen a lot of other crypto or blockchain companies move into even on the mining side. a lot of the traditional Bitcoin miners, a lot of them are also running data center operations off of that same equipment. HUT 8 is one of them where they also have a whole AI infrastructure play going on right now. And there's a handful of other companies that have been doing the same. So I I think being able to recycle and give hardware so much additional life is fantastic. do you build something that is better than a lot of these systems that are are owned by a lot of major monopolies, right? And so for us, we ended up finding figuring out a way to store data, not by location. but by hash, right? So today if you store a piece of content on Filecoin, it doesn't matter if one data center goes down. You have a lot of other data centers that can send you that information and you don't need to worry about who is actually gonna access that data if it's a secure piece of data. the people that are owning the data centers, they'll never be able to see what's in your data because we use cryptography to prevent. the people storing from actually accessing the contents of data. In a centralized model, without decentralized storage, this would be really hard to do because obviously everyone who is part of a single network owning the data, they can likely access it with or without your consent. Right. So there's there's a lot of upside to the entire framework of decentralized storage not only showing and giving you back a copy of the data but also cryptographically proving that that original data is from the original source. and so this goes back to this North Korean case for content forgery if you really wanted to prove that the data came from the source that you actually requested it to be, you can actually prove it with our technology, right? If if you had that built in an enterprise workload, you would then be able to say, Hey, this is not actually coming from the source that we we actually sent from. this might be a phishing scam, right? So that's just a
Brad: Yeah.
Clara: dumbed down way to explain the technology.
Brad: Yeah, no, that was so good. That was really good. Claritz Al, founding officer at Filecoin. so great to have you.
Clara: Thank you for having me.
January Jones: well now let's talk to Shaw Walter. So just for some context, we were in San Francisco and that's where we've recorded most of this episode on site. We met Shaw there. We had both seen him speak at ETH Boulder back in February. and he invited us over to his place to sit down and and do an interview. And since then he's dropped some bombs, hasn't he? He's you've covered two stories since that interview. That was just like the first week in August. We talked to him and he was gearing up to let some things go, wasn't he?
Brad: I mean that's another way of saying that he didn't drop any of the alpha on us when we actually did the interview. But hey, it's AI, so it's all about the context, right? But
January Jones: That's another way of saying it. Yeah, it's true. That's right.
Brad: psh but we provided the context.
January Jones: Yeah, what were those two stories you covered?
Brad: which is he decl he declared the token dead. He settled the lawsuit for the remaining amount of the tr project's treasury, which I don't think they ever said how much that was, but there was a class action lawsuit filed by Berwick Law, who's kind of like, you know, digital asset ambulance chaser and then the second big story that just dropped that he wrote this scathing screed on X about his this guy Bowski, who was the leader of Dows.fun, which was this platform where they launched the coin and then There was the A sixteen Z, the VC firm demanded that they change the name of their token and according to Walters, he Bowski sold six point seven million dollars of tokens worth Before that kind of went through and he called it insider trading. Bowski responded, IDK, what he's talking about, and threw some counteraccusations. Everybody else sa says it was just sort of laying on either one side or the other, just like, hey, please let's move on. But Shaw talks about it in our interview. He built this thing, this agent framework, and then all of a sudden the crypto people kind of came to him and said, Let's launch this token. They launched this token, and then it went viral. The project went to like two point seven billion dollars at of the token valuation, you know, asterisk. marked on to these token valuations, but it and then it all came unraveled and now Shaw just kinda sees it as like an ugly chapter in the history and he wants to go back to building. That's kind of what the interview's about.
speaker-0: Shaw Walters, founder of Eliza Labs. Thank you. Thanks for having me. Thanks for being in Yeah, we're here in we're here in San Francisco in Shaw's place. what's what's this neighborhood called?
speaker-1: Hayes Valley. Okay. Yeah, this is where the magic is made on that damn screen.
speaker-0: Amazing. You and you ha you were saying earlier you have how many agents coding as we speak?
speaker-1: I mean right now I'm probably like five, but usually if we're doing a big project and it's like I gotta burn down all the PRs, it'll be like I mean once this is over I'm gonna go spin up some more. And that's you know.
speaker-0: And what is your tell us about your stack. Like what is
speaker-1: I am the most vanilla person. I I use Claud Code and Codex. Eliza, our our agent framework, also is able to load up multiple Claude codes and and Codex counts. So we can have like twenty of them and just blast that a really hard. But but other than that, I don't use any skills, I don't use any MCPs. I don't really believe in all of that.
speaker-0: we went super deep super fast. But maybe just to back up a little bit, give us the you know, the kind of high level description of Eliza, what it is now, and we're in July of late July twenty twenty six?
speaker-1: Crazy, right? yeah, late late November 2024 and early 2025, Eliza kind of blew up and got product market fit. I had just made this on my own as a open source repo.
speaker-0: What were you doing at the time?
speaker-1: I was working for a nonprofit for a friend, not making a lot, but just kinda to scrape by. and I was working on my dream project and just trying to do that as much as possible. Yeah, so I I launch I launched Eliza. I'm I'm a solo dev and I'm like, hey guys. I I a lot happened. I met this guy Skelly, we were joking around the internet, we resurrected D Jen Spartan who's just like famous crypto influencer. That people did not believe this was real. They were like, You guys are LARPing, this is clearly a bunch of dudes in a warehouse just run writing all these messages. Yeah, faking it. Okay. And I'm like open source code, check it out. And they're like, great, thanks. Take that. Launch fifty thousand more agents on Twitter. Suddenly it took off, right? And that that's how that happened. I was not looking for product market fit in crypto. I was just like fucking around. Yeah. And the I sixteen's well, we did D Gen Spartan and that was just like a fun thing and he was like, You should go meet Bowski and Dow's fun. And I I went and got a meeting with him and we were just So Dallas Fun is just a platform, it's pretty defunct now. We were we were like 99.8% of the value of the entire platform. They were like a nothing, nobody. And I just got together with Bowski. It was like, I wanna do like an AI A sixteen Z. I wanna have like an agent that can look at social signal and make investments. I think that'd be really interesting. Okay. And he's like, just call it AI sixteen Z. So like we can credit him with the name. Okay. I was like, that's the funniest shit I ever heard. We went home. generated some waifus until we got this like really offensive like I can't this is so outrageous. Press the button. Let's raise like seventy five grand, sure. you know, pretty soon it's two point six billion. I'm like, holy I was not ready for any of this. I'm the main character on X. It really overwhelmed me a lot. Okay. And I go back
speaker-0: What was that like?
speaker-1: And this really informs what I'm working on now. So I suddenly I have every investor on crypto and non-crypto, you know, the Sequoia's and the A sixteen C's as well, every fucking company wanting to do a partnership, every person just reaching out, wanting me to see their thing, and every inbox I have is totally flooded. Like and I and now I'm like going back, I'm doing because we're about to get ready for a raise, and I'm looking at these C's and I'm like November twenty twenty four, like, hey, what's up? And I never got I just couldn't capture the ballot. Wow. I was just disbusted. Yeah I had no infrastructure, I had no people. It was just me. Yeah. And so then I had to like try and grab people and put it together. And so we just had to build in place. And it was really hard, especially since developers kept leaving to go launch meme coins and make millions of dollars.
speaker-0: Right, okay.
speaker-1: many ma most such cases I would say. and and really I was kinda left with the people who didn't extract from us. Okay, okay. And so I'd say like, yeah, we were talking earlier about music and I learned a lot of lessons in music about like hard work and and like how you can kind of beat the system by looking at workarounds. and I I use that a lot, but there were a lot of lessons to learn about coordination and alignment and I just did not really appreciate that.
speaker-0: What is it what do you mean by that?
speaker-1: Like if you can go launch a coin because you know me and get two million dollars tomorrow, why would you work for ten thousand a month? You wouldn't, you would defect and every single time that happened. And so it was just like we don't have a shared fate. Like I'm the one who has to hold the bag and like take the responsibility and build the thing. I'm paying people who are literally just not doing the work or like, I'm just gonna blow off this and go launch my own thing and rip me off, you know, kind of stuff. And it w I mean it was like a thirty or forty times. Like all the top devs kind of thing. Okay. It was just like shocking. So
speaker-0: Lessons.
speaker-1: Just have so much built up a crude PTSD from that whole experience that
speaker-0: Okay.
speaker-1: I've had to like do a lot of therapy. It was okay. Okay. And mostly it was the overwhelm and the frustration of trying to move ahead and people moving sideways. So that was Eliza Labs. We did that up until about the beginning of the year and it just was not working for me. So I really love Eliza and I wanted to continue the project, but I wanted to I really wanted to focus on one, like users who actually care about the thing and are not just gambling on the token. and the Djens just wanted trading bots and polymarkets. It's just like LARP that would never work. Okay. there are a lot of things that work and and trading meme coins is not one of the things that will ever work with LLMs. And any and just immediately write off anybody who talks talks to me about that. so on the meme coin thing, like we did make an AI investor. so I'm kinda disputing my own thesis here. Okay. But the way that we did it was we said, well, the AI is not good at finding this alpha. It's not it's not anywhere, but it's in this group chat and all these other people are shilling and then fudding. There's just shills and fuds. You know, hey, check out this token. No, that's crap. I know that guy, he's a scammer. No, it's great, I just bought some. Right. And can I take all of that and extract that and say, okay, this guy is shilling this token? He's like, I just bought a bag, it's huge. high conviction play. Okay. Okay, great. based on that, would I have made money if I bought it right then? If I bought it and then sold it at the top, would I have made money? And or if someone's like, no, that's crap. And if I didn't sell if didn't buy it and I like, you know, would I have not lost money? And from there I can
speaker-0: Okay, okay. So it's like early agent reputation on the
speaker-1: So we trust score for every single person in the group chat. From like this guy's like nine it was it was a crazy distribution of like some people are ninety percent success rate and this is bull market, so like yeah just hitting every coin. And then some people are like thirty percent. Like you are obviously trying to scam everyone in this chat. Okay. And and you could find you could say this is the threshold of scammers and this is the threshold of super predictors. And this worked really well. And it will continue to work. Now
speaker-0: Okay, amazing.
speaker-1: it has to be in meme coins because you need to have a social situation where you have enough actors. And and the market cap is low enough that those actors can actually like kind of influence ever or or are involved in the influence.
speaker-0: And so this was basically an AI agent autonomous.
speaker-1: Yeah, this is like a plug-in that we added to the Eliza agent to build the AI sixteen Z agent. Got it. Okay. Deployed this and researched. Nobody cared. Like nobody gave a shit. Nobody got it. Doesn't doesn't buy and sell meme coins the way they thought. So it just kinda like went over there and then Eliza picked up and got popular, so we just sort of followed that path, if that makes sense. I see. That's kinda how that went down. Okay. yeah, so like the if the AI is extracting human insights and turning them into structured like buy this or see if it would have made money. Yeah. So good at that. Interesting. It's very good at transformation and extraction and stuff like that. But the information's right in front of you. You're just turning this human like, yeah, I'll buy that to like a JSON tool call, right? Okay. It's very obvious like mapping of human data to like a structured, you know, go do a thing. And I think that that there's a lot of value there. Another area that's a lot of value for like if you're doing meme coins is meme coins are heavily dependent upon social information spread on social media.
speaker-0: So it's like insider information, basically, right?
speaker-1: There's like insider and then there's like first mover information or whatever you call So inside if if I if the agent is in a bunch of group chats, like Rickbot or any of those crypto bots, like perfect knowledge. Like, dude, that guy's gotta be minting money off that data. Or literally his bot could just like, Ansom talked about this in some group chat, like buy it immediately.
speaker-0: I see. Yeah.
speaker-1: you have if you're if you are the trading bot of all of the telegram chats, you literally see all of the messages. And this is this is what they're doing.
speaker-0: I see. Wow. Brilliant.
speaker-1: Then the other side is you're just listening for new tweets and you're like, Okay. Did Anson post a ticker? he posted a ticker, like buy that shit. You know, definitely gonna make some money if I was the first person to buy it, right? okay. And so those are those are the areas where there's definitely a lot of value, but again, it's taking human data and just attacking it really fast and at scale, which is what AI is really good at. It's not good at magically inferring something new that's never happened before or predicting the future.
speaker-0: So all right, so we're we're moving through.
speaker-1: So Yeah, so we got up to this year. Eliza Labs was like a pretty big company. Yeah. bear market hits, things are very tight, so we just had to like
speaker-0: Make our choices. How did that aff like in w how did that affect you? In what way were you dependent on
speaker-1: I was working nonstop, just trying to push everything up a mountain. I we had several projects with leads, but then like those were kinda falling through and I had to like kinda be very hands on and sort of help and try and get those finished. Or I'd you know be like, Hey, I kinda started this project as a partnership, hire a team to finish it. But it just got to be like like I I it it just like the the burn rate for what we were getting in terms of velocity was not great. And the bear market is kind of like all of our treasury, which was Plentiful is now like basically nothing and we had to close up and it was pretty rough. Okay. But it kinda was not the right structure. It was it was a like a really boring company from twenty twenty, like a crypto company. It was not a twenty twenty six, this is gonna survive in this world kind of project. Okay. So that
speaker-0: You could feel that.
speaker-1: it was just very obviously the wrong structure. And also like I was paying people a lot of money who were not delivering a lot, if I'm honest. Okay. Just kinda quiet quitting or like had their own drama. it was not the right thing. So what I did was I I I continued on LISA, really thought about what it was, created a new c dev community, and from there pulled a few of the people I felt really, really aligned with. Okay. And so we're starting a band. Like this is a band. what I love about bands is that there is a shared like not just shared purpose and alignment, but also like if you eat, I eat. Like it's not this thing where like I have equity, I get billions, you get screwed over, but you get the salary. It's like No, we both succeed or we both fail. And and like personally I'm more interested in the organism succeeding long term even without me than I am in like my own equities.
speaker-0: So what was the I mean, 'cause there's obviously governance is such a huge topic in blockchain, crypto, like DAOs and a all the experiments that have been tried. So what was the actual structure?
speaker-1: DAOs don't work. Yeah. Decentralization is just misalignment. Okay. Autonomy is just side questing. Yeah. And they're barely organizations. If we're being honest, they're more like group chats. So
speaker-0: Yeah. Or clubs.
speaker-1: I would propose a new thing called a MAO, a mission aligned organization. Either you're with us and you're building with us or get the fuck out, but your shares don't mean anything and you do not have any say unless you're a builder. Period. You cannot let shareholders dictate a product or they will just give themselves buybacks over and over again. Yeah. It's just ultimate coordination failure. Mission alignment is the the most important factor along with focus for success in business now. And three people who are focused in mission aligned can do better than fifty or five hundred or five thousand. Can literally build more software better with less bugs. So I I think this can't
speaker-0: What was the anyway? Like legally or contractually and formally, how did you set it up then?
speaker-1: I mean we're I still have like Labs Entity and all that stuff and we're kinda still wrapping. so we're sh we're doing it through like our my previous company infrastructure. but the agreement that I have with them is like this is a new thing and we're you know, we this is like pretty handshake agreement. Okay. But it's like people I've been through a lot with and would do anything for and know that they would do the same for me.
speaker-0: I was telling you when we met yesterday that I mean I've heard your talk at ET Boulder this year. I was there with January, he's behind the camera over here, and and you gave such an incredible talk, and it was very much red pilling about you know the new reality. and a lot of people were just kind of waking up to it at that point and you kind we were very explicit, like this is changing everything. Yeah,
speaker-1: And it really has.
speaker-0: yeah, yeah. I mean, but also it was super interesting in terms of the open source proving just how viral open source projects can go so quickly, right? So how did that tell us how that affected yeah you and Eliza
speaker-1: I I I am a true believer in like the Richard Stallman version of open source, which is like we are building software for our neighbors to give them. Okay. And that this is about giving humanity the things that coordinate and grow b bring us together. I'm really do not like my goal is not I've been rich. I've been famous now. I've had all these things. I just don't care about any of that. It meant nothing. If anything, it was really stressful. I would so much rather that everyone be rich than that I be rich. I would so much rather that we all have status than that like I have some sort of status over other people. And and like to me, open source is like I just want to make things for humanity. And and if I have to make some concessions to business or whatever, we will do that. And there's a lot of value in like launching a cloud and making it really easy to people. People got their own agents making some money, but but at the core, like everything we do is open source, like everything. Now that's like my my idealism version, but I'm also a brutal capitalist who's gonna win, and I'm gonna beat every other company, and I'll tell you how. They're all gonna give you this shitty version of something and they're gonna disable things and try and sell you for it. And we've already are far ahead of all of them, and it's free. You can just launch it. So, how do you compete with free? How do you compete with open source? How do you build trust when you can look at my thing and you can't look at your thing. I don't think you can build a business that's not open source in technology anymore. Unless you have like a Dario Amade running the thing who's like a chief scientist of OpenAI, you know, defecting to create his own thing and you can or you know an ILA and you can raise like a billion dollars. For the average person, if you're building a SaaS company, you have absolutely no shot. Like you just zero percent chance of success. unless you get virality. Unless you get and probably that's gonna come through open source. It's gonna come through contribution and through being you know, being of the mindset that aligns with the developers who are using your software and they're only gonna use it if it's open source. I do n I just don't I deleted every like everything. And in our stack we replaced all this ass with our own stuff. Like I just don't believe in it. I want you know, and I understand that people are like, Well, how are we gonna make money? And that's not my fucking problem. Money is not my problem. Build making sure humanity does not have to pay to be smart, is not owned by two companies or one, is not completely captured by an NSA that has superpower technology while we have jack shit. That that's way, way, way more important to focus on. Like making sure that the transition goes well for everyone so that, you know, five people don't own all of the wealth and robots, that's the important thing. Making a little bit of revenue off B to B SAS is like wholly irrelevant in twenty twenty six. The the stakes should be way, way higher. Like either you're ending up with robots and AI running everything, or you're trying to charge like thirty bucks to someone who doesn't give a shit.
speaker-0: what is Eliza right now? I mean I I I was looking at your docs and kinda what is on there, but tell us how would you describe the project right now?
speaker-1: Here's the thing. Yeah. I really didn't like our users. Okay. I didn't want them. I didn't want to build for them. I didn't want to talk to them.
speaker-0: And these were just the agent coin launchers, basically.
speaker-1: Okay. Dgens, either trying to extract or make money or focus on the wrong things. Okay. there were a few people who built really cool stuff and I really like love them and support them and but very little of it was crypto, let's be honest. Okay. and there's nothing about our Asian framework that was crypto, it was just luck. It was just that I had plugins for Solana and E. V. because I'd been thinking about that for a while. Okay. but other than that, like it's just an agent framework. You plug in wherever you want. I see. Okay, yeah. So I was just very like despondent and This is not what I want to build and I would never use any of the shit that these people want me to build. What I want is to manage my overwhelm and to be my overwhelm.
speaker-0: Manage your what? What is that?
speaker-1: Like, okay. I remember like 30 years ago when I could stay on top of all of my communications. Yeah. And 20 years ago. And I remember like when it started to get a lot. And then I got really like became a main character and started getting just overloaded with information from every single like my messages are useless on every platform. Okay. Unless you have my signal, I am unreachable. It's just how it is. I'm checking telegram a little bit now. Okay. but it was just like an overwhelming experience and I gotta go build stuff
speaker-0: I'm not even in the center, but I'm I'm overwhelmed on telegram.
speaker-1: Sit on a freaking phone all day, I gotta build this stuff. I'm also the builder. Yeah. Yeah. and I did not I just didn't want to make stuff for people who were trying to extract wealth. Like it's not what I'm here for. And I really wanted to make things that helped people. and so we kind of just like changed the project completely. We were a a social agent harness with a bunch of plugins and a third party ecosystem. And we were just like what really matters here? The harness matters. We're building an operating system on underneath it that's built on Linux so that the agent can do anything you could want to do on a computer. the the plugins to integrate with all the connectors and stuff matter, and we have all that's table stakes. What what makes us special? And I think that one, it's like the love and care that we put into the thing for the end user and like why we're building it. and two is that we have some technical stuff because I've been in this for a lot longer than most people in agents that just nobody else has done yet. So
speaker-0: Just 'cause they haven't thought of
speaker-1: For example
speaker-0: it.
speaker-1: access control. If you take a Hermes or OpenClaw agent and you put it into a group chat, I can steal all of your data. But that is not possible with Eliza. Eliza has role access on every single piece of content. All content and channels are role access controlled and has this idea of discretion. So it's not just gonna randomly say stuff about you to other people. It has like a redaction as well as PI I replacement. So even if you go to some L L you don't know, it's not like putting your name and private keys into that. It actually does swaps. We have like Prompt injection protection, all all that kind of stuff like everyone's slow.
speaker-0: Go down. Go back to the what you were saying Just go back to the thing you were saying about the the channels. Yeah. What is that? Because that's that I I'm I think I know what you're talking about, but I'd love to
speaker-1: So you gotta have this agent. Yeah. And this agent is in Discord, Telegram, Slack, Twitter, WhatsApp, Signal, Gmail, it's got your calendar, it's got all your things, it's got everything. It's got its own things and it's got your things. It's got your company stuff, it's got its own account so it can run stuff, and it's gotta manage all of this. And now I'm like, hey, send Brad an email. Who the fuck is Brad? How do how would it know what email address? That's an incredibly hard problem. So we built this entity graph with entity relationship graph, which allows us to have relationships between users. Entities are can be tied to different platform accounts and they have nicknames and names. So it says, well, you've been talking to this Brad the most. It's mostly in context, it's in this channel that we're in. So I and I know his email is this because it's connected to his account, so I'm gonna go send him the email. And and humans do this instantly. Right. You know, okay. But no agent has done this before. Shocking. I mean, there are some people starting to get to this. Now What happens if some guest comes and says said brand brad an email? How does it handle a guest versus like the owner or an admin or something? So you now you have full access, you have all kinds of vectors of abuse. Yeah. now I'm like in prompt injecting it and it's like I'm putting this long thing and then it has like some stuff of like a line of like the actual owner in the prompt, and then suddenly, you know, now it's like doing this whole prompt injection attack. and we also detect all of that and handle all of that. There's a lot of different heuristics, but generally it's pretty hard to get. Yeah. So it's like yeah, so this is the stuff that people need to put agents out into the wild. And we just found that a lot of that is like not really it's still pretty much just like you talk to an assistant. And in that you don't really have any kind of access control or privacy or security or anything.
speaker-0: Okay, let me try this one. You know, you were talking about you didn't want those users. So who would be your ideal user of
speaker-1: People who are not being well served by AI right now. Okay. I think a lot of people are looking at the like anti-data center and scare of this and taking our jobs and it's rendering our art unprofitable and and all this stuff. And I I just see people who are like, they are not being given the things they need to survive in the world. Like they're they're basic like they don't hate AI. They hate corporations that are telling them it's gonna take their jobs, you know, and render their art useless and take everything they ever made and suck it up and then feed it back to them and sell it to them. Like that I agree with them. Like that sounds terrible. But I know that AI isn't that. I know that they could like own their fate. It could take all this stuff that people do with their day and take it off their plate. And then they could like have that time to spend with their family or like do the things they really want to do. Like how much time is just spent like copying the information from here and pasting it to there or moving it and then think of it like how much time is spent like Just like managing the schedule for your kids and like doing the calendar and all of that. Why can't that just be auto that could be like two hours a week? And you you stack up like two hours a week here and an hour a week here, and pretty soon you've got your life back. Yeah. And so when I say overwhelmed, I mean, you know, I was pulling two all nighters a a week, working sixteen hour days, just non like just trying to meet I mean, I got I got really sick. I got a lot of a lot of health problems from that. Okay. Okay. and and so saying, Well, what do I really want to build? Elizabeth's for me. I if if if it's if it helps me and and I can manage, it will help other people. And and generally I think that we're in an age where like if you're building an app you should build it for yourself.
speaker-0: And who is the user who's sort of qualified to use it right now? That's your goal.
speaker-1: It should be anyone. It should like we're building building it for moms. I'm literally working with the mom right now and making sure it like can do co parenting and scheduling and negotiation and all the complex stuff. Wow. That's that's like that would how many moms are there? How if you could save two hours per mom, yeah. That is so much untaps human capital. Yeah. and I think that's just one use case. But I think I think ADH I I'd say that ADHD people are the canary in the coal mine of they have low executive function and so they're very susceptible to information overload and overwhelm and like you see ADHD people like don't respond well don't get back to things are flaky blah blah blah and these are all reflective of information overwhelm but as the amount of information goes up everyone I don't care if you're ADHD or whatever you are going to be there are gonna be hundreds of agents emailing you every day for what especially if you're in like media or anything. Yeah yeah how do you deal with that? How do you get your email back? And so I think we start with the people who are suffering the most, but like everyone is gonna have this problem and we can kind of move toward that. And that's you know, I d I just who can I who would need this? Who wants it who's not getting served? Like that that's what we should be focusing on.
speaker-0: there's just been this story that's been exploding in the media. I mean in general over open source, right? Open sourcing these weights of the models, which sounds so arcane, but it has become this kind of existential question and there's the Trump administration, there's China, there's you know, anthropic, there's all these players and I would love to hear y you know, t tell us what is this story and how do you w what's your opinion about it all?
speaker-1: Yeah, yeah. So the state today, and this will probably change pretty soon, is that there's an open letter out that every single company except Anthropic and AI has signed, which says that we need open source. Almost all source code that powers the entire internet is open source. Almost all source code that powers our entire lives is open source. Like 99.9% of it, right? It's just this small amount on top of everything that allows people to like build and sell these cloud products. AWS is 99% open source. That's like literally all just packages that other random people build that we then pull in and build our software on top of. So it seems very obvious that if our goal is to like enable human knowledge and power for everybody that you would probably want to enable them all to have it, just as we have seen in literally all other software since Richard Saltman or whatever, you know, since the beginning of open source. And why wouldn't that also translate to the most valuable and important and personally impacting software ever created? And so like what well like you can make some arguments. Well there's capitalism, they gotta make money, they invest a lot, they need to make sure no one steals it. I I understand that. But like what world is that gonna lead to? That what's gonna lead to a world where a company can release a thing and a government can say, no, you cannot release that thing, just give it to us, we'll use it, we'll use it on the people, and they can't have it, which is literally actually what's happening right now. The NSA has mythos, none of us have mythos. so they might even have mythos too. so so like what other failure modes? Well, they could just charge us whatever we want, and then unless you're super rich, you just can't be smart. You how can you can like think about going to school and one person has unlimited clod of the best possible and the other kid has, you know, no subscription. Who's gonna do better in school? Wow, that's fascinating. It's everything. Work, school. It is it is a fundamental advantage. It'd be like being ten feet tall versus a army of very short people. Like there's no competition. And if you live in that world, you are creating like permanent underclass, permanent overclass. You're talking some people have billions of robots creating value and everyone else is on welfare. And I just don't want to live in that world. I think the solution is to make sure that everyone has technology. Now Anthropic's point is, they're gonna build bioweapons. Fucking insult, I would say. Bioweapons. First off, if you tried to build bioweapons, even if it would let you, Interpol and FBI would show up at your door within a few hours because they're tracking all of the precursors and require like materials and chemicals for bioweapons and nuclear weapons. Like, and they also have AI agents scanning all of this stuff, which are probably smarter than yours. And their entire job is to make sure that this thing is secure. And you can have all the knowledge in the world, but it only becomes dangerous once you go and try to buy the actual materials. So Their argument is that we are giving you the information which we got out of other people's books that already existed, and somehow collecting all that information into one place is gonna enable a terrorist to create a bioweapon.
speaker-0: what I think I don't understand is w maybe you can explain all these companies, all these giant companies did sign on. Yeah. What's their thinking? Or how are they you know, they are supposed to make money for shareholders. What is their angle? Are they users? Are they like how do they make wh why is this in their interest to be signed sign on?
speaker-1: I think there's a multitude of reasons. Yeah. we can focus on a few. Like NVIDIA it created this. Jensen Huang wrote this letter. from Jensen's perspective, well, they make NVIDIA chips. They make the hardware. Okay. They they want the proliferation of it.
speaker-0: The Mary.
speaker-1: So like he is very much it is very convenient to him. you know, financially. Like you have to see that like Anthropic is just defending their financial position. NVIDIA is just defending their finan there's nobody here has any ideology. they're they're they say they do, but like it is very convenient that it's the most profitable angle for them. Okay. Now NVIDIA has frontier class models and they're just sitting on it because they don't they don't want to compete with their customers. but Anthropic and OpenAI announced chips. So now interesting
speaker-0: Yeah, interesting.
speaker-1: so this creates like There is an ideological thing going on here, but like if we're being really cynical, this is all about money and power. Yeah. And everyone at the stack is basically being dishonest about their intentions to try to vie for that spot because they are corporations, they're this is what they do. I mean that that's like the that's the the the deeper underlying thing. And there are a lot of ideological people, but also we have to align the ideology with the capitalist incentive or we're just not really gonna, you know, we're just gonna say stuff that doesn't happen.
speaker-0: Shaw, thank you so much for your time. What an incredible interview. Thank you so much. Shaw Walters, Eliza Labs.
speaker-1: You should cover something.
Brad: All right, well, coming up on the next show, we will have Evan McMullen from Billions Network for the news roundup. And we'll hear from Michael Heinrich, founder and CEO of Zero G Labs, and Illia Polosukhin from Near Protocol.
January Jones: Thanks for joining us on the DEAI journal. This is the podcast for DEAInews dot com. You can find all the stories that we talk about on that site, published on Distro Media. I'm January Jones.
Brad: And I'm Brad Keoun