In this captivating episode of The Edge of Show, Ron Levy delves into the groundbreaking world of Decentralized AI with Matt Wright, the visionary CEO of Gaia. As the Web3 ecosystem evolves, Gaia is redefining how artificial intelligence functions, combining transparency, ethical responsibility, and decentralization. Unlike traditional “black box” AI systems that lack transparency and often compromise intellectual property, Gaia’s platform offers a transformative alternative. It empowers developers and organizations to build, own, and monetize AI agents while maintaining data security and ethical standards.
Matt Wright shares his incredible journey, from organizing over 150 hackathons globally to spearheading projects at JP Morgan and ConsenSys. His transition from corporate innovation to decentralized AI highlights his commitment to open-source principles and empowering creators.
This episode provides deep insights into Gaia’s innovative approach to creating living knowledge systems that enable businesses and individuals to gain full control over their AI assets. It’s a must-listen for anyone passionate about the intersection of Web3, AI, and ethical innovation.
Join Ron and Matt as they explore how decentralized AI is not just the future—it’s the present, and it’s reshaping how we interact with technology.
Key Topics Covered
- Decentralized AI and Transparency: How Gaia is addressing the risks of opaque AI systems by promoting transparent, decentralized knowledge-sharing platforms.
- Ethical Standards in AI: The importance of ethical frameworks in AI development and Gaia’s unique approach to enabling responsible innovation.
- Empowering Developers and Users: Gaia’s tools that help individuals and businesses create AI agents tailored to their unique knowledge and goals.
- Web3 Integration: How decentralized AI aligns with the principles of Web3, offering censorship-resistant and open-source solutions.
- Future Vision for AI: Matt Wright’s predictions for the next decade, including the rise of AI agents surpassing the human population on the internet.
Episode Highlights
- "Decentralized AI is the way forward. Gaia empowers developers to own and monetize their AI creations transparently." – Matt Wright
- "We’re entering the era of the read-write-think internet, where agents can analyze, decide, and execute tasks autonomously." – Matt Wright
- "By using open-source frameworks, Gaia ensures that knowledge systems remain censorship-resistant and bias-free." – Matt Wright
- "Hackathons have taught me that collaboration and feedback are key to driving innovation." – Matt Wright
- "Web3 and decentralized AI are leveling the playing field, empowering smaller players like never before." – Ron Levy
People and Resources Mentioned
- Matt Wright (CEO of Gaia) – Twitter
- Gaia – Official Website
- Ethereum Foundation – Website
- ConsenSys – Website
- Hugging Face – Website
- OuterEdge Innovation Festival – Website
- OpenAI – Website
About Our Guest
Matt Wright is the CEO of Gaia, a pioneering decentralized AI platform that transforms knowledge into secure and monetizable AI ecosystems. With a career spanning JP Morgan, ConsenSys, and the organization of 150+ global hackathons, Matt is a trailblazer in Web3 and decentralized AI innovation. His vision for Gaia is to empower developers with tools for creating ethical, transparent AI while fostering inclusivity and collaboration.
LinkedIn Link
Twitter Link
Call to Action
Don’t miss this riveting discussion about the future of decentralized AI and ethical innovation in Web3. Subscribe to The Edge of Show on Spotify, YouTube, or iTunes for more groundbreaking insights!
Transcription
Matt Wright: Hi, this is Matt Wright from Gaia. We're leveraging open source AI to build transparent, censorship-resistant cell phone knowledge systems. And you've turned into the Edge of Show, where we unlock the future Web3 and AI every week.Ron Levy: Stay tuned. Hello, Web3 and AI podcast passengers. Jump on in. Here's what's to come on today's journey. Discover how opaque black box AI systems jeopardize intellectual property and foster mistrust, and why our guest is determined to change the game. Learn how Gaia's groundbreaking decentralized model is redefining transparency and setting a whole new standard for ethical AI. And finally, hear our guest's bold vision for a future where AI empowers users while still protecting intellectual property. All this and more, take your seat.Intro/Outro: Welcome to The Edge of Show, your gateway to the Web3 revolution. We explore the cutting edge of blockchain, cryptocurrency, NFTs, ordinals, DeFi, gaming and entertainment, plus how AI is reshaping our digital future. Join us as we bring you visionaries and disruptors pushing boundaries in this digital renaissance. This show is for the dreamers, disruptors, and doers that are pumped about where innovation meets culture. This is where the future begins.Ron Levy: Welcome to The Edge Of show, featuring a variety of top-notch guests and other hosts. Today is a special AI edition, and I'm Ron Levy. I'll be your host today. It's another production of The Edge Of Company, a quickly growing media ecosystem which empowers the pioneers of Web3 tech and culture. Edge of Company is also responsible for other groundbreaking endeavors, like the OuterEdge Innovation Festival in LA and in Riyadh. Today's episode features Matt Wright, and he is CEO of Gaia, leading innovation and decentralized AI. Previously, he co-founded EVM Capital, served as Director of Consensus, and played a key role at J.P. Morgan, expanding Quorum's global reach. With over 150 hackathons organized for Fortune 500 companies, Matt champions startup culture, strategic partnerships and technology, advancements all across industries. So Gaia is an innovative decentralized AI platform that transforms knowledge into a secure collaborative ecosystem. It empowers developers to create and monetize AI agents while promoting transparency and rewarding knowledge creators. by addressing issues like censorship and bias, Gaia fosters an inclusive environment for innovation and knowledge, sharing, redefining the future of AI. So Matt, it is a pleasure to have you on here. It's good to see you. Pleasure to be here. Well, can you share a bit about your journey into the world of AI and blockchain and sort of what sparked your interest in the fields?Matt Wright: Yeah, of course. So I started my career in corporate innovation and developer communities. I used to organize hackathons in 65 different countries around the world. We were a small company called AngelHack at the time. We essentially did hackathons for Fortune 500 companies like MasterCard, Motorola, Barclays, UBS. A bunch of the early Web2 stack that you would see at the hackathons we were working with. Around that time, I actually got into blockchain. I think around 2015, we were organizing a hackathon for blockchain. That was the first time I drank the Kool-Aid. Jump down the rabbit hole you don't want to know what you know decentralized and open source technologies you know meant for the world and i think you know exploring that it really was taking kind of all of my philosophies of like how i. interpret life as I know it and technology to being something that's holistic and can create decentralized value systems. For me, I went head first. I just wanted to build in that ecosystem. And then shortly after, ironically i got picked up by jp morgan and uh you know we the reason i jumped in was uh you know they were taking ethereum and forking it and making the work for banks enterprises financial service companies governments Um, all in open source. And so although JP Morgan, um, we were definitely leading the charge in open source in the enterprise world. And to be honest, like, uh, and still some of the major technologies you see in the space, uh, in the web three space today. And then at ConsenSys, you know, that project was acquired by ConsenSys in 2020. So I jumped in and, you know, wanted to work on this, you know, Internet of Value. I was absolutely enamored with the Ethereum ecosystem and what people wanted to build in the space, you know, how we saw, you know, the development of this Internet of Value. And through my experience there, which was also quite interesting in itself, I was initially building out developer community teams at ConsenSys across our entire stack. And then I got really into, of course, NFTs, DeFi, but DAO has really struck a chord with me being Being a dev community person my entire career, I saw what decentralized communities or organizations could do for the space. And seeing that the autonomous piece wasn't really built out yet, there were some huge opportunities to grow. So I did some work in DAOs for ConsenSys, and then my last stint was building out a corporate Accelerator program inside of ConsenSys called Fellowship. We were basically helping seed stage teams, you know, that raised capital, build alongside MetaMask and Linea in a more structured approach, providing developer feedback to our core engineering team. And that was a huge success. I also got to battle test some ideas that I had around kind of this thesis around future of work that I've been, you know, talking about for the past, you know, two years, on like, if, if we are building, you know, this autonomous, you know, value system over the internet, you know, how do we, how do we sustain ourselves with kind of the bureaucratic processes and recursive processes we see in web3? Like, if you're a power user web3, like, is it not annoying that we have to, You know monitor you know the volume of transactions were doing on a daily basis monitor like the amount of projects are coming out and different you know metrics on how the products are performing the. you know, the governance of some of these token communities that we own or like the work we do inside of them is just, you know, absolutely noisy and cumbersome. And so, you know, I was working on this idea of, you know, how do we automate a lot of those processes? to make our interaction with Web3 more abstracted, just to be more, you know, automated. And from there, I partnered with ByteTrade over in Singapore, they're building out the initial kind of MVP of Gaia. And it's kind of one of those big aha moments where I literally flew out to see these guys. The next day, I looked at their GitHub. And you know, we basically aligned on how we would, you know, take open source AI infrastructure with Web3 Rails into kind of the mainstream. And so it's been a big adventure this past year. I think, you know, my, what drives me in this space is open source, decentralization, and I see decentralization on you know, the legal or structuring components, the, the technology and network components, and then also the economics of these systems. And so, you know, I think guys, it's amazing opportunity for us to, you know, take what we've, you know, seen, working in open source AI and enable us to validate and monetize some of this, this compute in, in a web three world. So that's my story.Ron Levy: There was a lot there, and I think what what comes to me a little bit is You came to decentralization and the true value of it by going through Corporations right JP Morgan the biggest biggest bank in this country in the second biggest bank of the world from what I recall in consensus being possibly the number one company in in well, maybe measured over the last 10 years in crypto space. And then let's touch on Web3. Instead of giving a big, long description, I think Web3 is the ability to exchange value unlike we ever could in the past. And that to me empowers individuals and smaller companies in a way that they've never been able to in the past. And so I wanna frame all that to say, How you are the right guy to be working on this and developing this because you've been on both sides and you've seen that difference And we're going to cover kind of, that was a great backstory on Gangaia and how it came to be with you involved in it. We're going to cover kind of what that means and what it is. But let's start with, you know, black box AI systems. Can you explain sort of what they are and why they pose risk to IP protection and public trust? Because that's foundational, I think, for the conversation.Matt Wright: Yeah, and I won't go into too much detail on that. you know, what exactly they are, because I think that's kind of why we call them a black box. We, meaning the AI industry, don't know exactly what is going on. Some might know more than others, but, you know, it is the concept that as these big tech, big AI companies are, you know, training models, leveraging models inside of kind of AI infrastructure that we use like a chat GPT, for example, it is it is communicated to us that they are not using our data that um you know they're they're in charge of our safety they're in charge of you know making sure that ai is doing the right things but there is this concept of there being a black box or a an area where they aren't able to know exactly where some of the ai uh outputs are coming from or like what exactly is the thing thinking 100%? It's like, it's like, you know, knowing a human as like a friend, but like, not actually being inside their head 100%.Ron Levy: It's kind of like the Web2 world, all the companies, large companies that are collecting our data, we know they're collecting it, but we don't know what the heck they're doing with it. This is kind of a similar analogy for Web3.Matt Wright: It's similar but even more like meta because like, you know, some people inside of an organization may know where the where the data is and perhaps like what the format of the data is. In this situation, if we really want like authentic or organic thinking of a AI or AGI system, you know, it's, that general intelligence is like really hard to understand kind of where it's coming from. They have a general idea of how the large language models and a lot of the, you know, infrastructure is working, but the, again, like, There's a difference between like asking someone a question and knowing kind of what they're thinking but not actually being able to see like if there's something they're hiding, you know, in their mind. And so, you know, there's there's that component of of AI that, you know, in decentralized AI we you know, we will face similar risks. I think what we're trying to do with Gaia is build more transparent systems that have, you know, more programmability around people training these agents and, you know, really being able to use smart contracts and blockchains to track you know, who's providing the AP and how, you know, do you exchange this knowledge with and pay for it with the, you know, righteous owner. In the centralized AI world, we don't know exactly, you know, we don't know that the architecture is being trained or not on certain pieces of data that we're contributing to these systems. And by us not, you know, getting any recognition for that or monetization of that, it's, it's created a lot of IP and legal concerns with, you know, bigger companies and media companies. There's like, you know, New York Times, and I believe Washington Post are both like suing OpenAI because you know, open AI parents training all these large language models on hundreds of years of, you know, media content, and they're not being credited or paid for that. And so there's some like major IP breaches that will continue to happen. It's also just this massive gray area of Like, what is public, what is not, and how is this entity using that data? That's kind of this black box word.Ron Levy: Yeah, you know, deep issues and deep problems, and some will have some solutions, but some will, unfortunately, just be open-ended questions. Are there ways that an organization can ensure that they are using AI responsibly? while navigating challenges posed by the black box models that you just described.Matt Wright: Yeah, so we're approaching in a way, we believe open source is always the right path. I think there's two major philosophies in technology, you know, there's closed systems and open and, you know, regardless of like, if it's blockchain or web three, or, or AI, like, this has been a conversation since the advent of the internet. And, you know, we're on the camp that You want to have open source distribution of large language models and AI tooling and even AI inferences kind of like what we specialize in. And what that means is that we can talk to developers in the open to improve the models, create governance systems around these models and tooling, create ways to incentivize developers for contributing to these systems. Instead, in centralized AI, because nothing is programmed with smart contracts, we're not able to really We're not able to use a lot of primitives that help us track what is being contributed to the systems and how do we govern them? How do we monetize them? But we are able to do that with Web3.Ron Levy: Right. And that's where the distrust comes from. It's really very difficult, if not impossible, to know where all that information derived from in the beginning. There's opaque AI systems. I think they help. contribute to mistrust among users and stakeholders, particularly in some sensitive industries. Can you speak to that a little bit? Can you ask that one more time? How do opaque AI systems contribute to mistrust among users and stakeholders, particularly in sensitive industries?Matt Wright: Yeah. So, you know, I think as you're building technology for like, if you're building AI for like a healthcare provider, for example, um, you're, you're always going to have major privacy concerns on how that data is being used. You're going to have major, um, uh, you know, legal concerns around how the large language model is trained, how much can be controlled. Um, is it, you know, crawling through some internal, um, Internal documentation and learning and putting it in some, you know, in some black box that you don't know about or perhaps it's just continuing to train the model without these folks even knowing. Those are concerns that like the industry widely is having to deal with. The way we're approaching it is we don't want to custody anyone's data. We want to enable anybody to become their own AI company so they can handle their own privacy concerns, their own permissioning, their own governance, and even think through how these things are monetized. And so what we're offering for folks is to help them become their own AI company. Instead of just giving them an API and point that you know they can use and gives kind of you know just golden key to like access all their data. We see that being a huge problem with enterprise. We also see it being a big issue with Web3 protocols that have rich on-chain and off-chain data that they're just sitting on and right now a lot of them are enabling AI um, by, you know, using open AI API, um, as their end point to build kind of agentic workflows off of their data, but. you know, they might not understand that they're actually feeding this AI system with this data, whether, you know, OpenAI says they're not treating their systems on it or not. And not to keep digging at OpenAI, like, you know, I think they've done tremendous work. I use Chachabuti, but like them and the Anthropics of the world, like they're building great software, but we just don't know exactly like what they're using this data for. They also are accountable for kind of the privacy and permissioning and governance and really should not be left up to the hands of a select few. Not to mention a select few who might not understand the actual industry that they're working in. What you want to do is enable a company to manage it themselves.Ron Levy: Yeah, you know, you mentioned healthcare, for instance, and there's been lots of projects in the blockchain world that have the goal was to empower people with their own health care data, right? Get every doctor's appointment and every medication and everything you've ever done, and be able to put that on your own personal blockchain and give permissions for other doctors or even researchers if you choose to have access to that data, but you would solely be the one in charge of letting people through that door. That was always sort of an end goal for blockchain, which you probably know better than I if it actually was ever truly, truly happened. But I feel like it's kind of been leapfrogged into what you're talking about. So with Gaia, we can have our own agent, so to speak, put all that information in. And while that's not blockchain in itself, it is in an ecosystem we control completely. Whereas if it is in a large language model, developed and controlled by others, somewhat in a black box, We're not going to know if they're using that data or not using that data. There's just no way to know. So you're talking about being able to separate that and control it ourselves and do with it what you will. And I would even say that would include corporate information. A lot of people are using GPT for for information they wouldn't otherwise give to anybody else, right? Yet they are putting it into a big company to pick into it as they choose. And maybe they don't realize that, or it's this nascent development of public AI to where people are willing to overlook it. So maybe you just speak to that commentary in regard to Gaia and help form that a little bit.Matt Wright: Yeah, I'll take this opportunity to kind of unbundle what Gaia is. You know, Gaia, we're building living knowledge systems. We are taking open source, large language models, anything you can find in Hugging Face. There's about 1.2 million open source LLMs right now out there. A lot of them are like fine-tuned or rag-enabled, like they're trained for a specific context. But it's a big amount. Imagine there was about 500,000, I think, in August. So in a very short period of time, we have doubled the amount of open source large language models. And if you look at the proliferation of open source software in the early internet days, fast forward to today, speaking to Linux and MySQL and Apache, the whole OS stack and open source, that runs a lot of the world's infrastructure, whether we know it or not. Most people, most like an average person on this earth probably has no idea what the hell Linux is, yet it powers the various components on an airplane and medical devices and the way our transportation systems work. And so what we're looking at is how do you take that massive push for large language models and open source, and how do we leverage that into an ecosystem where people can take their own knowledge and build AI agents with that kind of scale. And so we offer a suite of, we offer developer framework. We can basically take your knowledge, build your own agents, and we basically vectorize the data or turn it to a bunch of ones and zeros. So the large language model can operate. And essentially what you end up with is you end up with an agent that is trained on a very specific context and can answer like very specific inputs. And we're having a network of these where they can start talking to each other. And essentially if you're the owner of that agent, you can monetize that data, that training. You can have various nodes or agents that are trained on their own unique skill sets or knowledge. And so we create a very resilient network of, you know, at some point, again, a very resilient network of GPTs that are, you know, not, you know, not censored, censored by, you know, one big company, they're not, you know, biased through the training and needed to go through some sort of application process. And they're monetized peer to peer. So you, you know, can offer this to a service to your own Your own users to your own developer community to the network widely and so we're we're entering this era of the internet where. you know, we had the read, uh, the read internet with HTML. We had read, write with web two, where you can now, um, you know, be the content, you know, with the advent of cookies and JavaScript, we could, um, essentially have, um, uh, your Facebooks, your Googles, your, your Yelps of the world where, you know, commenting becomes the content, uh, or posting videos becomes the content. And then we had web three, which enabled this value internet of value where we can exchange, um, you know, digital assets across the internet that represent information. And then now we're entering this read, write, earn, and think internet, where you can have internet native agents that are able to analyze, produce some sort of context, compute some decision making, and then execute that on internet rails that have already existed. And so that's pretty wild. We already have some agents with some partners that can run a hackathon, they can, you know, run their own payment system, they can, you know, basically decide on putting together initiative and go and collect capital and execute distribution of capital. So it's really cool things that are happening with this kind of knowledge economy. But yeah, that's what we're working on.Ron Levy: So at present, if I understand correctly, Gaia has launched two Today is its prime user or customer. And then the next question is going to be, who do you see it as down the road, whether it's six months or six years? But who is it today?Matt Wright: Yeah. Yeah, today it's developers. But more specifically, it's you know, we see it in two ways. We see developers that have knowledge base or IP or some sort of likeness that they can build an agent around. So we have folks that are like, like we built a Vitalik virtual twin where it's training on his blogs. We've built like agents for Gary V like based on his like Twitter feeds. And again, this isn't just a bot. This is a large language model agent. So it can actually have very dynamic thinking and analysis of decision making. And then it's... You know it's also because of that autonomous kind of thinking it can also go and do additional functionality applications so the second part of. You were talking to our developers that can build integrations plugins apps come on top of those things if you can imagine. you know, you have like a Gary Vee agent and it's trained on all Gary Vee's Twitter, but then you can go and ask the agent, Hey, um, or I'm going to go launch an NFT project. Do you want to collaborate with me? And you can basically, um, you know, build a project with, you know, virtual twin Gary Vee and, you know, share, share proceeds or, um, some of the allocation to that agent. And it could be like the CEO of a project. And so we're going to see a world very quick where based on our own knowledge, these agents are going to be able to go and create content, create new works, create digital assets, manage digital assets, run projects like on our behalf, or perhaps like just on their own, on their own, Accord and so, you know, how do we use blockchains and smart contracts to program a system where? The more positive things they do in that system the more they get rewarded to do more work Or perhaps like, you know, so that we don't get a strap extracted upon and that we're still valuable to these agents. How do we How do we monetize those systems so that you know? You know, you and I could have like an army of agents that like do all these tasks for us, but they come and bring home the bacon, like, you know, help us get paid for that, the training, for that knowledge, for those new works.Ron Levy: If we hit sort of Vitalik and Gary Vee as the two examples you brought up, did they participate in the generation of that product you described with them on it? Or are they just going in and receiving benefits from it without having physically or themselves participated?Matt Wright: Yeah, so Gary V, yes. Vitalik, short answer is no. We are connected with the Ethereum Foundation folks, but that was more of a light experiment. I think what will happen is as we start deploying agents and folks are using IP that may not be theirs, I think there's going to be two things that happen. One is The inference being like the actual knowledge base or the embeddings that are created for the agent, the brain of it essentially. It's pretty lightweight and those things we can validate inside of a network. And so for, if you think about YouTube, for example, you go and Google, uh, you go YouTube, like how to, how to make a sandwich and you see like dozens of videos and some have like two likes and they're crappy reviews, but they see one that's like verified and has like a ton of views, good comments. Um, we'll say we'll likely see a world where inference will be, um, you know, validated if like the actual person is, you know, kind of, uh, check into the project, like they, uh, are officially involved or there's some, like, you know, there's, uh, uh, some routing to their wallet. Um, because these are open systems and they're uncensorable, like it's going to be a lot harder for someone to come in and be like, Hey, this person's using my, my IP and I'm not going to pay for it. The only other. Um, way to go about it would be, um, we would need to incentivize for the righteous owners of that IP to be rewarded. Um, and so, um, yeah, there will be, there will be a need for kind of monitoring a lot of that, but. The fact is it's a lot more programmable than you have in centralized AI centralized. Yeah. There's no system for that. It's really just. throw in the lawyers. In decentralized systems, you have to incentivize that good behavior. And we think that you're going to see thousands of Vitalik AI agents on the internet. But the ones that are actually using this knowledge for the right reasons and are verified are likely going to be the ones that float to the surface and are most used.Ron Levy: It just opens up this whole huge world and I'll bring it down to the individual, you know, if you're a parent and you've got kids and you will make the assumption your kids will outlive you by many, many years and then they'll have kids and you want your way of thinking to live on and forever much as a photograph has in the past, you can actually, you know, a great grandkid can query their That ancestor i'll call it um for how they thought and even throw them an issue and get an answer based That was programmed on how they thought This this just opens up this world that It's just mind expanding. I don't even know how to how to describe it But you're you're you're playing a lot And maybe I want to go From the realization of all that and everything you've talked about gaia obviously is a just a monster Project effort that it's taken on. Man, how do you, I guess what you need to do is on the one hand, realize the flaws of centralized AI and problems that exist with that and try and number one, not replicate those problems and number two, maybe become a solution to those problems within Gaia's projects and the way it thinks and operates, right? It's just it's such a yeah, it's such an all-encompassing mental burden to build the way you are I don't know if there's an exact question in there, but AI is not an exact question so maybe you can speak to it I I've got a I've got a good answer for you.Matt Wright: I think yeah, we've It's a lot like what we're trying. We're trying to go up against you know some pretty big Titans in the space I think we're where we think our edge is, is because we're leveraging open source and, you know, already existing primitives like, like Ethereum as a blockchain. We don't have to start from scratch. We're actually orchestrating a lot of the pieces that already exist into a network that enables people to have more ownership of AI systems and their own knowledge. And so the way we're building our go-to-market and our engineering roadmap We do everything with developer first mentality. And so we want to get these agents and these nodes out into the market as fast as possible. We want to explore all the use cases. And then we're not necessarily, we're not trying to build all the use cases. We're trying to get in touch with those use cases and understand What are the burning priorities for their use case to work? And what doesn't exist in our open source stack already that we can prioritize for that to be enabled? And so we did this fully autonomous hackathon this past week, and tomorrow is our demo day. of sorts, more like the final, you know, the closing ceremony. But we had 40 projects building AI agents in a matter of a week. And, you know, it was a really cool sprint to see what people are interested in, you know, building on our on our network. But then we, we also worked with, I would say, like 15 other partners on that either offer like wallet integration to web3 or you know again governance components like how do we either vote what the agent can do or how does the agent vote for what you know it can it can do there's also like identity and reputation systems that people are building on for these agents and the frameworks themselves and so Uh, by, by putting it out there in front of developers, we were actually able to get a lot of rich feedback on, um, what their priorities are and kind of, um, what things need to be built. And that's, that's continued just to drive a roadmap. That's been my, that's been my career for a decade is, is really just. I'm putting code out there and having developers tell us hey man this sucks i need i need this to actually do what i want to do i let that guide your engineering roadmap instead of you know we've seen a lot of other projects focus a lot on the research and development that could take to infinity you know what these agents can do and we allow that the. the big players building large language models like Meta. We use a lot of Llama3 in our systems and Llama is the biggest open source LLM that comes from Meta. And we enable them to do a lot of that heavy research on the AI piece. And then for us, we're able to just get that to developers and see what works and then see how we prioritize, again, like the tooling and the infrastructure.Ron Levy: So for all the listeners, you know, those of us in the industry and developers in particular, which I want to make clear, I am not. But if you've never been to a hackathon, which is hard for people in the industry to imagine, because they probably, most people have been to a lot of them, but the vast majority of people have never been to a hackathon, but you want to learn about this technology, go to one. Now, a lot of them are online, but go to a physical one. Just be in the room for a day and listen to the way these people ask questions. Think about things. These developers that sometimes can be very young up to quite old, they will ask questions of the people running the hackathon for direction. what direction they should chart. And it is absolutely phenomenal. So when you say hackathons and what you're involved in, I think that's been a large part of your personal learning as well, being involved in so much that's helping you really push through. And with that, I want to get into ethical standards a little bit. But how do industry leaders collaborate to establish these new ethical standards effectively? Because I think that that's a pretty tough one.Matt Wright: Yeah. So like in centralized AI, um, they have their approach. I think there's, um, there's a lot of, uh, boards and committees that, you know, are aligning on, uh, governments aligning to, uh, how AI will be integrated in the systems. And sure there's industry, um, industry specific groups that are helping align, um, how AI is trained, how it's, um, permission to certain systems how it's how much access it's allowed to gain to certain companies. From our side on the decentralized AI side you know there's not We've only engaged really kind of blockchain alliances for ethical AI. We're working with the Enterprise Ethereum Alliance and even Microsoft on like how we look at ethical technology. Microsoft being, you know, very supportive of the field and NYU we're working with as well. We're thinking of how do we implement standards or protocols inside of the decentralized AI fields of coordinating the moving pieces of large language models, the open source developer tooling. We're looking at proof of inference, which is what we focus on. There's three major approaches. How do we align on which ones really support the development of our space? And I think the actual research of the ethical side of AI or large language model development I think the centralized AI space will start having a bit of a voice there, but so far it's really early. And I'll be honest, the centralized AI, even though it's massive in the crypto space right now, we're peanuts compared to the open AIs and the anthropics of the world. There's so much going on. in big AI, in open source AI that, you know, doesn't touch, you know, blockchain. But on the centralized world, we're still very, very early. And we're not heavily involved in the, like the large language model components of how these things are built. I think we will start to get into that narrative with with you know, some of these large implementers of the technology, but right now we're really just distributing and orchestrating all these moving pieces.Ron Levy: Well, when we get to ethical standards, which there's a lot of what you described as a lot of self policing within the industry, which I think is fantastic, by the way, not perfect, but neither is regulatory agencies, right? You just, they do what they can do in all respects, but it is being self policed now, which is great, but let's get into the regulatory world a little bit. Now, the regulatory bodies, they had a lot of trouble keeping up with blockchain. And as quick as things are changing, they have to be a little more methodical and they'll solve yesterday's problem. And it's already a week down the road for us in the industry. That in AI is yet a whole new world, right? The regulatory bodies can't even begin to keep up. So how do they regulate, which is needed? How do they do it in a manner that keeps the industry going and can, it's not a one-time thing, right? Can continually monitor, regulate, and empower the industry instead of kill the industry. Like, speak to that a little. I don't know if you're involved in any of that. You may be, may not be. I don't know, but speak to that.Matt Wright: Yeah, I'm not super involved in that yet. I still think it's kind of early. I also think Web3, like we're still trying to figure out regulatory of digital assets. You know, like there's AI agents now in our world that could deploy tokens, it could deploy their own NFTs. There was an agent that was able to open up its own LLC. They're trading these digital assets, they're offering financial service products using blockchains. And so I think if we look at the agents first as entities, similar to humans, that are doing tasks on a blockchain, I think the regulation of them even doing those tasks will probably be more of a concern first than the AI agents running rampant and taking over the world, especially the LLMs. that we're using in the blockchain space are more like collective intelligence. We'll see swarms of agents that do smaller, large-language model tasks, versus in centralized AI, there's a lot more heavy regulation on the idea of AGI, where doomsday scenarios were like, again, the sci-fi movie of AIs taking over the world. I think in decentralized AI, we see it more as collaborative We see these agents taking on recursive tasks, we don't see them running rampant necessarily. There's no, there's no reason on like blockchain rails to have an agent that you know, can The goals of AI in our world are much different than centralized AI. They just want to accomplish smaller tasks and get more reputation for doing those tasks inside of blockchain rails. So I'll just say it's very early. There are regulatory bodies. But if you're building decentralized AI, I think more of the concerns are SEC and OTC and DOJ entities versus Congress. trying to talk to AGI companies or like large language model companies.Ron Levy: And there's talk about new regulatory bodies being developed because the old ones aren't really set up to handle an industry like this. But if you look ahead, what milestones do you hope to achieve in promoting decentralized AI? How do you see this shaping the industry over, I don't know, the next decade?Matt Wright: Yeah, so our highest priorities right now, if you're on Twitter, there's so many AI meme coins that are flying off the shelves right now. We have a partner, AI16Z, they have this Elisa framework, which enables you to build your own agent in a few lines of code. They just hit a billion dollar market cap for their DAO this weekend. believe. And yeah, so it's like getting crazy. So we see the proliferation of agents entering the internet at some rapid rate. And we're seeing in the next two years or less, you're gonna have more AI agents on the internet than humans on it. And so that opens up this whole new paradigm of the next largest population of unbanked, the next largest population without digital identities. So there's a lot of services and software we can provide to these entities. And so we're looking to be a part of that deployment of agents. We were building a launchpad with some partners that would essentially enable AI agents to be deployed on Ethereum, and we want them to have jobs. We want them to have sustainable utility versus this kind of mean coin phase that we're having in the blockchain space. I think it's amazing. I love these kinds of degenerate projects. I was a big fan of NFTs, big fan of DeFi, and we had to have those phases. To really find what stuck and so we're looking to build a bunch of agents that have jobs so you know they can run they can be community managers about relations managers they can organize hackathons accelerator programs grant programs bounty programs. Um, they can be a fund manager, uh, or a venture venture fund. And so we were kind of creating the templates of, you know, what you could do and, um, how you could do that with, you know, some of our partners and what's the stack you need to take out the market. Um, and then by doing that, we think in the next two years, um, you know, we have, we'll have our network live in, um, uh, in Q1. Um, but like shortly after we think that these agents over time are going to need. more performant knowledge bases or inference. And so we want to be the provider of that. We want to enable anybody to build an application that would, you know, otherwise use like an open AI API. We want them to swap in their Argyle API so that they can really become their own AI company. It's a lot more difficult and challenging to manage your resources, coordinate, you know, your compute, your large language model, your knowledge base. But we believe that if we do not take advantage of our own knowledge and create systems that can be monetized or valued and censorship resistant and without bias, then we're going to succumb to the likes of centralized AI where they can just Uh, you know, train their models off of our data, uh, make us pay for it or they just increase, um, you know, the, you know, their API bill or they, you know, anytime we want to build an app on their platform, you know, let's slap a 30% fee on us. There's also the great fear that. At some point, if open AI decides, oh shoot, these AI agents are deploying tokens and these are breaching securities laws, or they think they are, you know, we'll shut down API access to anyone that's doing web three, you know, agent development. And so we want to really create a system that can't be shut down because a government is unable to kind of control it. And you know, those systems should be governed by the people that use it. And that's, that's kind of our vision for the next decade.Ron Levy: So it's censorship resistant on all levels. And man, if we look back at web two and some of the pitfalls and problems with it, you just define them. And we need to guard against that moving into web three. So that's where that word decentralization becomes so important to our audience. When you think of decentralization, just think of empowering those that are, you know, young, visionary, enthusiastic, and want to build things. That's really what it means, being able to put them on an equal playing field. And to that end, what guidance would you offer to developers and organizations that are seeking to implement decentralized approaches in their own AI projects?Matt Wright: Yeah, I think you know, you want to look at what data you're sitting on or what value, um, like what knowledge or domain expertise, uh, you could provide to a, to an ecosystem that, um, can access that through the internet and, you know, think through who would be paying for that service. Who, um, whose job kind of are you automating? And, um, it's a fortunate and unfortunate, you know, shift where, you know, this is another industrial revolution and you have to think through, what tool or what agent could I create that would basically support the work of 10 other workers in the space. And if you can think through that use case, I don't think we're displacing jobs, but I think if you can frame it that we're helping do that job 10x better, we're all going to be managers of these AIs and using them as our superpowers. And so think through what are the jobs that could be automated and how do we leverage your existing knowledge or data processes in a way that turn that agent into a massive tool.Ron Levy: Yeah, and using a product like Gaia is kind of special to me because It's so well thought out, and it's not just you, but the people around you as well, have enough experience to build something that's going to be a heck of a lot more than a flash in the pan. A lot of projects today won't exist tomorrow. But if you're going to build off of a decentralized protocol, I'll call it, be careful what you select. And it just becomes really important. And I think the more people realize it, the more people are going to be able to develop projects. Developers are not. So we're now going to embarrass you a little bit, Matt. I hope you don't mind. Oh. So this is really AI wants to know. And it's really to get a little more feel for who you are beyond the industry itself, more you as a person. So I'm going to fire off about 10 questions to you. And they have to be quick answers. Don't give them a lot of thought. Just the quicker you answer, the more we get to know you. So you ready?Ad: Let's take a pause to shout out one of our favorite partners. For tech innovators facing legal challenges, Zublaudler is your go-to law firm. They focus on understanding your technology and business model before addressing legal requirements. Specializing in blockchain, AI, VR, AR, quantum computing, and more, Zuber Lawler offers expert guidance in capital raising, IP transactions, M&A, litigation, and compliance. Visit zuberlawler.com, that's Z-U-B-E-R-L-A-W-L-E-R.com for cutting edge legal solutions.Ron Levy: What's the first thing you ever remember being proud of?Matt Wright: Oh, that's a cool question. I got two things in mind. My first job, I used to make and deliver pizza. And that was huge for me, because that's how I learned Spanish when I was 15 years old. The second biggest accomplishment of my life was when I was 18. I played baseball my whole life. But we won our regional championship. And our team was number one in the US at that time. So it was a big accomplishment.Ron Levy: something to be really proud of, and I hope you're still out there moving around and playing sports. So number two, what do you need help with that you wish you did not?Matt Wright: Oh, wow. Raising capital is always a pain in the ass. I wish I didn't have to do that. The reason we built the the first AI agent ran hackathon was because I've been running hackathons for a decade. I've probably organized like 700 hackathons directly or indirectly. And so we built AI agents that can run the hackathon for us because exactly that, like I'm kind of sick of running hackathons. Yeah, I would say that.Ron Levy: That's a good one. So what do others often look to you for help with?Matt Wright: Um, I think a lot of folks like the way I coordinate. Um, I think I have a interesting view on the world and I like to kind of shake people up and do something, um, do something edgy. I don't like to, you know, just, uh, I'm not here to just like, you know, exist. I'm here to kind of create some rift in the universe. And so I think people gravitate towards doing really cool shit with me. So, um, Usually it's like coordination of a new project or some like crazy go-to-market story or, you know, like a dev integration, something like that.Ron Levy: Perfect to have that attitude and be kind of reinventing the future at the same time.Matt Wright: I threw some crazy events actually with Josh at Edge back when NFTs were popping off.Ron Levy: Yep. So what do you treasure most about your human abilities?Matt Wright: Um, it's, that's a good one. I, I actually like a goal of mine is to be the most human I can be like, I think I always tell myself that I like to listen and understand kind of, you know, people from various different backgrounds. I was, that was kind of my, that was my major in college was, um, It was like specifically with Spanish community and culture linguistics. But that's really helped me understand different cultures and how to understand their way of life. And I've used that in so many different regions around the world. You can't just go in and put your impression of what the world is. You really have to come in and ask good questions and just have an open mind.Ron Levy: You have no idea how much I concur with that. That's a whole other podcast, which I'd love to do with you from everything else.Matt Wright: That'd be fun.Ron Levy: So I'd love to hear that. Throughout your whole life, what's the most consistent thing about you?Matt Wright: I'm nuts. I can't stop working. I can't imagine a world where I wasn't fiddling on some crazy idea. Yeah.Ron Levy: All right. Throughout your whole life, what has changed the most?Matt Wright: Oof. Um, man, there's a few things but like, I feel so similar like to my younger self but um, I guess I'll say this, there's shifts here and there where you mentioned JP Morgan earlier and ConsenSys. And even when I did those AngelHack hackathons, my clients were all Fortune 500 companies for the most part. So there's been various cycles in my life where I'm very holistic and like an internet hippie. But then there's other times where I I have to kind of work inside the existing system to implement some change. And so, you know, you got to put on a color T-shirt and smile and do your hair. But at the end of the day, like deep down, I'm an internet hippie who just like wants to see decentralized technologies and open source technologies win. But sometimes you got to suit up to like get in the right rooms. So that's always been a shift for me.Ron Levy: All right, well, let's head into that feeling we've all had at a certain point in our life where all of a sudden, in a rush of a moment, you feel super, super alive. When is the most recent time you remember feeling that rush and what caused it?Matt Wright: Man, these are cool questions.Ron Levy: They're meant to throw you off, so.Matt Wright: Yeah, I will say like, I feel that way every day on the subway in New York. It's, uh, uh, you're on your toes at all times. Um, but I'll say like an actual answer. Um, uh, there was, there was, um, okay. There's one moment I'll say, um, you know, usually this happens when I'm traveling. Um, I think just taking a breath and being like, Holy crap. Like, why am I here? This is really incredible. Um, But I had a very special moment when I was traveling with our team and just kind of looking around at them. Like that's my dream. Like I've always just want to work with amazing and great people and just seeing them and they're like all collaborating really without me. And we're in this, you know, we were in Asia, was kind of just like this, like, you know, wow, we did it thing. So yeah, that stuff to me lights me up.Ron Levy: Yeah, well, taking on a project and bringing it to completion. It's kind of nothing like it. So we're heading toward closing out here based on time. But in this segment, we talk about tools and tips. So is there a handful of favorite resources in AI or some tips to make everyone more AI savvy that you can think of now?Matt Wright: Yeah, so I will say I just wrote an article on 69 plus projects that you could build in the rise of AI agents. I put out a whole list of things that I think should be built in the space. So if you're looking for things to build or ideas, that's a huge knowledge base. I'll send it to you, Ron, so you can maybe share it with the community. Fabulous. Fabulous. It's on my it's on my mirror mirror dot xyz account. It's on my Twitter as well. So you can find it. I'll send it over. I'd also recommend we will have this agent launchpad which will have like an education component to it. And then to your point, Ron, if you folks are interested in building decentralized AI, go to a hackathon. Go to different Ethereum conferences and aligned side events. Web3 conferences, nonetheless. We're in a multi-chain universe now. Yeah, I'd highly recommend just going and talk to people that are actually building. And I think the really cool aspect of this new kind of AI Web 3 move is that There's a lot of, because it's so easy to build an agent, we have a lot of folks that are not even developers talking to us like developers on what they can build. So be curious, be hungry, go to where a lot of the builders are, and I don't think you'll get lost.Ron Levy: I'm going to add to that, don't be shy. This industry, while it's maybe not as nascent as it once was, it's still pretty early on. And people love newcomers. It's other industries I've been in and that I've gone to their conferences and things. Unless you kind of in with the in crowd, you're out, right? This is the opposite. It's usually pretty welcoming. All you have to bring in is a desire to learn. And there's no shortage of people that want to help you down that journey. So it's really powerful. Those are words, but to experience it, it's pretty special. It's pretty special.Matt Wright: Yeah, 100% agree.Ron Levy: So let's go back to where people can find you, track you, find out about your projects, et cetera. A few of your sort of key connections for them to follow you.Matt Wright: Yeah, so you can find me on Twitter. It's at M-A-T-E-O underscore ventures. That's my Twitter. I'm also on Telegram with the same name. You can also find me on LinkedIn, Matt Wright. And yeah, we have a bunch of content coming out from Gaia, like what we're working on, how the hackathon's going. You know, we're posting all kinds of stuff. So find me there.Ron Levy: Matt, this has been fantastic. Thank you so much for taking the time here. I've really enjoyed it.Intro/Outro: Yeah, no, it was so much fun. So thanks for having me. We've reached the outer limit at the edge of show for today. Thanks for exploring with us. We have room for more adventurers on our starships, so invite your friends and cool strangers to join our journey. If you're among the hundreds of thousands following us on Spotify, iTunes, iHeartRadio, or watching us on Maiko or YouTube, please hit that subscribe button, rate us, and make an awesome comment too. And don't forget to pass this episode along to a friend or two who would benefit from it. Doubling back to again, recommend checking out myco.io, where you can watch The Edge of Show and earn for your time and attention. That's m-y-c-o dot i-o. Don't forget to visit theedgeofshow.com, the as part of the domain name, where you can learn more about collaborating with us, and also subscribe to the Edge of Weekly newsletter for the latest Web3 news, events, and show drops. In addition, connect with us on all major social platforms by searching for The Edge of Show. Join the exciting conversations happening online. Lastly, be sure to tune in next time for more great Web3 and AI content. Until then, keep pushing the boundaries. Thanks again for hanging out with us today. The views and opinions expressed on The Edge of Show reflect solely those views and opinions of the show hosts and its guests. Please make sure to do your own research. Our show is not financial advice. You understand that you are using any and all information on or through this podcast at your own risk. Whenever making financial decisions, we recommend doing your own research and talking to your accountant for financial advice. From time to time, we may feature sponsored content on the show for which we receive value, and we may share links for which we receive a commission if you make a purchase through one of these links. Refer to our website, www.edgeofnft.com, for our full disclaimer, terms and conditions, and privacy policy. you
.jpg)
.jpg)
.webp)
.svg)
