REBROADCAST AI Livestream- The Rise of AI Agents: Identity, Autonomy & the New Digital Life

Unknown: Welcome back everybody. You have tuned in to the Artificial Intelligence Live stream. This is the stream where we talk about

Jon Radoff: AI models, applications of models, AI agents, all of the cool stuff you hear everywhere, but it's a focus on people that are actually building things. So I try to bring on entrepreneurs, CEOs, people that are actually shaping the future of this industry by either creating the AI models that we use or building them right into applications. So today I've got a really cool conversation with Mike and he is from a company called Yumio, I hope I pronounced that right, Yumio, which is building all kinds of stuff for an avatars and AI agents and the like. So this is your opportunity by the way to take part in this conversation not to just listen to it, but to be part of this conversation. We really pride ourselves on how engaged this community is. I love seeing you all the time. I already see the little hello messages coming in here. We'd love you to ask a question. We'd love you to post comments whether you're on YouTube, LinkedIn, Facebook, X, Twitter. We're broadcasting to all of these places and you can be part of it just by posting in there. If you want to join with your camera, we're crazy enough to make that happen too. So Oscar is going to be standing by. He's going to be watching for those of you who have comments and questions will field those up onto the screen. But if you're one of those people who wants to jump in with a camera, those happen once in a while, he'll send you the Streamyard link. And if you've got something pertinent to share for artificial intelligence, we'll get you right in here. So without further ado, I'd like to introduce you to Mike and Mike, let's start a little bit with just your personal journey. What have you done before and what got you to this point in time where you're building stuff with games and AI avatars agents, all of that.

Guest: Cool. Yeah, well, great to be here. Thanks for having me on, John Pleasure. And yeah, my kind of I suppose I can begin my journey with I was a video game producer working on Nintendo. Worked on some kind of cool games and cool IP. It was during the Wii U phase of Nintendo. So maybe not its most popular phase. But it was a cool cool kind of time to come up in the games industry. I moved across to Sega, a studio called sports interactive, who were known for again called football manager.

Unknown: It's a simulation game. Still a huge fan of that game and played all the time. And I'm still very close to team amazing team actually. And so yeah, I produced games there, focusing on MMOs and also some mobile games.

Guest: And I then started this company with my friend from Dublin in Ireland, where I'm from. His name is Ben Cleary. So he is from the film world actually. He won an Academy Award for Best Live Action Short in 2017. And more recently made a kind of a near sci-fi film called swansong, which is a really nice film. It's on Apple plus. Yeah, so we started this kind of journey together and our focus was to look at how we could combine these different emerging technologies to build characters and to build a supposed compelling relationships between technology and humanity. And that's something that we're still very much focused on.

Jon Radoff: And technology and humanity. Okay, well, we're going to we're going to definitely dive into that. I'm especially interested in that interface between artificial intelligence and AI entities. Probably can't quite call them sentient yet, but maybe someday and how that's going to interact with with us humans. So that's what you're building, though, right? Like you're you're building an avatar system if I understand I'll let you describe it in your own words. What does you me do?

Guest: Yeah, so we're an ecosystem for people to be able to craft the future of on chain AI. We want people to build spoke on chain agents. And we want to liberate those agents to be able to live across digital culture. So our own platform, but also across social media. And also into video games like the video game that we're building as well. So that's kind of what you me. So the agents that people create with you me. They've got they've got like their personality and their backstory. So it's like who they are, whether from what their motivations are. If they're 3D avatar, which enables the diversity is different worlds or sonified. And they've also got what we call skills. So these are their agent decadent abilities. It means that they're able to perform complicated actions inside of complicated space. So they're able to do things for you and they're able to do things for themselves.

Jon Radoff: Very interesting. I think we have some video footage right Oscar that we can share to show off a little bit of this stuff because I think it's fun to look at. And then I have a million questions. And I have also just a lot of curiosity around how game developers might work with this too, because that's the market we're in with the immobile. I'd ask her. I see you put absolutely.

Guest: And AI agent creator with a difference. What's up friend? Holy shit, you're talking. Yeah, and now you've unlocked 3D Explorer. I guess we should jump in then. You know it. Let's dice. Damn, that's wild. Unleash from browser land into this mad 3D world. You just created life. You're like a god. It's pretty trippy here. Let's run a community poll friends. Which one of these nutzo agents should I chat with next? Okay, let's go. Man, I need some new threads. Want to dig me some out? Sure thing friend. Nice collection. Fingers, right? Yeah, those kicks are savage. You just leveled up and we unlocked online gamer. Amazing.

Unknown: Gotta be a brawler. Get some. Upper cut sidekick. Boom. Adios scumbag. A little rotation for the final touch. Nice. Looking dope. Let's get back to the browser. We're getting some love. The future of AI will be built by you. Create groundbreaking agents and shape the world. Wow, this is wild.

Jon Radoff: Alright, Mike. So the avatar creation aspects pretty clear from that, but can't miss the little chart on the token that that character is creating in the process. It's I'm getting meme coin vibes almost. Am I picking up on some aspect of this?

Guest: And we probably differentiate ourselves have leave from the end coins in a sense that like mean coins by almost by definition, and to any utility and where is and for us, we'd like the concept of people building tokens and launching tokens inside of our ecosystem that people want to use for a number of different reasons. So one of the things that we focus on is in viewing our platform with the type of utility data, actually, yeah, clearly differentiate in the way from stuff like that. And yeah, there's a world further mean coin stuff, but it's not really the world that we're interested in. What we're interested in is enabling powerful AI through the prism of personal characters.

Jon Radoff: Yeah, well, I mean, clearly there's it's completely 180 degrees different than like the casino aspect of just like minting a coin and getting to see if people take it up or not, but I was noticing though this sort of individual ability to mint a currency basically for your avatars or the characters that you created and presumably that currency is driven by what like why would some what what's going to produce the buy pressure on that currency. That makes us get that green candle chart that we saw in the character that you meant it.

Guest: Not yet. Yeah, well, I was supposed to like to go into some of the details on some of our our on the kind of the token, token on my aspects, but some of its publicity inside of a white paper cell and probably shouldn't delve straight into it for a number of reasons, probably most of which you know yourself. And, but, but what I can say is that we're very focused on ensuring that when people use our platform to you interface with agents that are user created right we want users coming and creating really cool agents and then we want other people coming in and essentially using those agents and paying to use those agents and we want to ensure that the creator is not going to be able to use that. I want to ensure that the creators that exist inside of our platform and find the type of renumeration for the usage, you know, so that's where a kind of a token on the structures, particularly useful set of smart contracts that can kind of fairly create and distribute revenue is one of the best things that block you.

Jon Radoff: Well, it sounds like that white paper is coming soon, so maybe some of you in the audience got just a little bit of alpha there to keep an eye out on you me, probably you should follow all of the social accounts, which will post on the feed here as well so that people can find you if they don't already know about you and read that white paper when it comes out, but can we talk a little bit more about like how did the avatars get used because I we saw from the demo like they can talk you can chat with them. How do you see them being used by the community like other than the creator and your own individualized fun with that character what's going to draw in individuals or even applications and his applications a target for this like could I build a game and then consume an avatar and use it within my game, for example.

Guest: Yeah, yeah, so you're kind of you're right on track there so how people use the agents, the agents themselves are multi platform so people will come to you me out to build the agents, but the agents themselves live and exist in multiple different places, even at the same time. So of course we love people coming to our platform and using the agents and hopefully our platform has a lot of stickiness and you know all the KPIs are great like retention and stuff like that to ensure that people want to be on our platform and use the agents on our platform, but. What we're also creating is a structure by by which the creators of agents inside of the new me ecosystem are truly interoperable are able to navigate to other places, for example, and the one that probably won't be groundbreaking to to any of the viewers will be we want the agents to be able to exist across social platforms. That's something that we've seen already we've seen a lot of agents existing across social platforms. I think one of the things that we're going to see across the board probably in in broader culture is we're going to start seeing agents existing on websites. I think that you know the you know the Mr. Clippy example that we used to have for for Microsoft office or something like that and we're going to see more advanced versions of them in websites, you know. I know I know even that you know i'm glad close with some DCs who've told me that they're receiving decks that have links to chat like a GPT which is essentially a conversational agent so the investor could learn about the product that they're investing in so. I think we're going to see a lot of this across the board and our agents are able to exist in all of these multiple different places across you know the broader worldwide web. But the stuff that's pretty cool that we're really passionate beyond that what actually got us into this john to be honest with you from the start was that you know we came from games we were building agents to occupy 3D spaces. So those are the cars that you're seeing there genuine 3D avatars that are able to go into unreal engine or unity based experiences so it becomes a you could call it like an infrastructural tool for people to be able to build character driven experiences. Inside of interactive entertainment.

Jon Radoff: With the social integration and my understanding correctly that the that the agents communicate through the social channels that they're integrated with so you've got kind of built in. Social propagation mechanism with the platform as well they're given that everyone's going to want to promote their agents that they spawn on line exactly.

Guest: Yes that's right.

Jon Radoff: I for whatever reason the last week and i'm just going to continue the trend but in this last week I keep getting into all these conversations about dead internet theory. And i'm sure you've heard about it this theory which is like we're getting to the point where the internet is just going to be populated by bots the humans are going to be gone. I have a very specific take on this theory but i'm curious what you've heard of this what what do you think it seems like you're going to help. You're going to help us get to this future where it's all bots right.

Guest: Yeah probably i'm and i think well i think that we're at a we're at an inflection point where our systems. The whole nature of our digital culture is changing because of AI many of the systems that we've built before will have to adapt or die in order to be able to exist within the AI huge you know. I think that's a natural progression of technology broadly I don't think this is part of AI it's you know. When we see huge massive scale technological innovation it does mean that there's going to be huge disruption as well and so I think a lot of the systems that we built need to. Sometimes we need to get rid of them or sometimes we need to adjust them in order to be able to. Kind of advance into the next phase of what you could call like the agentic era I know that's a term that people bounce around and. So yeah I think that does create a lot of opportunities right it's quite a lot of really cool things but you're right it creates some challenges as well. Creates a few bummers man it's it's going to of course like I know people don't like to talk about the some of the negative things about inside of AI do I i'm okay to do it like it's. It is going to disrupt and subvert those of. Things and some of them are going to be sad you know we're going to lose some things that we really like you know we're going to gain lots as well so I just want to point that out and. In regards to kind of the amount of things that are how yeah like the internet is going to change dramatically social media is already started changing dramatically I think you may. If you look at the comments section these days you don't know what's real and what's not with the amount of kind of bots that are in there so. And. This is an area actually kind of. To as a potential segment that actually crypto and blockchain does offer some really interesting solutions it offers the ability for transparency the ability for to be able to prove things and really kind of unique ways so. You know a lot of people have been looking for kind of convergence between AI and crypto for the last four years you know crypto has desperately being trying to find like how does AI matter for crypto. And I think that that's it it's that the world's changing and we're going to need to be able to prove certain things I think that there's only one really really. Clear on a saleable way to do that that's blockchain tick.

Jon Radoff: Okay, all right well we're going to come back to all the above and I'm even going to try to answer my own question with my own version of that before we get to that number one. A bunch of you are posting in the comments hey what's the code for today so Mike we've been doing these codes we have a leaderboard that people can compete on for the beamable network leaderboard so we're doing our own token launch a little bit down the road. And we've this humongous community for it now over a million people millions like even after removing all the bots speaking of my bots we've still got like over a million people in this community. Yeah, I did. They're all they're all looking for the code to rise up the leaderboard a little bit more and claim a space in the air drop. I promise we'll give it out in a little bit i'm not going to wait until the very end but if you've shown up a little bit late wish you would have been here from the beginning but you haven't missed anything yet the code is still coming. Okay, so kind of pertinent to the topic that I was just on though so as we have AI agents one of the things I'm interested in is sort of the. The competition I guess would be the right word between the centralized approaches to AI and this will lead to the blockchain stuff I think as well, but. centralized AI is like open AI and. Ampropic and all these places that are essentially API gated where I can pay a service fee to that central company get access to an API and then access it versus decentralized AI. Not yet talking about blockchain i'm really just talking about AI you can run on your own device so models that have been trained and released their open weight models that I could download to my own machine granted. These are not as powerful as the frontier models that someone like an open AI is running but they're pretty amazingly capable. At the point that they are now after really a pretty short period of time of development back when these things first launched with chat cheaply people thought. Oh my god you'll never have this kind of power in your own personal computer now of course the models that you can run on your own computer are a lot more powerful. Than even chatching PT was initially over although now we have models like oh three and whatnot that are extremely impressive and you know not going to eat and buy what you can run on your own device but what what's your thoughts on this market of. AI that you can run on your own device versus through something that's API gated in the cloud how do you think about that as a developer. Of products that are going to presumably consume AI models that you want to be able to use thoughts yeah yeah great great question i'm.

Guest: We work with both so and we've worked across both as well so a lot of self hosting of that alams and we. We were kind of early movers inside of certain sort of crypto AI but we're relatively speaking in AI we've been. Office for longer than most and we've been working with like a lot of the big company some of which you mentioned there and you know we used to find you around. Alan's a lot in order to try to get the best out of the characters that we were looking to create the most amount of flexibility. You know some characters we couldn't get the type of personality we wanted or they were too politically correct or you know they weren't swear we couldn't create villains there was all sorts of problems you know. That have a lot of which i've been kind of solved I would say so. The rate of. Change inside of the LLM world is. i've worked across like technology for a while now you know I know you have to and it's it's amazing to witness when an industry is just like sprinting right. i've never seen a spring like what's happening in an hour in within the LLM world like it feels like every week I look at something i'm just like Jesus I felt like I can't believe this you know. i'm not going to slow down it just seems to be speeding up and. Our approach is to build flexible systems to be able to adapt to that type of change so and this took us a while to realize by the way initially when we were doing this we were kind of we didn't really fully understand that statement like how can we be flexible how can we. Feeding the proof what we're building when things are changing so fast. And we made some mistakes in the way you know to try to try to get there but our approach is to be kind of agnostic to LLM's to build systems that can. Essentially be modular you we can modularly bring in LLM's from apis or from self hosting sometimes self hosting could be really efficient for costs for example right. You know you could self host your voice model that could be a good way for you to keep your costs down. But maybe you're not able to get the same type of. Spongebob that you might get with like point more or four or you know. So so there's kind of like pros and cons our approach is just to be flexible as much as we possibly can and and and. don't try to predict too much if if that makes sense you know because every time you try to predict it it's like I just get wowed and and like no one so deep seek coming in January for example that's kind of like what hold on how did that how did that one happen I mean all of Wall Street didn't like crashed basically that. James look what along like an open source Chinese model that's as good as oh one what I was that happened we spent like hundreds of billions on this so. Yeah so yeah that's the kind of way that we look to mitigate risk in this like fast based kind of mental adventure it's um is to focus on what we what we do well. And sometimes that means you have to wait for some of the other technology to come for example like voice models we don't invest huge amounts of time. That voice models at the moment they're really expensive there's some really good players out there like a leather knobs this open source solutions access me. But we think that there's going to be really significant change there over the next year and we'll see like massive kind of cost reductions like we're kind of seeing across the board inside of a.

Jon Radoff: I remember when a lot of these models started surfacing and showing how powerful they were and then developers started consuming L I'm interfaces word for you referring to some of the other models things like generative art models or generative audio models like 11 labs you're referring to what there was sort of this initial take on the market which was all the value was going to be in the model creators. And not so much in the applications. But it strikes me that applications actually have a huge advantage that the models don't just from a business standpoint and we all love the fact that people are willing to send hundreds of millions of billions of dollars into these models so guys if you're doing that keep doing it we love all of you. But from an application developer standpoint the ability to essentially have your application get more and more powerful with you. How do anything other than just use the latest version or even switch to whoever's better that's kind of what you're talking about like it's it's such a competitive market where someone is constantly leapfrogging and it actually benefits all of us because if you're building stuff you just keep getting better at the same time.

Guest: 100% I echo what you said totally I mean when deep seek came out and people saw this as like this really negative moment for the AI industry I just couldn't I couldn't understand what I was like hold on so there's a free open source model that like the whole of humanity can use. And by self hosting so that opens up like just millions of developers being able to use these really smart model so absolutely that's exactly what's happening and kind of one of the things that you kind of better pretty much is that you know I think it's a reasonable bet is that it's going to get way way way way better and it's going to get way way better fast as well. So we don't really like as much as we all have to ruin businesses think about the now of everything and you know we work in crypto so the now is very important you know people have a short attention spans they want to see specific things from their from their investments and so on. And but one of the things that's very important I think building an AI is to also just recognize that hold on the future is right around the corner and it is it's it's going to have a very significant change so. So yeah that's why we try to kind of build the way we do we're kind of like well hold on what is the LLM what's what's what's what's chat to be five going to be like you know what's what what will be expecting from that what could we expect from six. And that has been a strategy which has has worked well for us since we got into the AI game you know and again really hard to predict some of that stuff and now crystal balls when it comes to those things but if we if you stay really close to it stay really close to the world. And so we're going to have a lot of research as well and it's very helpful it can be I think it can provide some long term viability because you see that the challenges that a lot of folks jump into AI they they really see it as like okay here's the technology now.

Unknown: But you know some of those businesses come in you can call them like GPT wrappers or sometimes people use that in a very pejorative way which I don't agree with but sometimes they they can be swallowed up fast by the LLM all of a sudden the LLM improves and it's swallowed up a part of the application.

Guest: So you know so yeah trying to say once step ahead is pretty important.

Jon Radoff: So a little while ago I asked you about device based AI versus the centralized AI and you kind of moved it to well this is a very competitive market and the fact of all this competition and different options for AI interfaces is what actually arms you with a lot of options and you can stay competitive. And I brought up this this notion of the dead internet theory so I just got to go back to it for a little bit so sure I think this whole dead internet theory is wrong and fundamental ways because it kind of assumes that we as humans will no longer use the internet that will just bots I don't think that's possibly sustainable. There's another trend that actually concerns me a lot which is just the general inshitification of everything on the internet. You see that in places like Facebook and social media where it just becomes increasingly extractive of the value creation and one of the manifestations of that is the tendency to silo things like communication. So I can do DMs on X but they can't really go beyond X I can go to discord discord doesn't talk to X Facebook messenger doesn't talk to all of the above. I think I have a dozen messaging platforms that I interact with on a regular basis slack linked in like this list just goes on telegram let's just goes on and on and on and a I can't keep up with all of these platforms. I can't even even if I choose one like telegram I have so many messages waiting for me to read I feel bad for people because I'm never going to get to a lot of these messages. I guess be persistent folks I'll if you if you rise to the top I might notice you but it's just hard for me to keep up with it. What is the answer to this it feels like a agents actually are the answer because I don't want to use the this isn't the way I want to use the app the internet I don't want to use a dozen plus applications that just do the same thing and unlike email where it was designed from the ground up to be decentralized and communicate and totally interoperable. Capital forces have kind of shifted us into all these individual silos which are just really annoying and their ad monetized in some cases and these entities have a business reason to separate us so I want an agent that I can just unleash on my computer and it just gathers all this stuff I tell it what to do it looks at all my messaging systems it surfaces it it tells me what's important it finds the redundancies I can tell it what I want to reply to. And it just goes back and it kind of controls everything I feel like that's a killer app like if you're an AI creator out there and you want to create an agent that I will buy and I think of like hundreds of millions of other people would buy. Just saw that one problem my messaging platforms and bring them under one agent to control now does that mean that in the future we're going to have a zillion agents online. Yes, it does but for me that's going to make the internet more alive that's the opposite of dead it's like now I can actually start to use the internet more regularly now our interactions on the internet may become increasingly. Intermediated by our agents communicating with each other but that's great let all our agents go out and mine there are various forms of intelligence and computation and report back after interacting with everybody I think that would be great. Give me that over the end rising model of the internet any day.

Guest: I think that's exactly where we're heading so I think they're going to be happy. That's my piece of sonar you building this Mike yeah well that is a significant part of what we're building for sure so. I mean it's kind of fun to think about it I know it's very recent but even to think about what had a history of usage of AI has been out of the last couple of years right so. When GP chat you be tea was kind of released most people saw people used to say it was like it was going to kill Google it was a search replacement it was a way to impart knowledge or it was like kind of a very basic assistant that you would use all right. The next way with users then started to see differently I'm currently the biggest use case the moment is that people are using it like a. A dedicated personal assistant or therapist so people are becoming extremely close to their AI and using it to help them coach their way through life and I think that that's not something that's probably going to go away and for folks of movies that they'll probably. Kind of reference something like Mike Jennings is her I think that that's a very kind of standard normal vision of what is coming in the very near tomorrow let's say. But interestingly actually a new generation of AI users is making are making them those to parent. Those are the folks that are very interested in using agents as these kind of multifaceted protocols you know they're even using mcps mobile context protocols in order to be able to get agents to operate at all corners of the internet and.

Unknown: But in that sense they're using AI and agents more like an operating system and I think what the cool thing about that is that it will enable a lot of folks like yourself to be able to.

Guest: Delegate a lot of the tasks that you really want to them but they're distracting me from stuff that you want to do more I think that you're going to be able to impart a lot of those tasks to. Two things like mcp's and you're going to be able to focus on the things that you want I think that that's probably one of the best things that AI can offer it's that it's going to give everyone that you. And that new executioner executioner ability to be able to make the kind of stuff that they want.

Jon Radoff: So Mike I want to talk more about the blockchain intersection is you see it because you made a powerful statement earlier that you that you really saw the future of AI and agent to AI in particular intertwined with blockchain and I'd love to hear your rationale for that I have some similar thoughts but I may have different reasons i'll let you. Talk about yours today first since this is your platform today before we get to you while you think about that it has come time to let our audience claim their code on on beamables leaderboard so we have just over a thousand people watching live so. So so delighted that all of you are tuning in today and watching this program learning about AI learning about blockchains intersection with it agent to AI LMS really fun conversation here with Mike but now is the time for you to enter that code so the code today is you me just like our guest here so that's why oh you m I. I know a lot of you do not speak English as a first language it's a little bit hard for you to hear so the place to look is at the bottom of the screen you'll see the spelling the uppercase why letter why oh you m I. I just like Mike's company that's what you enter to get that just go to our discord you type slash claim and then you me that's how you redeem the code every week there's someone who's like what do I do how do I get the code okay you that's probably because you're here for the first time. You might not be able to register for a discord quite in time to do the slash claim you've got to get in there you got to validate your account all that stuff be ready next time because we're given these codes out every single episode and it's one of the best ways to go up our leaderboard okay back to you Mike so I was asking my question about that blockchain web free AI agents intersection let's hear your thesis for this.

Unknown: Well my thesis is that agents are the de facto interface layer of AI agents are how we're going to be using AI it's going to be like the front end in some ways by which we access a lot of this powerful AI and the stuff that we're going to get agents to do is profoundly important to our lives.

Guest: I mean I just mentioned stuff like parapests there okay that's important. Well the people have in their lives but there's going to be tons of stuff it will have agents connected to bank accounts to crypto wallets to all manner of the most personal and important aspects of our lives and I think that's a lot of responsibility to give agents and without a public infrastructure essentially. These agents ultimately will be owned and tracked they'll be unverifiable they'll be. Totally open to the winds of centralized actors and we've seen some of the results of this type of. System being built with social media so I think it's very important that we engineer the systems of the future the right way in order to be able to get the best from the technology and align us within what users what are good for users and arguably what's good for humanity so as we answer these decisions to these agents I think we need we need agents that we can trust. I believe that went three blockchain technology is the opportunity to do exactly that so we can provide an open infrastructure where agents are able to act transparent transparently they're able to.

Unknown: They're having an identity of their identity their memory their provenance on chain you know actually being able to prove that they did what you asked them to do privacy controls I think putting these into the hands of users and having them owned by those users I think is.

Guest: I prefer that kind of world for AI I prefer yeah I prefer to build a kind of world whereby. The truity is decentralized and truly is in control of humanity.

Jon Radoff: It seems like in web three. With agents this is really the moment that composability has been waiting for some people talked about composability in web three for a few years now and for those who are. Hearing it for the first time or maybe I've heard it before but you're not exactly sure what it means it really just means being able to build software on top of other components but with blockchain you're referring to being able to do that at the smart contract level which means there can be value exchange. Within the protocol so that means protocols can talk to each other. Instead of having to be gated through centralized APIs where you have to sign up for things in advance and swipe your credit card now you can do a blockchain based value exchange so services can talk to each other and exchange value so in example with an AI would be your AI agent could be granted some currency to spend and it could actually go on and interact with other web three protocols. Which could be anything it could be everything from infrastructure provisioning to another AI agent to playing a game to I don't know play poker for you it's whatever you want it'll go and. Access to that currency though and being able to use that currency and spend that currency with other applications seems like that is a massive unlock in terms of software infrastructure once we get to the point of lots of software modules actually able to exchange value with each other by exchange value I really just mean be able to pass money to each other just like in the real world I could walk up to someone. And give them a $20 bill and get a service in exchange it's kind of the only way to exchange value that isn't intermediate by some government agency or an institution that wants to extract a rent from the exchange like swiping my credit card blockchain allows you to do that on a direct basis with individuals as well agents will now be able to do that. Thoughts on the composability dimension of this.

Guest: I think it's right I think it's important the frictionless transfer of goods and services is enabled by by crypto all be it could be a little bit more frictionless sometimes it's getting a lot better. But you know it's basically we've got this huge shared ledger you know for all of applications to work with so when you put that into the prison of video games it means that you can have agents existing or character existing inside of different games but they're interacting across different economy. You know so that's particularly important for what we're building actually because we're building a structure by which these agents are truly interoperable are able to go around into different video games so the game that you're building could have one of my agents in us and that same agent might appear inside of them when the Oscars building as well you know and they could trade. If the goods that exist inside of your game are on jane those could technically be traded between your game and an Oscars game via the agent so we call this a to a like agent to agent braid and this is going to be coming to the important there's. Another thing when we speak about crypto for this crypto is actually the most exciting area to be in this this is like it's very hard to set up agents at the moment that are able to trade on your behalf using bank accounts and the legislative and headaches to international had except you would be dealing with to try to do that would make it very difficult to test that type of thesis and where it's actually. It's not very hard to hook up a wallet to an agent and to get that agent to be able to perform actions on your behalf so it really is that the cutting edge of the particular thing and so the composability angle and the interoperability angle for us is is really really important and it's one of the areas that it's actually one of the areas that we always know that's about crypto broadly like when we wrote our first one page or selling Oscar before we came on air like when we've wrote our first one page or it was actually quite similar. In vision to what we're building that and you know we needed that AI was going to become. personable and we wanted to work with that and we also felt that we wanted to see an internet where people had more control and I think that we're still you know we're still we still have a good way to go there but hopefully we're all pushing in the right direction.

Jon Radoff: Mike there's kind of an adjacent question from our audience here you kind of brought up sort of the regulatory issues that would normally be around things like a trading bot for example but when you have blockchain technology sort of like just opens it up for anybody to jump in and build some kind of agentic system that interacts with that. What taric from linked in is saying is should AI agents have legal or ethical identities for example if an AI commits harm who's accountable the developer the user or the AI itself Terry because actually bringing up AI personhood I guess Oscar by the way the audience has spoken they're demanding that you weigh in and say something on. At some point in this conversation so I know that you have some thoughts on all things philosophy so we're going to invite you to do many thoughts and maybe I will share the. Is the galaxy brain holding all of this together like literally like if we didn't have Oscar. It would just disintegrate to its component parts so he is focusing his neural energies on just holding the space for us and making sure the camera points in the right direction and. Yeah their question surfaced but Mike we'll start with you here what's what's your thoughts on this question of identity and culpability I guess is what Toreka is bringing up.

Guest: It's a good question I think I'm not sure about legislating the real that's the first thing I'll say some I'm not a lawyer so I should be careful about what I say on that kind of stuff but what I can say and I think that it really depends on the situation right it really really does depend on the situation I think if. The LLAMs that have been built so far have gone through really rigorous ethical shacks and stuff like that's a lot of the time now of course you're up to do all sorts of stuff if you have your own open source LLAM. I think that the LLAM is essentially this database of knowledge on it's presenting some of that knowledge to you based on your request. Yeah I don't know it's a really tough question because it's there's so many different ways that it could go to depend on each situation there could be a way that the user decides to use in LLAM that like makes them really culpable for something nasty. Likewise the broad the obvious answer as well the people who created the database in the knowledge and the feedback that's coming through it's their fault on the flip side it's like oh the person who's distributing that knowledge is the person who's who's at fault. So yeah I think that's it's you know at the risk of sending a come down into question which I kind of am and I think it's I think it's kind of dependent on what whatever situation that you've come maybe we're going to legislate on that in the future yeah it's a good one.

Jon Radoff: Oscar are you an AI agent youngzilla thinks maybe.

Guest: But at this point based on all well based on all of the integrations that don't look at you be to like make life a little bit easier maybe well we'll fair one thought and it makes absolutely no difference whatsoever. Bed Internet theory it doesn't matter you're looking at pixels on a screen we're looking for engagement and your attention you don't care if the teacher that you bought at Walmart was made by you know a surrography machine or like artisanally made and being painted. Step from that I think that the real killer app when it comes to agentic AI human relations is. I want to be able to walk into the holodeck in Star Trek and describe the context that I want things to unfold in the same way that I want to be able to not have to download an app I want to be able to just tell my phone what I needed to do and have it build the app for me within the confines of that black glass. When I'm thinking about agents and I want them to integrate them into my life what I want to be able to have a conversation with an agent which will then create its own derivatives that were purpose built little agents box on the go out and do what I need them to do so look at how quickly prompt engineering died as a. You know as an area of expertise where before you have to be very very explicit very careful as a trick to do something now you have a conversion conversation and it just sort of like scopes out what you want. I just ask it to tell me how to do what I want to do and then it gives me a framework to do that so in the end I see agents as being. That front interface that we describe behavior and then that'll spawn little ages that'll go off into the universe and do their own thing and it'll probably all end in the end of the universe but for now it offers unlimited possibilities.

Jon Radoff: Alright so let's talk about and another adjacent issue to this brand and Harvey over on YouTube is asking about KYC so KYC is kind of a mess in blockchain. So KYC for those who aren't familiar we're talking about know your customer so. The main place KYC exists right now is when you go to centralized exchange so if you go to coin base they make sure that you are who you say you are and that you're not on like government sanctions lists as someone who might be laundering money or attached to a hostile government or something like that. So that's kind of the main lever that governments have right now is sort of the ingress and egress of money back and forth from the fiat system once it's on chain though. KYC is extremely limited because things tend to operate at a protocol level and protocols although there are exceptions do not generally check for like the KYC status of an individual that's left to sort of the final. So there are some financial institutions who have like who are subject to things like the regulations that govern money transfer so it's an interesting question though what happens in this new world where a lot of the value extraction as well as the value creation is going to be from all these swarms of agents communicating and ultimately exchanging value with each other.

Guest: Yeah, I think it's probably it's inevitable that we're going to see regulation legislation to reflect these kind of changes you know I think that it's probably something that's going to be important to be honest with you and it's going to change significantly it's kind of interesting at the moment because there's been a little like we've got yeah as you said we've got limited KYC at the moment centralized exchanges but you've no KYC to set up a wall right. You just flip up as many wallets people can flip up how's the wallets and but it's the onboarding and all boarding of of money that requires KYC normally and and that's sort of the right in an agentic world where agents are now transacting a lot of value exchange is that is the status quo and not fear saying you expect maybe more regulation I think that's going to be a lot of the same.

Jon Radoff: I think probably more regulation is somewhat inevitable because any time governments don't have control over something it's natural for government to try to fill that vacuum by by establishing render onto Caesar right but you know so I think one theory might be well like the clarity is fun I do that.

Guest: I think what's useful for people is clarity clarity like clear rules about how these kind of things should work so you know as a creator I I want that clarity well so will KYC be necessary for I want an answer to that question as well of course I can share my opinion we can all share our opinions but really what we're looking for is the ledges to the branch of governments and of the world. We want people to clear up these things you know and that's actually helpful but a lot of people seem like especially in the kind of like libertarianism people to see a lot of like the idea of like making rule as it like this terrible thing that's not necessarily the case I think like I remember speaking to a very interesting toponomic designer or a lot of that could spoke about how I think there was around the Wall Street crash that actually regulation was one of the things that really helped drive the kind of. The what came after that you know and actually created a boom on Wall Street people actually really wanted to just like here set me the limitations I find this is a creator a lot right like if I know the limitations I can build a way better shit them my doubt you know I found that as a game designer found that when I work in film so I accept the set the limitations and then we can make amazing in a bit of stuff. And sometimes in crypto you just have to accept the those limitations aren't there right and you kind of everyone kind of pushes in all these different directions that's the kind of crypto way but I think I think that more clarity is good for I think I'm with these kind of questions or good questions I'm sure if I'm the person to legislate on that you know so I need to be careful about what I answer about those kind of things but.

Jon Radoff: But yeah they're good questions yeah I mean that's sort of aligned with the CEO of coin basis take which is for years he said let's have let's have regulation that's at least know what the rules are that we're supposed to play by and then we'll play by them and for years the SEC over here in the United States has been like no we'll we'll keep you guessing and we'll we'll just be fine also it's against you to keep you scared. That doesn't work on still.

Guest: Yeah exactly that's stifled innovation you know and what happened in the States over the last like four years stifled innovation inside a crypto dramatically and there was like an outpouring of innovation to like almost everywhere in the world and which I think it's a which I think is probably changed very significantly over the last year and. So yeah I'm on I remain very optimistic about the future of crypto I think I think we will find that clarity I think that we will constantly work with some degree of blurred vision.

Jon Radoff: So to speak because that's the kind of nature of working at the aspects of this kind of cutting edge of technology and almost sociology as well yeah I think it can be okay to have a common understanding of what the rules are as long as those rules and regulations don't become so onerous that only large monopolistic entities are yeah.

Guest: In practical terms able to support all of the infrastructure necessary to even comply with regulation because then that becomes very anti innovation and well that's the boomer that's happening so out of year it you so I mean the states about their own difficulties but in the EU it's become so like like non-estateish about about what they're trying to do that really try to enforce legislation legislation to try to provide that clarity but it's. Major overreach which is stifling innovation and it means that a lot of folks who a lot of reading smart folks from around Europe are not setting up their AI businesses every year and because yeah it's it's a very concerning future for the EU when it comes to this so yeah.

Jon Radoff: Yeah see you're talking about it in the context of things like KYC and financial regulations for blockchain but you're rightly I think moving it to an even broader set of regulation which is regulation coming to artificial intelligence specific. Sounds like you're concerned about it specifically in the EU but how is there regulation that you've been exposed to that you've been concerned about or that you're concerned yeah.

Guest: Yeah no there's existing legislation lack of clarity about that legislation is famously does a GDPR stuff. You know when you're in Europe you have to click all sorts of things to agree to see a website often over kind of privacy concerns and those privacy the way that legislation that legislation has been built as has a very far reaching consequences for people who are trying to build AI. But as a result you might know so when when for example a new feature comes out in open AI or Gemini or like deep mind or any of these folks when they released new cool technology they almost never release it in the EU first and there's normally quite a cooling off period in order to be able to bring it over to the EU which is fucking terrible for big builders with AI but you want to be able to access the cutting out of the EU. So we've got relationships with some of these folks so we're able to kind of kind of get early access and stuff like that but most people don't have those relationships and yeah you want to work at the cutting edge technology you've got to be at the cutting edge technology you've got to have it at that moment you know no point getting some amazing and genetic technology six months later. You know we work in AI things it's a totally different landscape then so yeah.

Jon Radoff: So we're most innovative companies are just going to move to the locations where they can be allowed to innovate to so precisely these countries want to have a piece of this future then they need to work with companies right now just tell them what not to do we're getting towards the end of our hour Mike I want to kind of give you the final word in the conversation here so what's your vision for you me own where we take it next and what should people know where can they go learn more what what's their big take. What's their big takeaway here on you me and the things you care about today.

Guest: Cool well first of all thanks so much for having me on and it's been a really fun chat dive into these different aspects of the AI and I kind of already enjoy the free flying nature discussions like this and dive it into someone knowns as well sometimes and yeah listen I mean we're we're kind of right now we're head down we're bill and we're.

Unknown: Very focused on bringing our launch pad and to the world whereby anyone can go on and use advanced AI and create really cool characters and that are powerful and we want people ultimately to access this groundbreaking technology through the prison of easy to use and fun user interface personable characters and so that's something that people are interested in come over say how.

Guest: Come over say hi to us where you can find us at you me on your score AI and we've got some really cool things that are upcoming we're getting closer and closer to be able to take this into an open test so we're going to have some news on the soon I know this probably some folks in our community we're looking at this now we're like saying like when launch pad or when token or all that kind of stuff and I assure you it's that the launch sponsor is going to be.

Unknown: It's going to be a killing sooner than later and we've also got some really kind of interesting ecosystem and answering this upcoming as well so yeah things are being owned really well I think the broader crypto markets going well as well what about you John how has everything going with you.

Jon Radoff: Things are going really well they're very busy where I'm really focused on rolling out our nodes rolling out our tge it's all coming over the next several months so that's why we've got a lot of people here tuning in we've got over 1400 live viewers at this moment at the end of the talk so. People tuned in to learn about this subject and probably pick up the code but they stuck around we kept growing during the course of it so they weren't just air drop hunters showing up for a code and then piecing out on there they're here to really be in the community so we really love that our community is amazing at beam while we channel you guys to new games you you caused people to get like a hundred thousand wish lists sometimes on steam which is just insane like any game developer if you get good. 100,000 wish list on steam you just minted gold for yourself so we really thank our community for jumping on those opportunities that we've presented but yeah we're working on the token and the token launch and all that stuff the easiest way is to just tune into beamable network the the x page or mine J right off and follow those and you'll keep learning tomorrow i'm at permissionless in New York City so if any of you are there we've got almost 50. 1300 now so if you're here at permissionless meet me in person we can talk about all of the above but thanks Mike for taking part of this it's been a super fun conversation I feel like we could just talk for hours and hours and hours but sadly our hour is come to an end and we're going to have to say goodbye for now and thank you all of you for tuning in super grateful for all of you for being here your time is so valuable and you decided to be here with us so that's really fun and great. If you're watching this in replay go ahead and join live next time ask your questions and participate just as we did as we did in this particular conversation thanks everybody Mike thank you too thanks so much.