Tezos EVM (Etherlink) AI Hackathon

0:02 All right, we are live live to you from Brooklyn, the now media office. we are excited to present the first ever Tesos EVM Etherlink AI hackathon. we received some incredible submissions and applications and it was really tough to narrow it down to just a few finalists. and we've got an amazing star-studded judge panel for today. I'm going to pull up Ian and Jalil and I know Seth will be joining shortly as well and we can talk a little bit about the about the process.

0:36 Ian, Jalil, you guys are live with us now. How we doing? >> Hey everybody. >> What's happening? >> Hello Paris. >> Hey there. >> Let's see how we can get this layout. There we go. Better layout here. so yeah, it's been it's been a really fun one. you know, since we announced it a couple weeks ago, opening up the doors to to projects. I'm curious like from your perspective, everyone, like what were some of the things that like stood out to you from the the submissions, the finalists? I know from from my perspective, I was really interest I found it really interesting to see the the wide range. you know, there were a lot of like one, you know, kind of to be expected, you know, some ones that were really focused on financial rails for agents and and crypto DeFi, but I was really happy to also see a number of ones that were a bit more active in the cultural space, some art focused projects as well. So curious like >> for me, >> for me, for me to see everyone just like building, right? everyone using AI to like get their ideas out, whether it be something creative or something kind of more on the DeFi side. you know, just like the fact that the ability to build was like really, you know, eye opening as a developer is like seeing this kind of new new world unfold before you. so that's what was really exciting. The wide range of diversity and all that type of stuff was cool. yeah, >> I think I found personally, obviously as head of art, creative projects are always my top line, but I found it really interesting to see use cases that are not sort of the most popular, but you would you wouldn't ever think of them to the other. Like there was I think one use case about like if you pass away, what happens to your keys?

2:31 And like there's this like p and I think we saw a really good mix of something for the broad broader economy like prediction markets the most popular applications right now adding novelty to that but also very very niche things and I think it's amazing to bring together people and kind of like revisit and have these conversations to see how can we better support some of the builders out there today. >> Yeah. Can you guys hear me?

3:00 >> Yeah. loud and clear. >> I mean, I I think to to follow on what you guys said, I mean, to me that what you're seeing is and it's in what you already said, but but you know what like world of possibility. I think one of the most interesting things that that what people aren't talking about. People talk about what AI replaces, but they don't really talk about the things that simply wouldn't have been built otherwise. So, your point about things that are more fringe, I think, is a great point because like one of the things that we've all been doing for the past year is just building that we needed that no one built for us, right? Like we're so accustomed to software doing like 80 I mean I'm sorry more like 20% of what we want it to do and then the humans do the 80%. Right? we're just used to it. We you know we we you know if you I mean if you're if you work somewhere and you use a product like Salesforce man like the amount of work that the human has to do compared to like what that platform actually does for you is like it's you know it's a it's way out of balance. And so, you know, to me, like what you're overall what you're seeing is like, oh, people solving problems, right? And to your point, some of those might be more like for a smaller target market, but when the cost of building software comes way down, that makes perfect sense. And it and it's actually a great thing. You know, it's not it's not like a it used to be, you know, you go see a VC, right, and the first thing they want to know is like how many people is it going to serve? And I I'm I'm kind of into this concept of like luxury software where you know where it's not for everybody.

4:26 you know it's it's for the people that it's for. and I and I I think that you know there there are a bunch of things but also I I think that to me looking through all of those things it's it's like ideas I hadn't thought of. You know what I mean? So so you're and and and then you have to get your head around like okay is this useful? Who is it useful for? like you know and and and that but it's also but what it really does is it just shows you the world of possibilities.

4:58 >> Absolutely agree with that. Jal interested to hear your your perspective as well. >> Yeah. nothing much like for me I think the the coolest feature of going through a bunch of projects like this is to see contribution from all over the world. I think that's always so fun to see the little demo videos and the different accents. As a non-native speaker myself, it's like it's just a beautiful thing to see a collaboration from all over. and people that maybe like don't have the best access to a tech hub like you would in San Francisco, whatever. Being able to participate in something like this is, I think, something truly special and just beautiful about this this space in general. Well, actually that was something I was going to mention too is that that like the ambition like it's it's ambition that is off the charts which I think is this is where hackathons are different than they used to be. Right. It's funny because right now here at Ledger we have an internal hackathon going on. It started this morning. It ends tomorrow. I'll spend my entire afternoon tomorrow judging the the internal hackathon tomorrow and I can't wait. We've already you know seen them. But what what's crazy and we saw this we actually posted a video a few a couple months ago because a couple a couple product folks from the company and me participated in the circle hackathon. a couple months ago we built this thing called agent intents.

6:22 We built it in about 36 hours. And the crazy thing is the only time I've been in a hackathon where you actually increase the scope as you go along. We increased we increased scope like three times. And I think you see it in in what I saw too, like some of these are super ambitious and so you know, like it's always the case like sometimes you got some stuff in the cam before you start the hackathon, let's be honest. but you know, like because of what you can accomplish quickly, you know, like if you got a good spec, you know, Claude's going to crank that out in 25 minutes, right? So you if you if you've got it in your head, you can keep going and you can keep increasing scope and the level of of ambition goes way up. So I think you I mean there's definitely some of these here you're like whoa that's ambitious for a hackathon but that's what's like I was thinking about this like six months ago I was like what does a hackathon even mean now and the thing that we've tried to to do at ours internally is it's the first time I we've never had participation like this right because there's people from the finance team people from the HR team people from the marketing team it's not it's not for the engineering team anymore you know what I mean like everybody is sitting there right now with you with with AI tools building and that that's just a we we had last night we had an agent synonymous openclaw meetup here at Ledger as well and one of the presentations was a guy who just showed he was actually he like put up a slide that was to just show you how non-technical he was Seth was there and and like and then like here's how I did it you know and he shows like the concept and actually like I think for more technical cool people in the room, it was maybe a little boring and pedantic, but I was thinking about it from somebody who who had never thought of how they could build something and how it might be inspiring, you know, for them to be like, oh, like you can like, wow, okay, you can build that by knowing this and you can only know, you know.

8:19 So, anyway, that that's like that's just a I think that's something that we're we don't we haven't absorbed yet. You know what I mean? Like over overall, we haven't absorbed that. It's like a dopamine release every time it builds something. It's like you want to do more, right? So that's why everyone's staying up till four, five, six in the morning, you know, because you have that. I had that first with like mid Journey and Dolly, right? Like >> dude, hold on a second. This is the perfect introduction to Seth.

8:45 >> I'm gonna do I'm gonna I'm gonna do Seth's introduction, Matt. I'm gonna do Seth's introduction. This is this is real life. Okay, so I started All right. I'm g tell you guys a story that you'll appreciate. the the that I've told before, if you go this if you go search my name and Rick Rubin on YouTube, you'll find me interviewing Rick about about about vibe coding because I I I texted him at some point like I remember because I was at Consensus in Hong Kong and I was like, "Have you heard of vibe coding?" And he's he was like, "No, but I think I'm the meme." And that was kind of like how lots of other things unfolded. So I had been like, you know, coding using chat GPT is the way I would put it for a while. Even like the ledger opening, some of the stuff that we were all had on display was me just like not wanting to write JavaScript and making chatgpt do it. But I hadn't been using cloud code at all. And then Seth came to stay with me in the fall of last year and he was at our dining table like all day and all night with his laptop open. I was worried about him. I was like, I think >> very antisocial.

9:51 >> I think you might have a problem, man. Like, what's going on? And he was like, cloud code can't close the laptop. Can't close the laptop. And so, but that was that was how I got because I was I knew about cloud code, but I didn't really think like, oh, that's for me, you know what I mean? And then, but watching Seth use it, I was like, oh, I think it might be for me. And then it turns out like as an old command line, you know, as somebody who knows both Emacs and VI key bindings, it was definitely for me.

10:20 and and I live in it now. So anyway, that's that's my introduction to Seth. He he he was the addict who gave me my first >> Yeah. Well, Seth is the addict who gave me my my first my first hit of that dopamine that you just described. Just which is where why when you said it, I'm like, yeah, because like Yeah. Anyway, it's not the coding. It's the dopamine of the of like seeing it come to life >> like believing it, right?

10:49 >> Yeah. >> Live. >> It's unbelievable. Seth, I I know you did I think on the podcast, the led to I podcast you did with Ian. I thought there was a really great there's a great clip that I've I've shared with a lot of people where you talk about Rakman Rakmanov and the the the piano where the music is coming out of your hands and like you can't believe it. And I feel like that's that's a very shared experience for those of us who have gone down the vibe coding rabbit hole where it's like I can't like I I didn't put in the 10,000 hours, right? But I'm but I'm doing this, you know? and and so I I thought that that's been a really great like perspective and love love to give you a chance to introduce yourself and tell us a little bit about you know your thoughts as as you know we were kind of going through some of the submissions and you know and and going through the project.

11:39 >> Yeah. And it was great to see the submissions and like this obviously like we're early in this new primitive, this new creative primitive, whether we call it vibe coding or personal software or whatever it is. it's definitely early and I think we're still amazed at the of what we can do, right? We're in this phase. Even I was at the you know it wasn't Alcoholics Anonymous, it was Agents Anonymous last night. and there's this sort of like look I can you know I have I made an agent and I can call him and and we're sort of at that phase of this where like it all seems magical when you can spin up an agent and you connect it to Twilio and we did this on a Let's Vibe podcast grow with Farmer Fred from Proof of Corn. you can call him and talk to him and that alone is magical. But now we're getting to this point of like okay now what now that we can connect all these surfaces like I when I think about developing agents in my own practice there's like there's two things I think about like one is an agent that compounds intelligence we think a lot about like compounding investment and compounding whatever practice you have but like I want an agent or agents that are consistent ly refactoring what they observe and what they read and what they write back into themselves to get smarter, right?

13:14 so it's a kind of reinforcement learning I guess is the lingo. and then I want that compounding intelligence across multiple surfaces. So if you have an agent that can exist in Telegram or WhatsApp or iMessage or the web or X or Instagram is hard. but you know you imagine you have all these surfaces and you want your agent to be available to through all of them and that no matter what the agent no matter how the agent is dialoguing with the world or with other agents that the intelligence from and through those channels keeps compounding. And I think that's where we're getting to where there's like when you think and I mentioned this to Ian last night and I'm old I won't say Ian's old but he's older. but like you know my memory of Apple Microsoft is very different than what it is now. Like there was a time 20 30 years ago. It was like the Apple camp and the Windows camp and like Windows was Steve Balmer and it was like all about developers who had no taste but had massive scale, right? and Apple was much more closed but you could still develop on it, right? and it was much more curated and people would argue it was more closed but it was also more beautiful and so that's how I think about the cloud code ecosystem and the open claw open AI ecosystem right there we go next is kind of in between right >> but like we're in this I think like both you know both Apple and Microsoft were operating systems and they're Apple computers and you know Windows computers Right. And I feel like we're kind of coming into this similar world where there's like two camps. You're either developing in the cloud camp and the empathic camp which is I think more elegant but more closed or you're developing in the open AAI open cloud camp which is much more open where there seems to be a lot more developer energy but it's not the experiences aren't as curated and the aesthetics aren't as tight. and so, you know, where does that leave us now? Well, okay, you kind of choose one. People may try to go back and forth. People, you know, there are people that try to open up windows windows on apples back in the day and that was a thing. But, now it's like, okay, now what? Now that you can spin up an agent that gets smarter over time, that can talk across all these different channels, why?

15:47 What's the purpose? Just because you can do it, should you be doing it? And I joke that, you know, in Silicon Valley, where I recently moved out of, all the vibe coding projects are the same thing. They're like fitness apps or to-do lists, you know. And I want to see weird applications, and I want to see weird agents that are opinionated, that don't try to make everybody happy, that don't try to agree with you, that take a stand and have a, you know, have a deep skill and don't try to be everything to everybody, which is the case with Claude or with GPT.

16:22 >> Okay, that's my hot take. >> Keep it weird. >> Keep it weird for sure. I I don't know if you're saying that Daario is Steve Jobs, though. Let's be clear. Like, neither one of these guys is Steve Jobs. >> I think Dario's close. I think he's as close as we have. >> Might be more. >> By the way, by the way, my my t-shirt just matches my my tattoo. >> There we go. >> Wow. >> I love it.

16:50 >> Steve Jobs love. So, >> yeah. With a texture and everything. well go great. It's it's a it's an amazing group here in terms of in terms of judges. Seth was also the one who got me into vibe coding. so it's and I'm really glad we also we have Jal here. We've got Aaron who are also come from a technical back background. So you know bring that those years of in the terminal to the to the equation there. so I I think what we'll do now is I'm just gonna run through like how we're going to do the process and and then we're going to jump in with presentations. so we've got we've got our finalists in the the backstage here and yeah, really excited to dive in. each finalist is going to have five minutes to present their project.

17:48 I believe we've got five live presentations and three videos that were submitted. so we'll go through those. after that, you know, the judge panel will have the opportunity to ask a couple questions if if and and you know to for clarification or to to learn more about it. and then we'll move on to the next presentation. once we've done run through all the presentations, we're going to go we're gonna go on mute here, play a little music, throw up a little graphic, and then convene to talk through and then we'll be announcing the winners. I know Ian has has a hard stop. so we'll we'll we're happy to have them for as long as we can and and we'll we'll get things going. so one thing just a reminder reminder for the judges and also for those listening the judging criteria we settled on four categories that's AI innovation so creativity sophistication of the integration novel approaches you know technical execution the quality of the code the architecture security proper use of of the Teslink capabilities reach usability design accessibility real world application is this something people would actually use and then ecosystem impact. how this can contribute to the Tesla CBM and and Tesos's ecosystem. so really excited to dive in. let's let's get it going. I think our our first our first presentation is going to be with Krun. So I'm going to pull Kun up and I see he's ready to go. Here we go.

19:23 We got Kun on stage. and Kun you should be able to share your screen. Let me know if you have any issues with that. >> let's go and do it. >> Perfect. >> Yep, I see it. I'm gonna add it to the stage now. That's great. And Kieran, feel free to give a give a brief, you know, introduction to yourself, your project, and then run through it. We're excited to see it. >> Perfect. Sure. Shall I get started?

19:53 >> Let's go. Yeah. >> Excellent. happy Wednesday, everyone. my name is Key. I'm a creative coding artist from India. And thanks to this hathacon. this animation which was on JavaScript 3.js has now become an AIdriven system that's what I call and what you see live is this system which is in a resting state. this is live hosted on my website so that I can actually you know present to you all. Now what you can see in this system is primarily you know you have a central you know planet and then you have couple of planets which are orbiting. I was actually working on a living art collection. Over the last few years as an artist, I've seen that always collectors wanted to be somewhere involved, not just bidding and collecting my work, but also to be a co-author to my work. And that's where thanks to this Tzos EBM, you know, ether link hackathon where I could start using agentic AI and create an autonomous system which listens, reads and acts and makes collectors as part of the art itself. Let's get to the demo.

21:05 For our, you know, purpose, I have used a display here what you can see on the screen. now as I said, you have only two of the planets currently and once there is a blockchain activity like a bid comes in. You can see that there is one more. Now, a planet gets spawned into the system. Now that's the beauty of this you know AI where the agent is autonomously actually you know reading the signals from the blockchain activity which are sitting on you know ERC804 or ERC 721 and the display screen is capturing every bit like who is the bidder the wallet address what is the bid amount and you know immediately also they are becoming part of the art basically the canvas is now offered to the collectors this as we see more bids coming in, you have actually you know more and more planets getting spawned into the art and the one big challenge that I was facing as you know someone who love to develop infinite loops. This was actually a challenge for me because more and more new things are coming into the you know animation it disturbs the pattern. So that was one challenge once which I could overwhelm the way actually the claude code was able to help me sort that out. The rest everything was easy.

22:24 Now every auction as it comes to an end we know that you know it sets a pattern. So the system also actually you know understands that it is getting to a terminal stage and auction is going to end soon as you can see on the display screen. Now currently we have so many planets and whoever is the highest bidder they should be one who should be awarded the final screen and that's where it freezes. Now you can see that the system is locked. All the planetary orbits that are there this directly gets minted to the final collector wallet autonomously. All this is being done by the AI agent itself. Now what was supposed to be an auction mechanism which was done through blockchain is now no longer just an auction. It creates the art itself. This is celestial convergence where I have used 3GS and theos ether link chain has been used for the testing purpose. Currently this is on you know the test net but it is very much ready to be moved to the main net. Thank you. I'm open to any questions please.

23:36 How far does it grow? Like how many planets can you add? >> great question. So I tested up to 256. U it's just that the screen was not visible but it can go on to whichever extent. We can predefine the numbers if at all we want to restrict but yeah there is no limit. >> And the the amount of the bed does it affect the size of the planets or any >> that's correct. Yeah that's perfectly.

24:02 So actually speaking there is a legacy for this collection. The central whatever you know planet that you see comes from its predecessor artwork. So each and every art actually you know carries some of the trades like for example the bidder wallets hashed it actually you know becomes part of the next artwork. So that's how each of these planets when a new bid comes in like you rightly said the the color the hue constant and also the size of the planet is decided by the weight of the bid that comes in.

24:37 >> How did you come up with the name Kiran and how does that how does that tie into the the concept here? >> Yeah, I mean Kieran is my name actually to be frank. So >> Oh okay. >> Yeah. >> Wonderful. Thank you, Kieran. >> Thank you so much. >> Kieran, can I can you explain in a little bit more detail what the agent actually does behind the scenes? Like does the does the agent transform the visual by writing updating code itself or >> yeah what's it like what's what's the scope of the agent >> right? So AI here primarily focuses on first of all understanding the blockchain activity. Now like you saw initially there are only couple of planets along with the central you know planet which is sitting as stationary but once the blockchain activity actually you know starts happening where the bid comes in it understands reads and once it has the RPC call that is coming in from ether link it immediately spawns a planet based on you know the bid amount that has come in so that's how the first step comes in next it goes on studying actually all the bit comes in keeps creating these new planets finally Finally, when the auction is getting close to the last 10 minutes, it understands that now it's in a terminal state. Let's move the auction into the final stages and it prepares for the last step where it has to freeze the artwork and then autonomously move it to the final wallet or the collector who has actually know bidded the highest. So all this is done autonomously by the agent itself. I don't have any kind of intervention in this and everything is verifiable on chain.

26:17 Cool. Thank Thank you, Kieran. Any any other questions from from the judges or we'll move on to the next. Okay. Thank you. Thank you, Kiran. Thank you for presenting. >> Yeah. >> All right. I think we're g Dorian if you're there, I think we're going to go to you next. If that's good, I can pull you up on stage. All right. Great. We're gonna pull up Dorian now. And I see you've also got your your presentation as well. So, we'll pull that up. Dorian, welcome to the stage. Great to see you.

26:47 >> Thanks. Thanks a lot, guys. And, yeah, cheers to everyone. cheers to you, Ian, for having a great event yesterday at the Ledger HQ. I was there yesterday. It's an incredible rooftop. And, yeah, it was an incredible event. So, yeah, thanks for putting that all together and and to the Tezos ecosystem and Alexandra and all the friends there, thanks again for putting this together. a bit about myself. I'm a journalist and editor, and and more recently a vibe coder.

27:19 I'm I'm not really a developer by trade, but I'm trying to learn and soak up as much as I can thanks to all these incredible tools like OpenClaw and and all these incredible tools that are kind of allowing me to yeah, execute some ideas and just kind of hack away and learn. so yeah, a bit about my project. the idea was to essentially create a a minting function for hashtags on social media. and I had I didn't really anticipate it being that hard.

27:52 but in fact, it was that hard. so basically it's an autonomous AI curator that watches social feeds and mints identity NFTts on Etherlink, turning the imagine on TZO's hashtag into an evolving onchain identity gallery. and the core idea is to ingest posts from X and Farcaster at the moment. running them through a multi- aent GPT panel which includes like a chromatic narrative and pattern sort of recognition component and and then to fuse those into a single sort of interpretation and and pump that out into a sort of gallery. and from there it generates at the moment just a sort of abstract art and it turns it into an HDM artifact. and in within that there's some JSON metadata and that that gets pinned to an IPFS and then minted into an ERC 721 on right now the Etherlink ShadowNet. so basically I'm trying to figure out a way to capture projects that people may find interesting. So if you see a post online and and you you know would like to see that as an NFT the idea is to use that hashtag in order to basically mint that as an NFT. so what's going on in the back end there's some onchain batching. there's a solidity smart contract as I mentioned before. it's running on the T Tezos EVM and yeah and so I can kind of just play the the demo as I had it. I don't know if Can you guys hear the volume on that?

29:39 >> Okay, cool. >> >> imagine onos AI curated identity gallery. >> We can hear it but it's very low volume. I don't know if there's a way to >> Oh, I'm sorry. >> what is it? >> I can kind of just go through some of the I can just kind of go through it on my own mic here. so basically this is this is kind of the read me blah blah blah. Not really super important. and this is the actual functionality here.

30:14 we can kind of go through it. so it was just living locally actually on my I hadn't actually put it online yet. It was just living locally. and so so yeah, that's kind of the the gallery as I had it set up. and then I had I had taken as I mentioned, I had it wasn't it was not actually able to discover the post yet. I haven't actually included that functionality yet. But basically, I had posted a test a post of a a Kanye Pepe that I posted on my on my own ex account and you can kind of see what that back end of the AI is doing. So, it's just sort of like describing the image that was that was there and it's filtering that and kind of describing it. What I what I programmed the AI to do was to contextualize what it's seeing. and and contextualizing that within the let's say art historical cannon of the Tzos ecosystem. So some of the art that's been posted on there. So that that was kind of the prompt component where it was like okay take this hashtag and then describe what what is above that hashtag and contextualize that within you know the TZOs art ecosystem. So that's kind of what it's doing with this test post. It's, you know, it's saying, "Oh, you know, this is, you know, a cultural symbol, you know, meme juxaposition of Kanye whimsical character." So it is working in the sense that like it's able to describe in a way that is sort of like let's say like canonically accurate of what it's seeing. now where I'm running into issues is that like I'm not able to really like run this like every 30 seconds or you know just searching for this imagine on Tezos hashtag. so I'm I'm running into a few problems in terms of like how to actually like develop the functionality to the point where it will you know continually continuously sort of you know crawl through Farcaster and X. There are some I believe API issues. I do have the the X API that I'm kind of integrating, but there is some there is some well I mean I'm pretty sure I need to pay to to now go through kind of X's API. I think Farcaster will be a bit more friendly in that respect but I haven't actually gotten to that component. So right now it's just basically it it's just working in the sense that like I I've you know minted my single x post with the imagine on tzos hashtag and basically yeah so yeah in terms of like the back end it's just in it's running node and express and sqlite so there is a schedule there is a scheduleuler in place for x and forecaster ingestion but it for some reason I just haven't been able to get that actually you moving through my rest endpoint. so I'm I'm still kind of tinkering on that component, but yeah, I I have managed to sort of mint my my Kanye Pepe. So, yeah, that's basically it. And yeah, Hely Rick Ribbon vibe coding.

33:39 >> Awesome. Thank you, Dorian. Appreciate the the presentation. let's open it up to the judges for any any questions. Dorian, do you come from the TZOs ecosystem or are you like familiar with the EVM before this? >> I mean only peripherally like I like I mentioned before I'm not a developer. I've just been following some some of the projects like object and hikm and you know just some of the the folks that have been involved in in TZO. So I've been really inspired by generative art in fact and a few of the projects with respect to you know the generative art let's say canon that has been proliferating on Tzos and trying to figure out how to turn you know code into to to artworks and in this case you know how to how to combine a sort of hashtag functionality but yeah I'm not really an expert by any means.

34:37 No, >> I' I'd love to hear like what your experience was navigating. I I see in your project, you used a couple ERC standards that you implemented to run your project. Yeah. >> Yeah. >> What was that like experience like to like try and understand what the standard does and how to implement it yourself? >> Yeah, for sure. like of course you know ERC 721 is kind of the standard for NFTTS but I've been really interested in learning more about like you know 804 you know and sort of different smart batching you know whether it's like ERC8211 as well. So I'm trying to like learn about how to like basically mint add a chapter plus like state checks as well into like a single atomic transaction rather than kind of like breaking it up. So like the learning curve is I'm not going to lie is massive. Like I'm definitely not you know I don't have a background in computer science or engineering at all.

35:35 Thankfully I do have some very good friends who are who are quite adept who I can sort of bat ideas off but and I did actually do the the Ethereum hackathon a couple weeks ago in Khan as well which I you know it's a great opportunity to learn and like learn from the amazing projects like Kieran's as well I thought was super well done and very inspiring the way the planets work. So, for me, it's just about learning and gaining new skills and, you know, meeting cool people and yeah, that's that's basically where I'm at.

36:05 >> I love it. Dorian, throughout the process, any any kind of interesting moments or like challenges that you ran into in terms of putting the project together, getting it across the finish line. >> Yeah, like I could go on. Like for example, like I wasn't able to upload it now on my local host and like my open claw is down. like it's not discord. So I had to play the video. So I was like stressing out. I'm like I got to play the video. They're going to think I'm like completely dumb. So like definitely like my open claw is down right now.

36:36 It's not talking to my Discord bot. So I'm kind of just like oh my god. >> When open claw goes down. It's a bad it's a bad couple hours. >> Especially when you don't actually have the skills to like diagnose what's going on and you're going from like perplexity like why isn't this working? And then you're like going back. >> Yeah. Yeah. So, you know, like it's Yeah, it's definitely like, you know, I I have to say you know, there there it's just it's just constantly debugging, right? Like you're just figuring out why is this not working.

37:08 >> I love it. Keeping it real. I mean, we we all know that. We all know that pain point with Open Class specifically. >> Yeah. Yeah. But yeah, but again, thanks for putting this together, guys. And it's really cool to to meet everyone and I'm really looking forward to hearing about the other the other projects. Awesome. >> Thanks a lot. >> Yeah, thanks for thanks for joining, Dorian. Appreciate it. We're gonna drop you back down. Oh, thank thanks again for joining. all right, we're gonna we're gonna pop over to one of our video submissions next.

37:38 >> A very exciting video. >> Yeah, very exciting. >> I hope you're ready. >> You know, like I I was getting everything ready. I didn't even get like my normal coffee this morning, but then I watched this video and I was like, "Yo, I'm good. I'm good. I'm good. >> The coffee." So I'm just going to give a little background to so this is Maderu lineage an experimental art DAP where AI autonomously creates titles and interprets art works on Etherlink.

38:02 when a user requests a interpret a reinterpretation the AI reads the parent artwork and generates a new derivative minted as a child NFT with an onchain link to its ancestor. over generations a visible tree of AIdriven creativity emerges. built by the team behind Maderu.art, a Tzos L1 art platform. this is their first bridge to EVM ecosystem. and so yeah, they're we're really excited to dive into it. I'm going to pull it up right now. Share the screen. we've got it right here. Boom. We'll pull that on stage and do a little full screen here.

38:43 Oh, wrong full screen. Oh, no. Is it the right one? Yeah, I just fine. There we go. Boom. >> Imagination is connect and comput. >> What if creativity could be inherited? What if an artwork could pass it to a child endlessly autonomously? This is the lineage. AI artificial intelligence imagination is infinity connect human and computer singularity.

39:31 We make future artificial intelligence. We are the team behind me. The cattle art platform. Madero lineage is our first experiment bringing autonomous AI creation to easily artificial intelligence. Imagination is infinity. Connect human and computer singularity.

40:01 We make future glowing bright. The AI generates a new artwork, titles it writes its own interpretation. No human input. Zero. Now we mainten the AI reads the parent tigle and interpretation.

40:46 that generates a derivative. The child permanently link contain to its parent intelligence imagin. Here is the family tree. 10 generations of interpretation, ranching and evolving. The genesis work, a quiet landscape. Five generations later, an abstract painting ethically unchained connected.

41:53 We have everything. >> This is only possible on Ether New Zealan decision. 500 millisecond finality makes the lineage feel alive real time. On a real mainet, nobody pays Korean for an artwork on either link. The genealogology grows naturally.

42:24 >> Art infinite love. Infinite code >> mad lineage is also a bridge leader art has been serving the TZO cell one community for use. This project brings that philosophy to EVM inviting artists into the ether link ecosystem and in the near future users will walk through portals into each other's NFT galleries.

43:10 The human roises from artist and genealology of AI interpretation. Thank you. AI AI NFT NFT >> I'm from >> AI. >> All right. All right, >> we we know where this is going.

43:43 >> Yeah. Yeah. that look very creative submission. I I mean I enjoyed it, you know, as a as a Sunno user myself. I I I appreciate the the creative audio and the music would approve the song. >> Yeah. Yeah. I know Maderu it's a time zone. I think they're in Japan right now. So I I don't think they're they could make it. But if any of the judges have any comments or anything we'd any any further thoughts?

44:14 >> I think like one thing I want to point out, A I'm still tripping from the video. It's very cool. But B is I am wondering sort of the usability like the use case like as a normal user like do I wake up and I want to >> use this? But what I do enjoy directionally is kind of tapping into the L1 and the platform that Madera has built on L1 and then seeing how it naturally extends to sort of Etherling, Tzos, EVM and has like that conversation that relationship and maybe like broadening the scope of the the reach of the initial platform like from an ecosystem perspective. I find that really interesting.

44:59 >> Yeah, absolutely. Yeah, I think that the the I think the visually it it was like very entertaining, right? Very like the the music I think hit really well. I wonder like the why the what's the I I wish they were here to explain, but like what's the purpose of the lineage? What's the purpose? like is there a deeper message or is it are we just like running our generations on loop?

45:34 >> Feels like a work of art if you think about it like >> yeah it could be just a work of art in itself in terms of the the whole system you know >> but yeah >> I I would have loved to be able to see see the actual project code and the project running. that was my biggest red flag about this one is that there wasn't actually any code that I could look at or >> and it's like where's the it was like a visual. It's like an Instagram video almost, right?

46:06 >> Yeah. Yeah. Yeah. Yeah. It's I mean it's Yeah. fun to see all the different manifestations of people's tastes and that's >> awesome. Well, no, I appre appreciate that. Thank you. Thank you to Madera for for submitting that. we're gonna hop over to our next live presentation. we've got Jenny. Jenny Kruzy. We're going to pull pull Jenny up on stage and, Jenny's going to show us her project. So, Jenny, we've got you.

46:37 are you are you there? I think you're Yep, there you are. >> GMG GM. Everyone, >> GMGM, welcome to the stage. We're happy to have you. >> Thank you guys. Am I audible enough? >> We can hear you loud. >> Yeah, we can hear you well. >> Okay. So, I'll proceed to share my screen. >> Yeah. Once you share your screen, I I can pull that up.

47:14 there we go. I see it adding to the stage. Boom. >> Okay. Okay. So, I built Okay. First of all, Jenny, a beginner data analyst that just took interest in vicoding and since then I dumped my data analysis books and it has been by coding all the way.

47:45 So I really appreciate the Tesos team for this hackathon and I must say this is my let's say biggest build so far. I've been voding lots of stuffs. So I'll just dive into what the project is all about. So Sovereign Genesis like I call it is just an is an autonomous AI agent that manages its own treasury.

48:18 >> It post development bounties when it has surplus funds and makes mages PRs. It matches PRs of contributors whose bounties whose PRs are valid and makes payout to these contributors. it also invests surplus funds. As you can see, the treasury balance is there. It has a life support buffer so that it doesn't go inservent bounties when it has surplus funds. All you just need to do is fund the contract which I'll be doing right now. You fund the contract and you would see it's posting bounties live.

49:05 Sorry about that. So I'll just find the contract with Tesos. I hope I don't rob I think I have some Tesos for gas here. So I'll just fund go ahead to fund the contract. Okay, this should proceed. Send.

49:45 Okay, I'm funding the contract. It tells us you would see As you can see this this is the agent. You can scroll down to see what it just did. Now it shows that it received five XTZ and it's posted some bounties. So I will just go ahead to show you how this agent manages PR autonomously and how it makes payouts. So I'll just submit one PR and see if it gets merged.

50:30 Okay. this build works better if if it's an external contributor, but I can still manage to show you how even though I'm a the owner of the repo. So I will just go ahead to create a new file to submit that particular PR and see if I would get any payout. So I would proceed to name my file So I'll enter the file content. This is just the normal way you pull request. So I have the code snipets already down to save time since this is a timed presentation. So I'll just copy the code snipex.

51:30 Enter the contents. Commit changes. I'll I will not commit directly to the main branch because I'm trying to make you to make this presentation as an external contributor. This is what an external contributor would do. The only thing I'm not doing that an external contributor would have done is this repo. But since I am the repo owner, I won't I cannot be able to the repo. So I would now propose changes.

52:01 So let's see. So checking for mibility I will scroll down. Okay. Then I will now enter my description. What I Okay, first I would create pull request Am I still audible?

53:03 >> Yeah, we can we can hear you. Well, >> yep, you're good. Great. >> Okay. Oh, I would I was supposed to enter my wallet address at that description. So, but I didn't. So, the sovereign agent, my agent said no wallet address found, cannot process bounty payments. I'm sorry. I think I was a bit tensed, but this still shows the use of the agent that the agent actually reads and understands that no wallet address was found. So, I would go ahead to add a comment with my wallet address.

53:42 I think I I also have that saved. This is what I would have pasted there. This is what I did the tax description with my wallet address. So, I'll paste that and close the the the bounty. I think the bounty number is 16 if I'm not mistaken. Yeah, 16. It's 16.

54:12 Closes 16. I just want to show how the agent makes payouts, the automatic payouts. Okay, I think I've done that. So, I should be able to see that the agent has made payouts. Let's check my repo.

54:49 Okay. Not yet. I think there's an issue. This is what I was trying to pull. Let me check the issue. The fit was passed but there was no wallet address. Okay. So I should match.

55:35 Okay. I think this is done and I should be able to I should be able to see that that there has been a payout. I hope so. Still no payouts. >> Okay, Jenny. I think we're we're we've hit the the five minute mark. So, if there's anything you wanted to kind of wrap it up with and we can we can get going. >> Okay. I I actually had have a video where I demonstrated this, but I think I missed out on adding my wallet address, which you can see that the agent said no wallet address found.

56:20 So, I tried to make a new commit and I think the I I I I don't really know what's what's going on. >> That's okay. I think the the purpose of of this build this build can be useful in the Tesos ecosystem for managing treasury and all of that because funds are and it can also be useful in development posting development bounties funds are long are locked in contracts h contributors don't need to be afraid of being paid or when they will be paid payment is automatic and I have some paid transactions here on XT XTZ and it's on presently on shadowet test net. So I think that is the purpose of this build going forward. I would if I would continue building this which I would like to do. I would also I would integrate multi-ign agent rotation. I would also integrate maybe onchain reputation scoring disputes resolution model and cross chain bounty payouts.

57:33 So I think that's all for my presentation. Thank you. I'll be open to more questions. >> Thank you Jenny. >> Thank you so much. >> Any questions from the judge panel? Who's the tech? >> I just wanted to say, Jenny, I I really liked your project when I went through it. I think it solves like it's it's hard to was initially hard to see what it actually does, but when I understood what it does, I think it's a really cool experiment. there's like it touches many different aspects like governance.

58:07 It asks like a real question. Can an AI agent rate these PRs and then issue out funds? and I think it's a very interesting like the gut reaction of many people would be oh no, we need like a human to do that judging because the AI will be biased or whatever. And I think it's a very valid experiment to like attempt to see if an AI agent actually would be less biased to who makes creates a PR and actually does a really thorough review of the code changes and stuff like that. I think it's a really really cool experiment. I really liked it. Yeah, I think >> Thank you very much. I think you should try it. Try submitting a PR. It actually works. Everything works. I think >> I know. I I checked it out. It does work. Yes. Yes. And it's a >> I think I'm a bit tensed.

58:57 >> You don't don't worry. It was great. one question I have for you is right now the the AI creates the bounties, right? >> Yes. So it decides what to what the features like what the application build should actually be and that's where I think it would be interesting to understand a your reasoning and if you considered anything else like like a like the project creator to create the bounties and then the eye to basically just just be for this one purpose to to judge the pull requests.

59:32 >> Okay. I think I have that written down as a longtime implementation where the let's say it's a DAO the the people in the DAO like everyone people in the DAO with you know the way DO works the more tokens you have the more you can be able to vote they can be able to vote what the agent actually creates. So the agent just doesn't just creates random bounties. >> Right. Very cool. I I have that in mind for implementation.

60:08 >> Yeah, I think it's a really interesting concept from the perspective of like what a Gent wallets can do to to monetize themselves, right? Both from the builder side. Hey, I have a platform. I need these PRs done almost like a open source perspective, but also other agents. So, like me or Jal's agent can go out and do handle these PRs and that agent will then have an income stream for himself. he just goes out and does PRs from a bounty type of mindset.

60:35 So, I think that conceptually you're you're you're knocking on something important soon. >> Okay. Thank you. Did you ask a question? I think I think the voice is rugging. Couldn't hear perfectly. >> Oh, I didn't ask a question. I was just making >> Yeah. Yeah. I think it's more commentary, right? >> Yeah, it was more commentary. All right, I think I think that wraps it from the judges. So, Jenny, wanted to thank you. Thank you again for presenting and and appreciate you walking us through the project and >> thank you everyone.

61:14 >> Yeah. No, thank you. >> Thank you, Jenny. >> So, >> thank you. All the best. >> Yes. All right. Next up, we've got Enoch. I'm pulling Enoch to the stage. Enoch, how we doing? Yeah, I'm good. I'm going to share my screen right now. >> Okay, perfect. >> Yeah, >> I'll pull it up as soon as you do >> share screen. Share screen window. coming screen.

61:46 >> Okay. >> Can I see my screen right now? >> I see it. I'm pulling it up right here. Boom. Okay, Enoch, the stage is yours. >> Okay. I'm just going to start like this. A friend of mine last year hired a designer and what happened was that you never saw the designer again. That's more like a classical scam. So, another instance was that now flip the case. Imagine you're actually the designer and you don't and the buyer just come all of a sudden and said that okay, he's not interested in doing the project with you anymore. So, that's something that is really a point of particular concern.

62:27 so that's what habitita is here to solve. As you can see this is the page. Are you seeing the page right here? hello please. Hello. Okay. >> Yeah. Yeah. >> So this is the page right here. Write and scroll and watch the chain enforce it. I'm just quickly going to walk down through the page. So this is just about quick information about it. Okay. You can just scroll down. Okay, this is it right here. So, I'm just quickly going to test it by clicking launch app.

63:02 And as you can see, I can connect wallet right here. Metam mask. Okay. just going to connect right here. So, now I've connected my wallet. So, I'm just going to use the first example to show you how it works. So let me just Okay, I think I just need to change this to like 0.1 and the wallet address to change the address. So let's just say just for purpose of testing it say 5 minutes. And now let me click on start.

63:38 So this telegram bot is just more of like a way to get notifications from your avita escro because once the escro is created you can get a notification right here and once the once it's approved just any event you can just get it right here. So you can see show watch address escrow you can put the escro ID right here and help. So I'm just going to click on my wallet right here to get the address. So I'm going to watch my wallet so you see working in real time.

64:05 Okay. So watch then this is my wallet address. As you can see it's watching that wallet currently. And now I see I'll pay 0.1. Okay. When they deliver the PG logo in 5 minutes. So I'm going to pass the intent. Okay. Just going to wait. So this thing works in two phases. I'm using the open5 4 mini to pass the intent from this. And what what I'm trying to say is that this particular text that is here for it to become this particular text is GB 5.4 meaning that it's going to pass it. That means that if you type something that doesn't make sense here, it's going to give an error. That's just the whole idea of it. Or let's say the wallet address is wrong. So it's going to first of all for make a request and check if the wallet address is actually something correct. That's the idea behind what I'm seeing right here. So GP 5.4 min is going to pass it to something that looks like this. and GPT 5.4. You know there's difference between GP 5.4 mini and GP 5.4. So GP 5.4 is responsible for unlocking the verdict. So let me just deploy and lock without wasting much time. So as you can see it's going to tell me to sign the transaction. Then once I've signed the transaction it's going to lock it on chain.

65:20 So as you can see it's already locked on chain and you see that I've gotten a notification on my telegram bot. Okay. So that's just basically how it works. So this particular link now I can send it to the seller and or I can come as the buyer right here and place the like paste the verdict. So let's say I'm the seller. Let me just put that put it as an instance and I come back using my link right here and I use it.

65:44 Okay, for the purpose of demo I'll just put you can see that the deadline is actually written on change. So it's not something that anybody can just go and edit or make changes to. So once the deadline is passed then what was going to happen is that it's actually going to lock it in such a way that it's only a refund that can be requested. So instead of locking the money forever what I'm just going to do okay what I'm just going to do right here is I'm going to paste a proof or a verdict of that actually did the logo. So let me just put something right here. This is just for testing but it's actually making a request. So you can see confidence proof include a PNG deliver. So once I execute the chain right now, it's actually going to sign the transaction and release the forms. So once it release the phone form, you can see the notification on the telegram bot and it's actually something that you can view the transaction and see how it went on chain. It's using the test net XDZ. So I actually feel this is all for now regarding my demo and I think I've been able to show you how it works end to end and I believe that's about it. Let me just show yeah right here. So you can just quickly ask me some few questions if maybe I need to clarify some things.

66:56 >> Awesome. Thank you for for a great presentation, Enoch. team judges, any any questions? Any thoughts? >> Thanks so much, Enoch. What do you think will be like one of the most popular use cases for this? >> I think I mentioned it in the start of the presentation. What I said is the use case is that let's say you have a Okay, I think there's a noise in Bra. Okay, let's say you have a friend. Okay, let me not just a friend. Let's say you have a business deal with someone online you met randomly and I brought an instance that someone one of my friend last year got scammed and the reason why he got scammed is that they both agreed on a price $50 and the friend that paid him the $50 he never saw the designer again. So it's not like the designer scammed him and he didn't bring the work like there was no proof that he actually did the logo that he requested for. So that kind of thing can be avoided with platforms like this that you can just send the link to the seller and the fund is no longer with you. So it's not like the spire actually control the fund fully. You understand what I'm trying to say here? So the fund is like locked on chain pending the verdict that the AI is going to use.

68:00 That's just the whole idea of the routine. So it's going to prevent the issue of buyer seller issue. So u the seller can let's say the we turn it around and it's now the problem of the seller such that let me say the problem from the seller such that the seller just decide to say I'm not I don't want to work on this anymore whereas the designer has used a significant amount of his time to actually design a logo work on it so it's going to solve I think this is a very one of very good use case for it I think I've been able to answer your questions you can just ask me any any question Okay. I hope you had heard me correctly.

68:38 >> Yes. Yes. This this was a great use case. Thank you for talking us through it. >> Thank you. >> Yeah. I can definitely see the use case from a perspective of, you know, anything in kind of like the creative world where you're doing services for someone and the trust layer is not there. smart contract is kind of like that perfect trust layer. >> Yeah. >> To to kind of resolve that. I guess what I would encourage you if you're going to continue to keep building this is to maybe try to bring some of those elements of that process more into the platform. For example, like if I'm doing a logo often time, you know, this is what could happen. They the the client sees the logo but then doesn't like it and then doesn't want to pay.

69:20 >> Yeah. Yeah. That's part of the initial plan in the you can see that there's the first place the buyer has to chat that okay, what does the buyer want? So in that particular section, let's say I'm building this for that. I'm going to include something that the buyer would have to include exactly what he wants. So once the seller upload that kind of let's say it's an image or a file. So the AI the AI is going to check if what the seller is upload what the buyer is uploading oh my if what the seller is uploading actually match what the buyer wants. So if it's m that's where the confidence score is going to come from.

69:50 So when the buyer says that he want he doesn't like the logo it's not going to make sense since this is what the buyer said he want in the first place. You understand? So it's just checking and comparing what the buyer said he wants initially and what the seller is actually providing before it actually unlocks the funds. So that's that how that how the idea is going to work. >> Yeah. Is there is there you know is there direct communication between the buyer and the seller in this example or do they have to communicate through the agent because in my mind that be the only way to like safeguard that the buyer doesn't like lie in like not receiving the logo for example.

70:29 >> Oh okay. I would say that there might be other communication. Let's say we have a communication on Slack and we are both we are talking about how to I'm looking for a designer that will help me to design my logo. So, and I created a link and I sent it to the buyer. So, the buyer knows that once he has delivered the project, he will drop the the proof right there. So, there's no room for scam or anything like that. And the buyer can freely check it on chain if the wallet address actually match what his wallet addresses. You know, the funds are actually locked on chain.

71:04 >> Very cool. Enoch, thank you so much for presenting. really enjoyed the presentation and yeah, thank >> thanks for taking the time and joining. >> Yeah, I hope >> Thank you, Enoch. Byebye. >> All right, I will bring Enoch down. All right, next up we have Kawha. Kosha mentioned that his his internet isn't so great to do the presentation. So, we're going to play his demo video and then we're going to bring him up to answer questions. So, I'm gonna pull I'm gonna share my screen and we'll just watch that demo video. this one is for slashmarket built on Etherlink. We'll get this full screen and we'll get it added. There we go. Now media add to the stage. All right, we are going to run this and then we'll bring up Kashall to to discuss Hi everyone, I'm Koshel and this is SL market. We are building the agentic D5 where the capital allocation is no longer passive. Instead of static decisions, we introduce the meritocratic decentralization where the strategies compete, performance decides and influence is earned on chain. Slash market is a dual chain arena across the thesos and the Italian where AI agents compete in real time to optimize the capital efficiency. Agents continuously evaluate the validators using the fees, uptime and risk and every allocation decision is verifiable on chain as a proof. So instead of static delegation, E strategies continuously adapt and improve. Our vision is the agentic future of the T5. Financial decisions like baker selection and the risk allocation should evolve through the competition. By combining the TESOS as the source of truth with the Italink as the execution layer, we create a system that delivers both the security and speed. Why SL market? Today staking has the two key inefficiencies. First, the capital gets locked while E comes later reducing the capital efficiency. Second, the stake either concentrates in a few operators or spread without the strong performance accountability. So users must compromise between the liquidity decentralization and the performance.

73:22 Our strategy introduces the competitive agent governance. AI agents continuously evaluate bakery performance and dynamically route the capital towards the stronger operators. Alongside this, the eel tokenization allow users to separate the principle and the eel exposure enabling more flexible financial strategies. So instead of static delegation, capital becomes the adaptive and programmable. How the slash market helps? as user deposit the XTZ and receives the corresponding / XDZ which then can be splitted in the principal and the YT tokens which cues the users the instant liquidity exposure.

73:59 Normally when delegating to bakers, users must wait the multiple reward cycles to access the ET which keeps the capital logged and limits the liquidity/market provides the immediate liquidity while AI agents continuously route the stake toward the higher performance bakers. As beta bakers receive the more delegation, governance quality improves on theos since ether link evolves through the theos governance. This also strengthens the etherink's execution environment. So users gets the faster access to eel better capital efficiency and help to improve the network at the same time.

74:33 Let's see how the slash market works in practical way. Heat. Heat.

75:26 Heat. Heat.

76:23 I designed the financial primitives that remain composible across the chain aligning mechanism design with the real market behavior. Previously I have architected and built end to end onchain applications around the RWA and the payment resili infra into it. With these concepts I have bound into the multiple hackathons. So this tells my capability to build the slash market on ether link. Looking forward for the further ecosystem support to build the slash market. Thank you.

76:57 Great. Thank you. Thank you, Kashal. We're going to bring you up to the stage now. Thank you for for submitting your presentation. Kashal, are you there? >> Yeah. Yeah, I'm there. Hello, guys. >> Stage. Thank you for submitting. >> Yeah. >> All right. >> Very powerful presentation, Kasha. >> thank thank you. Thank you. May I ask were there any specific platforms or dabs that you were inspired by when creating this one?

77:29 >> actually I'm evolving this idea from many days. this is not what I actually inspired from someone but yeah I have taken the inspiration from some applications like how the eel tokenization works from the pendal and how the eMM pools actually work there. how they are different from the unis swap style amm pools. So yeah, I'm actually studying the yield primitives that how they actually work and I wanted to actually see that how actually the predictable yield and how the non-predictable yield is actually going to come. It's actually I am revoling on that idea. I'm actually turning the incentives on the eel side.

78:13 Yeah, I'm actually basically this is mix of of my idea and taking the inspiration from some better known applications in the ecosystem. >> Thank you. >> Very cool. curious. I know I know you mentioned that there were a couple you were entering a couple different hackathons, you know, to to build the kind of the the the the support you'd need to to deploy this. how do you kind of think about the the road map for SLMarket going forward?

78:44 >> currently I can seriously tell you that I have built this into the two to three days combining the research and development mainly the research I have taken the major time of mine and the development I have actually worked through my parallel agents working upon and I have built that into the one day. So yeah this is actually I have developed the idea through the hackathon and further road map actually I think that as I evolve actually I study on the concepts in the existing MVP what I have built if I actually develop this through the evolution I can actually build further this on the ether link as I what I outputs I get from the hackathon actually >> cool thank thank you Kashal any anyone else from from the judging panel have any questions?

79:40 Okay. Well, Kashal, I appreciate you joining and thank you for sending for or submitting this and yeah, great great to connect. >> Yeah. Yeah. Thank you. >> Thanks, Kashal. >> All right. Next up, we've got our last live presenter. We've got Ho Hoyu. I want to make sure I don't mispronounce that. Hoy. we're going to bring Hoy to the stage. and then I see also shared his screen so we can add that. welcome.

80:10 >> Hello. Hello everyone. You guys can see my screen and hear my voice. >> Yes, >> you are loud and >> thank you for having me today. And right now I'm going to share my my project. Let me know if you guys can see my the the whole screen right? >> yep we can see it. >> Okay. So hello everyone. My name is Fredo. so I have been a Taylor developers for more than four years and coming to this hackathon I come up with the idea Asian link the full stack automated defy training system.

80:45 So Asian links come to shown one of the biggest risk in defy right now the issue in abate defy is that 99% of cross lake arbitrate opportunities disappear before a human can react. So normally to to catch those opportunities, you need a bot. And to use those boss, you need to give them your private key or it means the boss can control your entire wallet. And if the bot has a bug or the creator of the boss is malicious, your entire wallet can be drained. So it linking comes to solve this problem by splitting between the brain and the bank. The brain here means the AI agent will constantly search for the best opportunities arbitrate across all the its but this AI agent will not hold your private key. Instead, we use a smart contract vault using standard ERC 4626 to securely hold your funds or hold your use your funds.

81:53 And let's talk about the brain here. In the brain, we use a dual engine decision logic with with two layers. The first layer is a histic logic that will constantly scale for the best abate opportunities across the next moons. And then this layer will fit into the Gemini AI agent the the abate opportunities and the AI agent here will evaluate the mont browse following three step. The first one to estimate the slippage. The second one to factor the exact impact of gas cost on the mont. And final the third step is to generate the optimal payload that guarantee that the abate is mathematically profitable.

82:44 And so here is the demo. after playing the demo let me know if I can see I can hear the the audio of the video. not I'm not hearing the audio. Let's >> Okay. sorry. So I will share another screen which is the the YouTube screen. So hopefully everything can hear.

83:17 You guys can hear the the audio. And on the left side you can filter by chairs by race protocol by hear >> you also have filter like filter by liquinity participants period Saturday and also by admin or other users in future we will also support other users end users can create their own viles and we will go into the vinted by the admin here on the top you can see the total value the API the 24 hours you revenue the how many loos This file can I check form and scrolling down you can see the value information the view revenue generated today and you can also see the execution lock here you can see that the Gemini AI agent just approved the through host from WOS to MDC to WDC and right to WOS and this $0.2 $2 profit expected. And you can see here the transation from that from this file and you can also check the the link of the attction here.

84:30 You can also check the BS that this file can be from. And next you can I will guide you to the deposit into the file. For example, you can deposit one to the V here and it will show the two-step process. First, approve the tokens. So, I will skip a little bit.

85:03 So first I will pull the token and you can see that was successful and you can next deposit to the file here and then you can check your position here your to look and your earn token so far and also check your earnings chart here and also the transation here and you can also withdraw from the BS with all of your total value and also your earn earn reward.

85:35 So yes, so there's one final slide that I want to share. so what's next for Asian link? I'm sorry. I'm sorry. I'm sorry. So what's next for Asian link? So beside that the project is already live on mainet and I'm right now getting feedback feedbacks from users and trying to update the project. There's are also four another feature that I want to have on the agent link. The first one is that custom vows it means any users can create their own val with their own AI agent strategy and second one is running the vows in TE which is just an execution environment. it means using T we can be sure that our AI code is untambered by anyone and the second feature is that I want to also support social profile rankings. Each user can have can have their own profile with their own checking of their their phones their profit and we also have rankings to rank between the phones between users. And then next one very interesting we I want to have the crosschain arbitrate. It means we also can arbitate from other EVM chains not only so we can increase the opportunities for users. And the final one an interesting feature is from to user can describe their own b in plain English and GI agent can translate it into like deployable deployable smart contract. So that's all for Asian linking. Thank you for listening. So let me know if you guys can have have any question.

87:26 >> Awesome. Thank you so much. Really appreciate the presentation. I'll open the floor up to the judges. >> What was your kind of inspiration behind building this? The inspiration is that I see there's not many kind of like ambit automated ambit defy project in Italy. So I kind of come up with this one so that the users can have like very great user experience just try to depositing their their balance into the phones and automatically get their profit like with only one step >> when you're doing these kind of arbitrageages. What's the kind of what's the risk that you run?

88:26 >> The risk right? you say about the risk right? the risk is that right now the our AI agent logic is actually running on virtual machine on cloud. So not only the admin or me can verify that the AI logic is currently running is correct. So what I mean for the next step is actually running the v the the v here in just a sec environment. to and everyone can verify that the u the problem or the logic of aation is actually correct.

89:03 Yeah. I I have one question. when you built this did you think about putting the decision making like you have two agents right? one makes the decisions and like suggests trades basically and then there is a separate one that executes it. do you think consider putting the decision- making in a smart contract rather than off-chain database systems?

89:37 >> that's actually very good question. So I actually think about this like the decision making will be like totally onchain. So any any decision that the AIN make will be u verifiable on chain. So that actually that actually will be like work in the future and also I mentioned about the T right. So in T every execution of AI agent can be verifiable and unticious participants. to so this one is like actually in the future work so right now everything is running on a virtual machine or cloud so it's actually one limitation here >> interesting yeah I think it's cool it's it's it's it's definitely like a I'm not too familiar with the arbitrage world but it's very interesting and I mean there's definitely history in people >> yes very good money on arbitragees.

90:49 >> I I think the biggest upgrade for me would be if the decision maker would post a message onchain and then the executor pulls that and then executes based on on on that onchain storage because right now the owner of this database can >> can like literally like allow a the a trade to execute even if the decision was different. if they just change the record in the database is my understanding at least from from the from the project but I think the idea is is fun.

91:22 >> Yeah, thank you. Can I add like one another comment here? So actually the the the prom or the logic of agent is actually store on chain you can verify that on the smart contract. So besid than the logic or the code is running on a central virtual machine on ground right now rather than that the the the strategy or the problem of the AI agent is actually verifiable on chain. >> Very cool. Oh dude, thank thank you so much for joining. really appreciate the the the presentation and yeah man, thanks for taking us through.

92:01 >> Absolutely. All right, we are >> we are now to the to the final the final video and then we'll jump into the into the the judging period. >> The toughest moment. >> Yes, indeed. so I'm just going to share I'll share this my screen and we will get that going. Share screen. This is the Etherlink no code agent builder that enables users to build, deploy, and interact with onchain agents all with zero code. We have a wide set of tools available ranging from token transfers, swaps right up to complex wallet analytics. Our business model is to have hybrid open platform with usage based monetization.

92:58 Our target audience would be developers, Dows, Web3 startups and web3 protocols. Our primary revenue streams would be usage based fees, subscription plans and enterprise licensing. So our current road map looks like this. We're planning to launch in the main net and then we're planning to let users use their own API keys under the hood and we're also planning to integrate more tools for our agents. Now let's take a look at a live demo. We first enter the application by logging in with Google or connecting our wallet by using Privy.

93:35 Yes, a wallet has been successfully created for our Google account and we have to either add our own wallet or create a new wallet and fund it for our agent for making any transactions on chain. So, let's create our own wallet now. And this is our only chance to copy the private key and let's fund it. Yes, we are funding it with one XTZ and we have received it. Yes. So now let's create our first agent. We have 10 tools as of now and we are actively working on adding more tools as we speak. So let's drag and drop transfer and connect it in series with deploy ERC20.

94:26 We can create any possible logic with these 10 tools and they act as the agents onchain functionalities. So let's name the agent and give it a relevant description. We're just going to copy and paste the name itself because it's self-explanatory. And let's now chat with the agent. Yes, we are transferring a token and also deploying an ERC20 token because that's the workflow that we mentioned for the agent when we built it.

95:00 So let's now run it and yes we have the results and both of the transactions seem to be successful. Now let's view both of the transactions in the explorer. First let's view the transfer transaction. And there we go. We have transferred 0.01 FNC to the specified address. And the deploy ERC20 transaction is also successful. Let's view that transaction.

95:34 So all of these tools are contracts deployed in the Etherlink Shadowet by us. So yes, we have minted 100 of test coins. This is how we interact with the agent and we can also interact with the agent through curl requests. This means that even your agents can use the created agents in our platform as a tool. Now since fetch token price was one of the tools, let's fetch the current price of Bitcoin by making a curl request to our agent. And there we go. The current price of Bitcoin is received and it seems to be 71,000 USD.

96:17 Yes. Now, let's create an agent with AI. Why build drag and drop tools manually when we can ask AI to do it for us, right? Yes. And we have our flow. Yes. This is our AI generated workflow. Let's save this agent just like the one we created manually. And let's give it a description as well.

96:53 Yeah. And let's create the agent. It's the same process all over again. It's just that we create the workflow with AI and we can also interact with that agent just like how we did with the previous one. And both the transactions airdrop and the deposit tokens per yield seems to be successful. Yes, we can also view the transactions in the explorer. And yeah, it works successfully. So what are you waiting for? Now build onchain agents with complex agenting workflows all with zero code. Thank you.

97:32 >> Nice. Nice. Clean. >> Clean like >> very clean. Very clean. >> Pretty clear. >> Okay. >> Node tree. >> Good workflow. >> Good workflow. >> Quick chats. I think it's maybe is a little deterministic in terms of like you have to build a kind of a lot of different agents to do general like if you wanted a general agent that was able to do transfers, mints, right? Do I want to really be creating a bunch of these little agents that do specific things?

98:08 But I think overall super solid demo. >> I like the no code element. You know, I think like one of the things we found is there's so much like friction from like in to interest to onboard, right? And so like that I know one of the things that we're that one of our judging criteria is reach. So like making this easier for people is something I always love to see. I think from a UX perspective it was among the like the one the stronger ones for sure especially if someone has zero understanding of how to use a lot of these tools >> right the like I think the node tree workflow builder the fact that you can chat with it also to create the the the workflow is cool >> all right is this >> all right I think so we have reached the judging period so we are going to take a quick break here We're going to put we're going to bring the we're going to put up a little graphic, play a little music for you all, and then myself, Alex, Aaron, Jalil, etc. We're gonna we'll huddle separately and and get get get get through. It was it's going to be really tough, honestly. But there are some some really strong projects. And so I wanted to first off just thank everybody who presented, everybody who submitted.

99:27 you know we this is the first time that we've done a hackathon like this and we were honestly blown away by the the quality of the projects and the amount of people who submitted >> the amount of entries. >> Yeah. The amount of entries. >> Thank you to everyone that wasn't on the live stream today. >> Yeah. Yeah. It was also really tough to pick the finalists too. >> Yeah. There were several dozen. We could have gone for like six hours, but you know, probably probably best that we keep it a little more compact. So, we are going to be right back. So, I'm going to throw up this graphic right now. Here we go.

100:00 >> All right. Five minute. >> Five minute. >> Five minute break. >> Break.

101:59 Heat. Heat. Heat. Heat. N. I want a a a A hey a a a a Heat. Heat. N.

103:41 Hey, hey, hey.

104:36 Good. Good boy. A a hey there.

105:39 Hey, hey, hey. Happy.

106:56 Dancing.

108:11 Heat. Heat.

109:06 I want you.

109:40 Hey, hey, hey.

110:19 Heat. Heat.

110:57 Heat. Heat. N.

112:24 everybody. Heat. Heat.

114:05 Hallelujah.

114:55 Hey hey.

115:42 Hey, hey, hey. Yeah, hey baby.

116:33 Oh yeah. Heat. Heat.

117:14 Doo doo doo doo doo doo. Down.

118:00 Hey, hey, hey.

118:37 Heat. Heat. Baby, hey.

119:13 Heat. Heat. N. Heat. Heat. N. Heat. Heat. N.

119:53 Heat. Heat. N. Gary.

120:31 down. Do Hey everybody.

121:22 Heat. Heat.

161:28 We are back. GMGM. Can people hear us? >> Oh, we have so many comments. >> Yeah. Yes, we can. Oh, yeah. This is all Look at that. Boom. Okay. This was This was really tough. Like what was this? This was a lot tougher than I thought it was going to be. >> That took us what, like 45 minutes? >> Took us like 45 minutes. We But you know, we >> six jury members. >> Yeah, we had all the jury members. We had to all weigh in. and it was really tough, honestly. Like super competitive, like really welldeserving, excellent presentations.

162:04 before we get into into it, like guys, like what was like what what was the process like for you guys? Like it was it was tough because there was like it was like apples and oranges, you know, there were some that were more finance, some that were more culture and then but then also like some that were like, you know, further along than others, but you know, >> sometime like it was like just a balancing act between all the different criteria.

162:28 >> Yeah. Why? Go ahead. >> Oh, no. I was just about to add like also are the smart contracts deployed? Is it is it is there you know is it possible to actually use it right now? Is it accessible? is the technical and code detail actually reflective of what it promises to be? Thankfully, we had some really great experienced jury members who were able to assess all of those things. >> Yeah. No, glad we had Aaron and Jalil for the the real technical side. You know, we brought the vibes really really important.

163:01 >> You've been a developer since like for like 15 years, I think. >> Yeah. Yeah. Yeah. At the end of the day, it should not stop everyone from life coding because that's the point. It's supposed to be it's supposed to be ideation. It's supposed to be things that inspire you and even learning from each other like what directions can we take to bring something new. >> Yeah. Are we ready? >> Really was really really tough. >> It was really tough honestly. Like I wish we had I wish we could get out more prizes. It was it was really tough.

163:32 but ultimately we we settled. We got everyone's feedback and we set we we charted it out. We settled on on on three projects that that did a really great job among many that that that really raised the bar for this. So we're going to start with third place, the third place project which will receive $1,000. we are happy to announce that that is going to slash slashmarket. congrats Kashal.

164:05 so Slash Market is our third place third place project. and yeah, it was it was tough. It was tough, but >> what do we like about that one? >> Oh, man. >> That's so Yeah. >> Well, yeah. What I liked about Slash Market is like really I said thank you. Oh, there he is. It was it was like I mean it was like from a backend like perspective really impressive like so like all of those different contract addresses >> >> automated market maker automated market and all that sort of stuff.

164:41 >> so I I was impressed. I also think that like yeah just like solid presentation and like you know good job answering the questions and so that was great. congrats Kashelle. I see you're still in here. Big congrats. our second place project which will receive 1,500 that is none other than the no code agent builder >> which no code agent builder which was submitted by Sam Felix. that one you know we love the UX on that one.

165:20 >> and you know >> wasn't able to make the live stream. >> He couldn't make the live stream. he sent us that presentation. >> I think this one's like a general a general purpose sort of easy entry point for people looking to create you know automated workflows agentic workflows on chain and on >> yeah very very intuitive walk through and e easily accessible to people who are not exactly technical from the start. Yeah, and you know, I think I mentioned it before, but I love the the just making it easier reducing the friction to like get involved and and get involved in this ecosystem and getting involved with agents. So, you know, I think that really like that really hit the mark in terms of the reach perspective and the opportunity there. so, congrats to Sam as mentioned he and his colleague who did it. they've been in in close communication with us. They were in transit at this time, so they weren't able to make the stream, but they sent over their the demo. and so >> the final, >> without further ado, drum roll, drum roll, please. Drum roll, please. This was this was tough. This was really tough. the first place prize with a prize of 2500.

166:35 we got to give it to Arbiter Enoch. Congratulations. I thought that the escrow I thought the escrow you know capabilities there were like very like a strong utility there like a clear like compelling need and a use case. Excellent documentation throughout. I mean, live demo video, a live a demo video, a live demo, a live telegram bot. you know, like >> we were all able to use and check. >> Yeah. Like up and running and functional. You know, we checked the back end with the contract, you know, it's like that was a really strong strong entry and there was a lot of a lot of like attention to detail on in terms of like that that front end >> the front end even the front end of the landing p of the marketing site.

167:21 everything really polished, really smooth. I definitely see the use case. having run a creative agency for seven ten years, like the that initial trust problem with a new client is something that's super real. and I think that this helps address that. >> Yeah, absolutely. Congratulations, Enoch. I see you're in here as well. You're very welcome. so yeah, Enoch, Enoch, Sam, Kashall, congratulations.

167:54 we'll be in touch in terms of, in terms of those those rewards. But I really want to I really want to emphasize like every single finalist here should be really proud of themselves. it was it was we we didn't think it was going to be this tough, you know what I mean? And like I also want to encourage people to continue applying and continue building because across six people we had sometimes completely different like completely different opinions on what is sort of like the one that deserves the most. And I think it's it's good that we had like majority votes, but you never know. Like you never >> Well, people vote people for different things, right? So it's like >> I think every But I I've checked this, but I'm pretty sure that every finalist got one vote at least.

168:42 >> Among the finalist decision, which >> Yeah. So we we went through just to give you guys some context. We had the judges submit like their their their initial like one, two, three. and they, as I said, they really vary. that was on that one. and then we all got together to kind of discuss why we felt that way, you know, talked it out under and then like kind of settled on a consensus, but every single every single finalist received at least one vote from there was no like there was no one who who was who who didn't get a vote. So like it was it was really competitive. It was really tough. everyone should be really proud. like Alex said, you guys should continue to for those who the finalists who didn't end up having their their selected and also applicants who didn't end up getting selected as finalist, that was also really tough. That was that was an insanely challenging and competitive thing to narrow down as well. but you know, I I think you shouldn't be discouraged. there's opportunities to come in the in the ecosystem and so if there's a a best way to stay in touch on those things or or to be involved you know >> yeah I think as soon as you're you're showing traction because at the end of the day that's the biggest challenge in all of these things like the idea is great in concept but how do you show real world use cases and I think that's really important so I really encourage everyone to continue following TZOs following the updates there's more to be released in terms of how Teslas X road mapap is developing but really appreciate everyone getting to know getting to participate and deploying all your newly built things either on mainet or shadow net.

170:20 >> Yeah, thank you. Thank you guys so much. It's been a really a really fun process all the way through. You know, when Alex and us started this, we were just like, let's make sure we have fun all the way through. was really great to you know come up with the branding, build the site, see all the applications come in and to even the people that didn't weren't selected as a part of this finalist. A lot of those projects were really really good and interesting too conceptually and otherwise. so yeah, it's been it's been really fun. Congratulations to the winners and we we'll continue to do more more stuff like this on the now media side. Yeah, absolutely. And one thing to keep an eye out for too, obviously we're gonna be mean announcing the the winners on socials, but we're going to also be running an article this week of a recap of the whole hackathon and we're going to be spotlighting all of the finalists in it. So, you know, we'll have obviously the the the winners there, but we're also going to have a blurb, you know, like a for each of the finalists. so you'll be included in that article. you know, and and you know, can point, you know, have a have a kind of a press piece there, showcasing that. And you know, we we really I know we've we've said it a lot, but we really do mean it. Like this was significantly more competitive than we thought it was going to be. we received so many amazing projects and even even beyond the challenge of of narrowing it down to finalists. so many deserving so many deserving finalists here. so really excited to, you know, still also use our our channels to help shine a light on you even if if if if you weren't if your project wasn't selected this time around. you know, we have your, you know, we're watching. We got our eyes on you now. And so and you know, we had more than one more than,200 people tune into this. So we're right at almost 1300. So we're not the only ones who have their eyes on you now. So, I wanted to, yeah, just thank you guys for for getting involved. you know, here at at Now Media, it's all about digital culture, you know, the intersection of culture and technology. And so, there's so much exciting things happening. There's so many exciting things happening on on the AI sides, particularly with its intersection with creativity and creators. And so, stay tuned for a lot more on that front.

172:33 And Alex, if there's anything you wanted to to speak to about, Tesla CBM or anything coming up. >> Yeah, 100%. Well, again, thanks all the,200 people who tuned in today. from an arts perspective, of course, you have the arts on Tesla's account that's updating weekly on the latest news, latest opportunities. We always have ongoing opportunities in the ecosystem. So, I would just suggest following the social channels. We recently had test which featured quite a lot of announcements and in the ecosystem. So that's why I really suggest if you want to dive a little bit deeper to watch the live stream that Cointelegraph covered. the recordings are all there or reach out to me and you can learn more and dive deeper.

173:17 >> Amazing. >> I'm going to plug the thing you got this week tomorrow. >> Oh well and also hello from New York. in other arts news, we have the Superverse/Object digital art festival taking place this month for two and a half weeks. Tomorrow is Sasha Styles reception. So, if anyone's in town, please do come by in the Lower East Side offline gallery and then next week we have Osunachi solo show and also followed by a group show at the end of the month. So, please, please, please do stop by. It's going to be open regardless of the opening throughout the days. And I'm here this week. So if you want to meet up, ha.

173:58 >> There you go. >> Sounds great. Sasha Sasha and No, both legends. >> Yeah, legends. >> Absolute legends. >> well, thank you all for tuning in. if you have if if you want to if you want to go back and rewatch or revisit this this this stream, the live the full live stream is streaming on on X and on YouTube. so you can go to the now media accounts there. feel free to, you know, share it, feel free to, amplify it if you like and stay tuned because we're going to be doing a lot more on the streaming front. the now media account, interviews, things like that. and so I'm, yeah, I'm I'm, I'm really proud of, of just how this all came together and, you know, really really grateful that, so many talented builders decided to, you know, show up.

174:48 Round of applause for everyone. Thank you. >> Thank you everyone. Really appreciate it.