HSuite Ecosystem Talks — Chapter 3: Use Cases

Ajouté : 2026-05-30

[00:00:02]live.

[00:00:04]>> All right. Uh good morning, good evening, good afternoon wherever you guys are. Uh welcome to our uh chapter 3 8 uh YouTube live and uh today we have uh together a few more uh new panel on the board. One of them is uh Luis from our treasury committee. Uh Luis, if you would like to say hi to everybody.

[00:00:30]Hi guys, help to be here. Happy to be here. Sorry for my English, but I'm here to help.

[00:00:38]>> Uh secondly, we have uh Jasm here, one of our senior committee members.

[00:00:43]Jasm, if you like to say a few.

[00:00:45]>> Hi, thank you for having me.

[00:00:48]>> Uh and then uh lastly, both our founders t uh Topachi and Tom. How are you guys today?

[00:00:54]>> Start with you.

[00:00:56]>> Yeah, doing well, thanks. just uh keeping things ready for the countdown of the launch that we're going to have soon of the wallet and all the other things around.

[00:01:09]>> Yeah. And uh Tom, how are you?

[00:01:12]>> Hey guys, thanks for being here.

[00:01:14]Everything fine. I'm just uh looking forward for the new chapter.

[00:01:21]All right, thank you for those warm introductions and uh myself Brzinski, one of the senior committee members uh is your host today and uh we'd like to start today with um talking about use cases for V3. Um I know last week we were mentioning uh there's some um vibrations coming from the end someone's mic.

[00:01:45]All right, thank you guys. Uh so last week we were talking about V3 and um I think there was still a little bit of unsettlement with understanding V3's concept and I thought this week would be a really good time to share with our community uh use cases specific use cases that we can talk about and maybe uh give amplify our um understanding of what exactly can be built. So uh today we'd like to start with uh Jazzim uh and there's a use case that Jazzim um has worked on that was also on um Africa's hackathon uh which won a prize.

[00:02:24]Um it is basically a microl lending uh project called um sorry if I butcher the name uh Silsilla.

[00:02:35]>> Yep that's right.

[00:02:37]And it is a microl lending platform pretty much a in layman terms a pawn shop uh in the digital world >> and um it's built on hyera using the smart no technology provided by 8uite uh and it's uh basically a a project on chain finance and uh working with real world assets. So Jasm if uh you'd like to please uh give us a little more information about SIST and uh how you came about the idea of this project and uh what the project is a lot really about and how it's uh imple improvised with smart nodes.

[00:03:12]>> Yep. Yeah. Thanks Brazinski. I think let me just start by giving a bit of context. Uh you know um about maybe six seven months ago we were looking at what can we do with smart nodes? kind of use cases would be u would be really uh something useful because I mean we can build a lot of things on smart nodes and and we have been for the last uh two years but um you know it's important that we find use cases that have sustainable business models and that's what we what we are really focusing on so I think we came across uh a particular use case actually from an inquiry that um that that a prospective client came and asked because uh their issue was very simply that in the Islamic porn shop and I think in in porn shops in general uh because because the the the the clientele of those porn shops don't have an ease of getting into well let's just say they're not very well banked that uh they are kind of taken advantage of and in this case the uh the pawn shops charge roughly about 24% perom uh for I mean although the tenor is much shorter it's between six and six and nine months the typical APR is about 24%. So you can imagine for people at a certain strata of society they uh they actually uh getting penalized and uh having to pay high interest rates. So this is the problem statement that people came to us with and they said, you know, is there a way to to to make this take away some of the friction and allow uh us to allow these pawn shops to be able to tap some of the uh liquidity and therefore if there's more liquidity coming in uh then there'll be more competition to lend money and therefore the interest rates will go down because one of the other one of the problems was that even though a particular pawn shop may offer a rate that is say lower than 24% say it's 18% a a year. So he's more likely to get the business. If he doesn't have the cash, he can't fulfill the business. So uh c you know liquidity was a key uh proposition in in in solving this challenge. So what we looked at is how can we bring in more liquidity for uh for the industry and what we came up with was that uh the pawn shop receipts they get tokenized and put and put and taken up by liquidity pool and therefore this money coming in from the liquidity pools goes to the pawn shops and they then have money to lend. So it's sort of like uh mortgaging of their uh of their asset which is the pawn shop receipts to bring in more liquidity. So of course once it's tokenized then they you can create a secondary trading of that uh tokens. So that was the basic concept and and actually it when we as we as we got down through it and we uh built it out it became clear to us that actually this flow kind of fits with any lending whether it's microl lending or even bank lending because you're basically taking an asset tokenizing the asset and then giving it out to a group of investors and those investors are then uh trading it or or en enjoying the the returns from that asset.

[00:06:33]In the Islamic sense, one of the bigger challenges for us was was that it had to comply with Sharia principles.

[00:06:42]So Sharia because it was aimed as a Islamic product, those Sharia principles had to be enforced. And this is where we came in with the uh uh smart nodes and their their particular um advantage over smart contracts in their basis that they could enforce uh logic and rules uh on the workflow to make sure that uh we were compliant with the rules of Sharia along the along the process flow. So I think you know what the silad architecture did is uh demonstrate how the smart nodes function from a governance and compliance and make sure that the execution is done in compliance with those rules so that going forward uh the the the um the investors as well as the borrowers can be assured that they're not falling foul of the of the rules of Sharia and this is very important for And yes, thank you for mentioning we did win the first prize at the Africa hackathon. Uh we we won a we won the first prize in the onchain uh finance and real world asset tokenization track as well as as well as as fourth prize in the um in the cross track. Uh >> congratulations man.

[00:08:03]>> Yeah, it was a bit of a surprise but it's not easy to accomplish this. Not easy. Very good.

[00:08:10]>> And can you say um why using smart nodes or smart engines uh in this uh architecture leverage the solution or make the solution better uh or faster or cheaper? Uh can you explain a little bit more the advantage to to use smart engines smart engines?

[00:08:33]>> Yeah, sure. I think one of the key propositions that that has always been the major uh advantage of smart nodes over smart contracts is that you can take out the the logic from the from the contract and put it outside the outside of the contract and still run it. And that's that's a very important proposition in an enterprise use case because if we keep it within the smart contract then it becomes difficult for my changing of the logic and and that second the second reason is that u a lot of the data sources are not there are off-chain data sources and if you have off-chain data sources you can't interact with the smart contract. So those those are really the key principles that made it you know the right choice for us rather than using smart contracts.

[00:09:25]Well, you mentioned that it's um changing the logic um because if you do it once uh in a smart contract uh environment, it will be forever and with the eight suite uh smart engines you can change it, adapt it or upgrade it uh without changing all the whole thing. It is it >> yes that's exactly it. I mean I think you know in in the context that we are working on you when you talk to the customers the customers may have different requirements they may have a slightly different flow uh in terms of the way it's applied to the process may be slightly different and make making sure that those rules live outside of the execution layer makes it easier for us to adapt and uh in our proposition we also included an AI portion and that AI portion basically um took a pawn receipt and that receipt was then scan you could scan it and extract the information and then use that information to run it through the rules. Now in in in this context that that off-chain data won't really work with a um with a smart contract. We'll have to find a way to tokenize it first. I guess put it on chain then only run it with the smart contract. So it was a bit more complex than uh than if we use smart.

[00:10:52]>> So uh smart nodes and smart engines will will allow allow it more complex logics, more complex solutions. This is it we're talking about more you can do more complex flows and uh this is something very difficult to >> I think the the yeah am I also >> okay go ahead.

[00:11:15]>> Go ahead.

[00:11:20]Oops, we've lost Tom.

[00:11:22]>> I think uh Tom's mic may not be working right now. Uh while we wait for Tom, >> uh yeah, I I I have a question for Jim.

[00:11:31]Uh Jim, so regarding your um project so uh is any of this live on our mainet or is this is this still a proof of concept with test net?

[00:11:41]>> We're still working on the test net right now. Uh we're looking we're we're talking to a customer a particular um uh de Islamic development organization that have about 190 branches. So we're working with them on that and not just on the on this micro lending side but because they also depend on donations.

[00:12:03]We're also looking at how we can help them do uh give them a system that helps them track the impact for their donations. So there's it's a twoin one type project but still in discussion.

[00:12:16]>> Okay.

[00:12:18]So so one last Yeah. But one thing I wanted to add to this is that you know um I think this this this point shouldn't be uh understated in the sense that >> there are a lot of legacy systems that are web web two-based systems right and those when you go to a client you can't ask them to turn their entire organization into a totally decentralized application right so you and and and give up whatever they've had in in terms of their systems before so this is actually where the I see a big value for the smart engines is that they sit in between the the web two system legacy systems and the web 3. So when these businesses want to go from to they want to start uh applying tokenization or other web 3 rails into their current enterprise applications a bridge like smart engines really makes a big difference.

[00:13:14]Go ahead, Tom.

[00:13:25]>> You having a problem?

[00:13:27]>> Oh, yeah, probably.

[00:13:29]>> Sorry. Uh, >> yeah.

[00:13:32]>> So, >> mine will work.

[00:13:34]>> Uh, do we have Tom back yet?

[00:13:36]>> No, no, no, >> not yet.

[00:13:39]>> Okay. Well, while we're waiting on Tom, uh there is a community question for you, Jasm uh that I'd like to share and ask uh your insight on this. So, uh Helloar 6859 is asking, "As a traditional Islamic bank that is offering Sharia compliant products, how will this technology benefit their future business with a complete overhaul without a complete overhaul?"

[00:14:03]>> Yeah. Well, exactly the point, right? So because they because we can bridge the smart nodes helps to bridge between the traditional uh say core bank if they're Islamic bank with a core banking system they're not going to give that up but they might want to start experimenting with tokenized products. So the bridge between the the tokenization aspects and the and the um and the core banking system is what smart nodes provides. In fact, in fact, in Malaysia, Malaysia has just launched a uh tokenized suk bond and that has to be Sharia compliant as well. And uh that is one of the challenges that they're having because how do you you can have web 3 but you can't have it completely in the sense that in sense that a lot of the settlement still has to happen on on legacy systems accounting the um the say the uh you know if it's using if a web app is being used for mobile banking or something that's typically built on their legacy systems they're not going to rebuild all of that. So um yeah, I think the technology is very important for that uh transition period especially.

[00:15:13]>> Okay, thank you for the response.

[00:15:15]Hopefully uh we were addressed to the community.

[00:15:19]>> I don't know if Tom is back. Tom, >> no, I don't think so.

[00:15:27]>> Hey guys, can you hear me now? Oh, there we go. We're back from technical difficulties, guys. Go ahead.

[00:15:32]>> Yeah, go for it. This is not smart engine. That's all.

[00:15:37]>> No, no. I was just messy. I was just messing up with the network here to to go faster and eventually I I broke the router or something. I had to just turn it off and on. Anyhow, uh it's like the IT service. Did you try to turn it off and on again? Uh but now it works. So that's okay.

[00:15:58]>> It's typical first step.

[00:16:00]>> Yeah. I I have I have like few questions for uh Jim like um um well I had a question but I guess you guys already answered this one but um it's like um um you you you know in the shaharan compliance >> um there's like several steps that you have to take right So eventually you can also integrate smart agents in a way that AI can evaluate the pricing like for the like the no rebar rules or something.

[00:16:41]>> Mhm.

[00:16:43]>> Yeah. That actually that's part of our solution. Part of our solution is actually using AI to do the first level of pricing. Uh well the sort of filtering right. So you you you you the first level is that you receive the um the instrument which is the porn receipt call it a pawn receipt >> and that that pawn receipt then has to be evaluated by AI to see to to extract the key um fields out of it and then it will evaluate it according to Sharia principles like is the asset a permissible asset meaning it's there is a real asset uh because you cannot in in Sharia You cannot char rules you cannot um provide money or investment into an asset which is nonexisting right it has to it's not a virtual asset it has to be a it has to be a physical asset so that is permissible is also because it's not speculative it's not gambling and uh you know that that those rules have to be there first before we even go to the next stage so um and then Even the second even one of the other things we do in our in our in our flow is we take a oracle of gold price and we look at the amount of gold and we check the gold price and see if the valuation that they've given for the um for the item is is suitable based on the current gold price because maybe it's not maybe the price has moved down a lot and they've given a loan where there's not enough uh what do you want to call it collateral for the um for the for the loan. So these all had to happen uh before we even get to the point of tokenizing. So that's built into the flow.

[00:18:30]>> Yeah. Yeah. That's cool. That's cool.

[00:18:34]And I was also thinking that maybe the fact that uh I mean you can safely establish the rules and then you can also safely update them uh when something changes to be fully compliant.

[00:18:49]Like have you been thinking that you can use also the smart engines in a way that um like let's say you have a wallet which is held by those rules but then the owner can either be like a very limited uh DAO or like a multi-IG made by humans in such a way that for example I don't know an advisory board can um agree on changing the rules to make the com uh more compliant in time or something like this. Because if you think about it, if you encoded the rules in a contract, which technically could be done, then the problem is that uh either you make it upgradable and then you have the the single key, admin key and all those kind of issues like that depends on a human and a single p of failure or um or you you need to deploy a new contract. So you break all the references and such, right?

[00:19:51]>> Yeah.

[00:19:51]>> Um, so so yeah, I mean just out of curiosity, you think that this this can be automated via like the DAO engine or you see it more like an advisory board of manual signers because th this is something very interesting. Our technology aims to be like uh self-sufficient. It can do it all automatically but also it can be uh customized in the way you need. So if you want a human bridge as a final multi-seek to change those rules you can also do it. That's my point. Yeah. No I think you know Sharia there is every Sharia based project has to have a Sharia advisory board. That's that's the rules right. So, so the advisory board uh obviously in the beginning they will set the rules, they will look at your flow and then they will tell you uh what are the rules where you should implement it, what are the what are the risks that they're trying to make sure that it complies and then of course there's a continuous monitoring as well to make sure that the con that your system is continuing to implement according to those rules and as well as what you say is correct is that the advisory board maybe I think on a quote ly basis make sure that they they may they may update rules and those that rules then have to be uh carry forward into your new compliance. So how do they how do they vote on those rules? We can definitely create a dow for them to do that if that was the way that they felt comfortable about it. uh that I don't think multisync is possible as well but uh I I think the DAO will make more sense because it it supports an advisory I mean an advisory board uh type of arrangement better >> it's not about it's not moving money so the wallet is less important it's more about uh the the the governance >> yeah I see I see can I shoot one more question guys because this I really really like this use case actually.

[00:22:03]>> Yes, go ahead.

[00:22:05]>> Yeah, >> I didn't have the time to dig deep so I'm doing it live. Um I was also thinking about like the the the ownership of the assets so to speak meaning uh if you if you mean a token let let's say funible or nonfunible right then either no one holds the owns the keys meaning you might have compliance issues in the future let's say you want to do KYC you want to do freeds in case of fraud what whatever right so if you if you do with contract a either you put no keys or you centralize them or you just can't enforce compliant on the way right so are you going to use like the full set of keys uh and the smart engines in a way that that this can be of help for this kind of things I don't know if the question is clear >> yeah uh well it's a good question I'm not 100% sure how we will do it I Definitely from from our standpoint um you know we we we see ourselves as a platform provider.

[00:23:17]>> We're not we are not we we deal with licensed entities as opposed to being a licensed entity of ourselves. So I guess it's up to the client whether they want to hold keys or don't want to hold keys.

[00:23:29]>> If they they will we will have to give them both options and they will decide on their own >> what which way >> that's Yeah. But that's my entire point actually with smart engine you don't even have to I mean if they want to hold keys they can hold keys clearly it's it's their choice but if they want to make it truly like totally decentralized yet um customizable and upgradable in time they can basically set all key like supply freeze wipe KYC whatever right and then the problem is that if the platform holds the keys Then retail and investors I mean they they they hold a token that can be controlled. Right.

[00:24:16]>> Right. Right. But uh if the platform doesn't hold the keys then like the post mean compliance becomes like uh hard to to enforce you see but if you set all the keys using the smart engines and those keys will obey to the same DA which is the the um the advisory board or whatever then you can guarantee to your customers that you don't hold those keys and if any changes will go through this procedure which is cryptographically signed, provable, auditable and so on. So there's a big value there that you can leverage. You see my point?

[00:25:00]>> Yeah. Yeah. No, I I see your point clearly is I think uh you know I I think initially they my my guess is they would rather in initially they want to keep some control. So they want to keep the keys but then later they find out that keeping the keys is is problematic in its own right right who who who has the keys who has the access to the keys within an organization that's not easy so I think the value of being able to say that you guys don't have to hold the keys and the smart engines will take care of that complexity for you while providing you with much more security than if you held the keys. I think that's a great proposition for them.

[00:25:43]Yeah. Yeah. Yeah. I think so.

[00:25:45]>> Cool. Thank you, Josh.

[00:25:48]>> Don't worry.

[00:25:48]>> Just uh one qu I will come back and uh when you start to talk about it, you mention to changing the rules, changing the system, the application and you mentioned about the offchain data. Uh I was in a presentation this days and uh a person from chain link uh was talking about this that uh chain link use oracles and when you talk about offchain you are talking about oracles external oracles that you will uh consult or you will um take data from.

[00:26:30]This is the same idea when you are talking that your application could con could act have access to to offchain is like oracles. Is this the what you want to say?

[00:26:43]>> Well, I actually not I think we're trying to avoid oracles frankly because I think oracles typically have a single single point of failure. I think we've seen a number of hacks that came through the oracles uh architecture. So, we're trying to avoid the oracles. I think we're trying to what we're trying to do is rather than create the oracle, we want to be able to interact with the data without having to and without having to tokenize it first or put it into some sort of onchain u uh sort of uh um architecture. So that's again only possible because we're using the smart engines. If we weren't using the smart >> using the smart contract, we would have no choice but to use oracles because that's the architecture of a of a contract. Uh I I I I mean Tom might be able to speak better to this, but I from my technology level that that's the the value proposition.

[00:27:40]>> Okay, Tom is this is with you. What what's your take on it?

[00:27:46]Sorry guys, I just lost like for 40 50 seconds. My line is very horiz. That's the question.

[00:27:56]>> Oh well, uh yeah, in a way I mean they can be helpful but they are kind of dangerous as well. Um the problem is not the oracle, the problem is the architecture behind it. Um meaning um let's let's suppose your smart contract needs to know what the price is, right?

[00:28:21]It doesn't have any way to go query the outside world. The only way is that you you feed the price and you feed it with the oracle. So the contract has this feeding process like as a single uh source of truth, right? instead when you do it from a smart app you can query at the same time uh I don't know coin geeko uh coin market cap all the end points that you want and you become the oracle of yourself you can make the average of the prices out there you can choose your priority to pick up with if some doesn't respond you can trust the other one you know you decides what the outsources are and you can pick more than one uh to compare. So if one gets polluted then you can still have a majority. I mean there's lots of things that you can do when you just uh uh um create a layer like the smart engines which is all the web two interaction is it's it's still there. So it's not that that that you don't use oracle is that you basically use data feed that you choose and and you are capable to select. So it's something that a smart contract cannot do. You know >> you you can say that we are simplifying the architecture not using other >> we we are uh we are looking at the other direction. You know the smart contract cannot query the outside world. So it has to receive some info and usually there's one oracle that gets there and feed it. You know yeah maybe you can make a smart contract that gets feeded by different oracles and then internally decides but then you have the problem of how big it grows the database the costs and all those kind of things. So we just do the other way around. You can choose your smart app to query seven different uh pricing services. Let's say crypto pricing services and charts. And then you can decide you can choose pick the average price for your internal balancer or do whatever you want to do, you know.

[00:30:42]So it's it's it's yeah, it's like the smart engine. It is is like a different way to look at things basically you know because you are uh using the web 3 you are onchain you can be auditable you can interact with many ledgers and such but yet you have all the agility of typical infrastructures and restful APIs and external services uh paid services if you wish to use them and so on.

[00:31:20]>> Okay. Well, I think that was a really great uh conversation that we had at a very high level. Um for all the community that's uh hearing us today, these are type of conversations we have internally with one another on a day-to-day basis. So, uh really appreciate everyone's knowledge and inputs uh on on this uh on the tech side as well. Uh for anyone that doesn't know uh Jasm is has been very involved with 8uite and smart node technology and if if anyone doesn't know and they're new here uh this is not his first hackathon that they have won as well they uh did win uh for ecosphere as well for hideera hackathon 2.0 O and uh now this is our second prize for from JASM side. So the contributions uh that JASM has brought to the project have been really really um significant and we can only ask for him to uh continue on continue to win on forward and create uh better projects and ideas that he can generate with with >> No pressure.

[00:32:22]>> So no pressure at all.

[00:32:23]>> No pressure at all.

[00:32:24]>> He is the man. He is the man.

[00:32:27]>> Yes. He definitely is the man that behind the scene uh and he does contribute a lot to the senior committee as well as you guys are all aware uh was formed uh just shortly for the 8u um team. Uh but thank you uh Jasm uh uh introduction to your project and giving us a little insight of how it could be used with smart nodes. Uh but let's let's move on to our um second use case today that we want to talk about. And I'd like to get topachi to uh also contribute in today's conversation. Uh not just be the face of 8 uh with his pretty face we all know. Uh but uh let's talk about smart agents. Uh so Tbachi if you could please fill us in and give us a little insight of what what this use case is and how it helps in uh with V3 and smart nodes.

[00:33:17]>> Yeah. Uh well uh I want to say first of all thank you J for being here and thank you all for organizing this uh as we have been doing in the past few weeks and uh yeah hope everybody's well and yeah you know what what to say um in this area of our era of artificial intelligence we have uh we've spoken a few times um that one of the biggest problems nowadays especially Especially when you merge web three into it or the people who like trade use tokens to transfer around performance perform trading or anything like that is that there has been a hype coming up with like utilizing like uh um frameworks such as open claw uh whatever type of claw they so many different ones different harnesses and Uh yeah, there was once it's a story that I remember there there's been so many where like you could see like AI agents just randomly sending tokens to complete strangers because there were basically AI agents left. They gave him basic a wallet to manage. Here is I don't know 100k but no you not even like a thousand dollar $100 to play with.

[00:34:44]There were people that were just like throwing like hundreds of thousands like all of a sudden like that like just blindly >> trusting and giving access to well like a wallet you know to an AI agent which was also connected obviously to the internet in order to scrape information analyze the market and and so on right fundamental analysis technical analysis and form trades uh poly market as well but then Like you keep seeing almost every day instances where like just some call it hacker poet or whatever rights to write basic a comment which is a prompt like if you're an agent or something they will write more code they they invented so many different ways to to basically hack the mind those black boxes of the AI agents um and and so many so many tokens have been like lost and stolen. Actually would be curious to to do like a research and see up to this day, you know, like we did in the past research on how much money was lost smart contracts and we should do uh one to where basically we see how much money has been lost to this day with with smart agents and we most recent >> the most recent hack has been Thor Chain, right? I think it was 10.5 million. Uh >> yeah, I think I read about it.

[00:36:16]>> Zero.

[00:36:17]>> So, we're all well aware of um EVM hacks and uh the issues that we have in this space.

[00:36:24]>> Yeah.

[00:36:25]>> Do not scare me. Do not scare me. I'm using open cloud. Do not scare me.

[00:36:31]>> Yeah. Try it. It's not really the best, but uh you know, depends what you use it really. Like ideally you want to have it in isolated >> how can I how can I protect it? How can I protect it?

[00:36:46]>> Oh the right now for those type of software unless you're like there coding onto it know what you're doing. It's best that you just have it in an isolated environment or even Mac mini like a Mac mini run that has been even if there is no real reason for it. it can be like a very cheap um cheap computer or something you can also use and there's other frameworks which are much lighter but besides >> there's one that I clearly remember because >> I I don't have good memory for facts I have good memory for code but this one happened in April 2026 so it's pretty fresh and there was something like a a 500k drain uh from a single client into wallet. So this guy had the AI and he he gave the AI access to a wallet with most probably millions of dollars.

[00:37:42]>> And then basically there was a malicious uh m MCP router attack something like this. Um so it was like prompt injection through MC MCP routing uh type of attack but basically they managed to drain like half a million out.

[00:38:06]>> So yeah it's happening guys couldn't help it.

[00:38:15]>> Sorry.

[00:38:17]No, I was just going to tell >> I was just going to tell Lewis not to give his give access to his wallet to his.

[00:38:24]>> Yeah. Yeah.

[00:38:27]>> And how could how could smart agents and as V3 could help us uh to not be so vulnerable vulnerable about it.

[00:38:38]H yeah I want to arrive to this point that uh thanks to the V3 the the core framework of the V3 with all the validation and the set to rules uh the same thing that automation could do which is like checking the smart the smart engines do the perform checks and sign only if it's supposed to pass can also put on top a smart agent which basically You can consider a smart agent like a a wallet or more so an account that acts can act on its own as if it was a human like they're actively doing things. Uh the the things that differentiates and that makes the big fish is that there are rules defined by uh the person who enacted the the agent you know. So for example uh if you want this agent to trade doing a certain thing token and certain amount of money uh 100 a day maximum and all those type of things what basically happens is even if the agent tries to do something funny to withdraw $1,000 or place a bet of $1,000 the what happens is it will check a set of rules because the account will be bound also to the smart nodes, right? So what happens is the smart nodes uh the smart engine validators will check that the rules have been followed by the request performed by the agent and it's so it's rulebased only acts when the request and action fits the rule book basically already defined and can be obviously totally autonomous depending on also how much like somebody wants to trigger this agent every five minutes every 10 minutes what type of agent if it's going to be closed or some other ones. So that the expenses may vary as well. But the thing is the most important thing about this is that um the owner stays in control even if he can basically go to sleep safely knowing that nothing can happen outside of the the rules that have been set. Now this doesn't mean that um it doesn't mean that it's perfect. It's like 100% secure because there is no such thing as 100% security in in software uh in the world of software but um it can be as secure as the set of rules are and this is for this is in general rule for the any smart that will be developed because just like writing a faulty contract uh can cause like hacks and stuff like that writing a faulty uh validator which can could still be happening. We will see how the instances could be we we're like studying on those things. still much much more secure than than contracts and there is much less exposure several different attack vectors but still the validators must be good in the sense that like >> that's a very good things right it's like if you deploy as a agents and the rules are an empty set of rules then clearly the smart agents will be requiring to withdrawal and send the money to a stranger and smart validators would agree Agree. That's that's what you're saying, right? So clearly we offer the technology stock, but then it's it's so um customizable that it depends on the the way you use it basically. But let's say you can create a smart agent and you can tell okay you can trade uh 500 bucks every day. You can trade on those pools on those chains only.

[00:42:36]uh and uh with this spread and um when you withdraw money from your own wallet, you can only send back to this treasury, >> you know. So if you do s such a thing that that the the the agent can hallucinate can try to send the money to an another wallet that you don't control or can can try to place uh uh uh on the order book more than is allowed to but then the the the smart validator would recognize that this breaks the rules and then they don't sign. So the multic get stuck get blocked. You know >> the agent can can can go crazy but the wallet will will block that behavior.

[00:43:24]That's the thing.

[00:43:26]>> I I mean I just got a question. I mean I think this is like really important because uh we've seen that AI is a non-deterministic user in the sense that you can ask the same question or you can give it the same instructions multiple times and it may get it right or or or do what you you intend and then one time it won't do it that way. So >> yeah and also is deliberatating ignoring rules. you you can give 10 rules but maybe only seven will be applied you know >> yeah exactly that you know in that in the prompt and stuff right so you're bringing a deterministic kind of guardrails to a non-deterministic technology which is very powerful but you know it needs to be reigned in in sometimes I guess >> exactly yeah uh >> guys I have a two-part question real quick uh for for Topachi Um, >> so we're we're talking about agent wallets. Um, my question is what is the difference between an agent's wallet and an owner's wallet and are they completely separate on chain?

[00:44:38]Um depending depends on different things you know because depending on how I mean basically yes usually yes because you don't want to because if the owner wallet was the same one there could be like um you want separation of concerns basically that's usually the best practice so to speak but depends still on how you would want to deploy this agent for what type of um scope per purpose. um you know because there could be a wallet owner which may be a let's say a canonical wallet um where certain things happen and then you could deploy like let's say own as this wallet which is like the parent and then you have like children wallet which are the agents and those children wallets could be deploying like different uh chains let's say doing different things following different rules And this is thanks to our multi-chain architecture basically. But yes, I did basically are separate. They're not the same thing.

[00:45:49]Okay. Thank you for that answer. And the second part of that question is if the owner revokes the agent, what happens to the fund still on the treasury?

[00:45:59]>> If the owner what >> if the owner revokes the agent?

[00:46:05]>> Oh, >> what happens to funds still in the treasury?

[00:46:09]There's a there's a block mechanism.

[00:46:14]Um you can pause the agent if you want to stop the activity because something uh you want to change or whatever you can pause it. But in the moment you revoke, revoke is allowed only when there's no funds left because once you revoke basically the the the wallet has to has to be destroyed and so on. Yeah, for security reasons. So, uh you you you have to perform two steps. The full withdrawal and then the revoke.

[00:46:50]Make sense?

[00:46:51]>> Okay. Yes. Yes.

[00:46:53]>> I I have a question. How can you can we use the wallet tech tech u architecture to create identities for the agents and map that identity to the to the role or the permissions? Is that is that what this is doing?

[00:47:10]>> Can you ask this again just please? Is it is is this is this a way of creating an identity for the agent and then setting rules for that or permissions for that identity and yeah there >> is is that what you guys are doing that that's >> yeah exactly exactly because the the the smart agent will have uh his own personal key which which is like his own wallet but he's not activated created as a wallet. It's just a private key that he use for signing stuff, but then all the funds, all the stuff are into the multi-sequence as usual. So there's like there's never really a single point of value in in in that chain. Also going back, if you can go to the slides back because we were talking about revoking um but actually you you can I mean the slide is not super perfect but you can revoke the agent meaning you you want to destroy it but you can also revoke a specific permission. So if you are the owner, let's say your agent was able to trade on five different ledgers on all pairs, then all of the sudden you want the same agent to only trade on XRPL and SA and only on USDT uh whatever other token, you know, USDT soul or USDT whatever else, then you can basically revoke permission or give permissions you know and then we also wanted to uh create a middle layer. So um if the agent is is going to be out of boundaries instead of directly reject the action as unauthorized or not allowed. Uh we we are have been working on an approval list.

[00:49:18]Um, so he presents you with a list of options that are like border lines between rules if you want to enable this and then you can either manually approve or reject in case you want to have more like fine control over the actions or the trading or whatever.

[00:49:48]Okay. Uh can we continue with the slide presentation guys? Uh as we only have about 10 minutes left and then maybe we can do the questions afterwards.

[00:50:01]Sure.

[00:50:03]So yeah, I mean like like this slide says it's a very simple decision flow.

[00:50:08]So the smart agent does not jump directly from intent to action. And we've seen like we explained a little bit how it works but steps are pretty clear. If the agent wants to do something then there is this rule book on chain decentralized fully which uh checks and limits what it what it can and what it cannot do. And of course it goes as deep eventually will go as deep as the rules that you're going to put.

[00:50:34]Uh because always coming back to this uh this um discussion about the the validator is that like you can say like trade is doing a certain thing like but then like something happens the smart agent you will be surprised maybe you say you can only trade a 100 of a certain token right on a certain time range but then maybe the smart agent comes with a situation where okay but you did not specify that I cannot do this thing and there is like a little hole in the validator which would allow for strange behavior that you didn't intend to happen but uh for for the type of framework and with all documentation that we will be giving and such all the work that we're putting in we are very we're setting up especially for the validators and set of rules very good samples on eventually how deep this thing can go how customizable it can be like super granular and modular And yeah, this is open the validations.

[00:51:38]It's it's cool because they're not so fixed and they're very scalable and uh it's it's really great. It's really great the level of granularity that you can achieve.

[00:51:54]>> Pass the next question. Yeah. Yeah. Just a question about it. uh will we provide some basic rules if uh if I'm a not professional developer and I would like to improve my my agents and my personal agents um we will I will be able to download or to use some preconfigured set of rules to make my my business uh >> I mean at the end of the day at the end the data rules. It's just like a JSON uh it's like a sort of JSON format that is posted into the Hera consensus service.

[00:52:36]Um so th those will be things where you will find some official examples from uh our documentation for sure on our GitHub as well. uh everything.

[00:52:49]>> Yeah. But the this is also something very uh special about our architecture and design.

[00:52:58]Um, which was not really intended to be.

[00:53:02]Uh, but we like it the way it is because in the moment you decide that rules have to be transparent and they have to be immutable and timestamped and such then basically you are um writing messages into a header consensus service topic. Okay.

[00:53:24]In the moment you do that is like the the header consensus service topic become a kind of registry of rules because all the rules that all the entities have been written there. Okay.

[00:53:39]Uh the thing is if I create a very good rule for my own smart agent in order to use it I have to write the rule in the public registry. So you can read it and you can use the very same consensus time stamp to create your own agent.

[00:53:59]>> You see?

[00:54:00]>> So if if there's a a super final fine granular rules that is running a launchpad or a DEX or whatever those the rule those rules will be public. So will be available to be reused either through the smart app or either you can use the rule and build on top of it.

[00:54:23]>> Okay. And the user could could uh see it if I'm uh oh I would like to verify like when someone do with a smart contract codes I can go there and see the the rules. It's very clear because as you said is a JSON it's easier to understand for for a D be because because the way you have to think about it guys is that we are building the centralized registry. So when you create a smart app uh and then let's say you create a smart top which is a launchpad okay and then after 6 months you have different launchpads created different sales campaign and such okay every time you start a new launchpad you are creating basically a multi-ig wallet with the rule of a launchpad make sense so for every entity that that the smart app creates um accounts, agents, whatever. We we keep track of those entities.

[00:55:33]So from our registry, you can browse the smart app and you can know how many entities has been created, what the rules are. You get the link to hash scan because the rules are on data consensus service. So you can't see like the you don't see the source code but you don't need it anymore. I mean the project can be an open source so you go and you read it but if they are closet source you don't have to go read the code you just have to look at the app look at the entities like okay this is a launchpad okay this is the rule okay then I can trust it because it it's is it it has to behave this way so rules are always public and very easy to read because it's basically a JSON object in very human re readable forms. It's not like go read solidity, you know.

[00:56:33]>> Oh, this is great. This is great.

[00:56:35]>> Yeah.

[00:56:36]>> Yeah. Because today you uh if you don't know how to read a code and you just uh blindly sign something you don't know you are doing if the code is right or wrong. Very good.

[00:56:50]Okay. Let me move the the present.

[00:57:00]>> Well, we are in Brazil. Any questions?

[00:57:04]Any other question?

[00:57:06]>> Uh yeah, I I think I'll have one uh after we complete the slides. Uh I think we're going to be a little short on time today otherwise.

[00:57:16]>> Okay.

[00:57:17]Topachi would you like presentation?

[00:57:21]>> Yeah, speaking with the agents about the smart agents, you can see that there can be what shall we call also two formats uh because they can be paired with different product shapes and depends very much so from the use case, right?

[00:57:38]uh there is for example a smart agent with uh you can pair with a host function uh for autonomous worker and there is no need like of a customer interface in this case h it's mostly for like an internal domation like if you need some type of functionality for your company or anything like that nothing that is like required like that needs to be commercialized basically uh conversely there is also the one where you can basically pair with uh with a smart app. Uh because smart app usually will be the ones which can also be commercialized. Um basically >> I remember you told us about uh when the agent could buy an NFT on V3 and then he can start to to do the job. Um this is something in your your mind that uh because of course everybody that will uh create a smart app they will need to to buy an a subscription NFT um and then the the agent could could do this or the the user needs to go there and buy the subscription and then start to to work >> how um okay you go you go the idea I think the idea will be ideally eventually is to so we want to make things as simple as possible for like users to use right we want to arrive to a point like while while we were doing v2 we were targeting uh from smart contract developers to worldwide developers now we're targeting the whole world those who basically have like a wry wallet potentially they uh interacted with crypto even noninteracted but giving reasons to where you come to this one-stop shop where is our wallet let's say and you could start maybe chatting with let's call Adam or an agent where you could basically just use natural language to you want to say okay I want to start doing this thing um so I'm probably going to need the NFT subscription in this this case um I might want to do this app and then the the smart Well, the agent will tell you, "Yeah, okay, that's a good idea. I think for this app, you might need the $20 a month thing because the use case is still blah blah blah." Instead, like for this, you're going to need the enterprise and uh I suggest you to get that one. And uh from that moment, uh we would see that everything comes like as smoothless as possible. And then yes the NFT uh with the agent could also be there could also be an interaction direct from the wallet itself basically.

[01:00:35]>> I mean I as as for now that uh I mean it's not it's not open to the public yet. Um um we are already testing the from from the CLI. You can basically mint uh test net subscription, develop your smart app and so on. Uh the final flow will be that from our wallet you can buy your smart app subscription and then you send this subscription to one multisig wallet that gets generated and then you can start coding. So it's basically couple of clicks to be settled. Um and uh we are working also on a smart app uh showcase which is basically going to be online for the people to click and play with it but also the the uh the source code is going to be open to the public. So basically they can check out the showcase which offer both back end and front end. They can uh use their own cloud code or whatever AI they use to develop this smart app and then they can just test and deploy. So yeah it's going to be it's going to be fun guys. Yeah, >> that is great. Uh, so >> anything else?

[01:02:09]>> Before we close out, uh, I have two questions and then we'll close it out for today. Okay. um for either Tom or Tbachi whoever thinks they can answer this best for the community in general in a basics layman's term I know you guys have given an answer on this on a surface level but I would just like maybe an in-depth example of it like what is the simplest real thing someone can build with a smart agent today that's the first part Ah the the easiest thing that comes to mind it's uh trading box defy trading box.

[01:02:51]>> Okay.

[01:02:52]>> I mean simple but like the the the easiest and fastest maybe would be a a fund manager.

[01:03:03]>> Example. Okay.

[01:03:04]>> Yeah. Like a f funds funds manager. uh funds asset manager meaning >> okay >> you have different uh wallets in different ledgers let's say you have bitcoin you have sana you have polo you have h bar you have whatever and then the smart agents basically um every I don't know 12 times a day he can check prices and he and evaluate spreads and market condition and just do the best to preserve and grow your packs basically.

[01:03:48]So let's say BTC is going down then he sells BTC on on I don't know change now and and his swaps for soul because it's pumping up those kind of things which is not really trading it's more like um uh banks management but yeah that's very very easy and fun to watch.

[01:04:14]Okay, my my second question to that would just tie in is how many agents can one owner run at the same time? Is there a limit?

[01:04:23]>> Um, nope. depends on your like what are using also in terms of like interference for the eye but no >> yeah I mean the limit is uh your your own spending budget because clearly if you have a thousand agents running 24 age 7 day a week then it's going to cost you like hundred of thousand of dollar to anthropic or something >> so >> and also those And also all the calls obviously for like you know when when you have to do something that interacts with smart apps and for the the deterministic framework you spend money for that also.

[01:05:09]>> Okay. Well last question for JM just to involve him a little in our conversation as well. Um, Jazzim, do you uh do you think so Select could use the smart agent as its host function uh to automate the token minting and repayment flows?

[01:05:27]>> Um actually I would probably not start there. I would probably start with the um as I said we're using the AI already but we haven't put in guardrails for it.

[01:05:38]uh we're using the AI to do the first level filtering to make sure that it complies with um the uh requirements of Sharia. So maybe that's a good place um to to put in something like this for the guardrails of uh of of um of the agent that we have because right now it's not working based on rules.

[01:06:03]>> Okay. Well, thank you very much for that uh answer, Jum. Um, we are uh out of time today, guys. So, I would like to thank Jum, Topachi, and Tom as our speakers for today. And, uh, Louise, thank you very much for co-hosting with me. Um, one last thing, uh, next week we do have another live. Uh, it'll be our fourth episode, uh, this month. And um I would like to mention that it's going to be a very big uh live because we're going to be talking about our wallet details uh which is coming out eventually. I know the community must be very amped up because we have been uh we have began our marketing campaign and uh we're on day three of that. So there's a lot to look forward to next week and um we will have uh wallet tutorials as well and demos and uh TE's are very excited for this uh to to share with the public what is there and what what the potential is. Um so thank you very much everyone. Uh we look forward to seeing you guys again next week same time 4 pm UTC on a Thursday. And uh Luis if you would like anything else to say last remarks before we close it out.

[01:07:14]>> No man just excited. Yes, we got to get there.

[01:07:18]>> Okay. Well, thank you. Thank you everyone. Uh thank you for uh thank you to the community for hearing us out and those that are being >> listen to our recording as well. Uh and we'll see you guys again next week.

[01:07:29]>> See you next week. Pleasure. Have a happy weekend.

[01:07:32]>> Thank you guys. You guys all have a great day.

[01:07:34]>> Bye.

[01:07:34]>> Thank you. Bye. Take care.