Hashing It Out
Hashing It Out

Episode 95 · 2 years ago

Hashing It Out #95-Cartesi Augusto Texeria Erick De Moura


Build scalable DApps using a fully-fledged Linux operating system and mainstream software stacks. Run complex computations off-chain, free from blockchain limitations and fees, while retaining decentralization and security. DApps with Cartesi are easier to build, scalable and more powerful..

Links: Cartesi

Sponsor Links: Avalanche and Panvala GitCoin Grant

Panvala Gitcoin Grant

HIO Panvala Tweet

HIO Panvala Community Grant

As per current prices, we made about 6200 from Gitcoin grants.

Now its about figuring out how to allocate it to the community
based "on ETHsecurity input"

Multisig Address

  • Corey - corpetty.eth (0xe027688a57c4A6Fb2708343cF330aaeB8fe594bb)
  • John - maurelian.eth (0x0Bf4C238a25b66cd869331a692dFD0322708D7fB)
  • Dean - 0x42D30BB3bf2da85dB054214Bfb4113c9a4D5A945

Gnosis Safe
- deploy txn

 ## Gitcoin Grants results
| Coin amount | USD | matching |
| ----- | ------------ | -------- | -------- |
| DAI | 157.76934383 | 158.92 | ?? |
| CLR7 | 148.26934383 | ?? | ?? |
| PAN | 7,809.855 | 418.74 | ?? |
| PAN matched | 51502.24 5693.73 | 5.88x |
| --- | --- | --- | --- |
| Total | --- | 6,270.83 | ??? |


The Hashing It Out Social Media


Donate to Hashing It Out


Hey guys, welcome back. Before we get into the show, I got the sponsorships and announcements to kind of a let you know what's going on. First Up, we have avalanche. Avalanche is a wonderful sponsor of the show. Really happy to have them. I'm really excited for the platform. It's now hit mainet that's live and ready for developers to build it all. Limits avalanche is a highly scalable open source platform for launching decentralized financial applications. It's the smart contracts platform that confirms transactions and under one second, supports the entirety of the etherium development tool kit and enables millions of independent validators to participate in full block producers. As full block producers, it's quickly gaining momentum, with more and more developers exploring the platform every day and a large, vibrate community of supporters and users ready for a new era of cryptocurrencies, blockchains and decentralized applications. And my personal opinion, it is a novel approach to doing blockchain and it's a hell of a consensus algorithm. So build on avalanche, builds out limits. Learn more at Oval Labs Dot Org. That's Oval Labs Dot Org. As for the Pan Valla Community Get coin grant endeavor we went through over the past couple months, that's over and we're very happy with the results. So, if you didn't know, you weren't listening. We as hashing it out. Participated in get coin the net, the last get coin grants around, particularly as a what's called a Pan Vala League, which you meaning that if you contributed to our grant on Gitcoin with Pan, you got additional matching, matching not only from the gitcoin grant platform but also from the next distribution of pan as it gets distributed over a certain period of time, and so, based on the number of people who donated with Pan and the amount they donated with and some other mechanisms which you can read into later if you'd like to, we get a lot more matching based on the the signaling that people give by no donating to us with pan. And to that end, we made around six hundred dollars with with with with funds who are donated directly to us from a merit of individuals, and I can't thank you enough for for contributing to us. And then an additional fifty six hundred dollars, or fifty seven hundred dollars in today's pan prices, matched by pain, matched by the Pan Vala distribution. Now that's a tremendous amount of matching from a relatively small amount of donations, right and that type of amplification of donations from signaling is what we what we tried to do and we set out and I think we were successful. So thank you all for participating. The next step is for us to figure out how to use this money and distributed across the security and infrastructure ecosystem of etherium. We're going to be doing that by engaging the atherium skewed security community and seeing what how best to allocate these fun which will update you with. So more news on that. Later we will be participating in Panbala get coin grants again, so be on the lookout for that news. But until then, onto the show. Now entering work. Welcome to hashing it out, a podcast where we talked to the tech innovators behind blockchain infrastructure and decentralized networks. We dive into the weeds to get at why and how people build this technology the problems they face along the way. Come listen and learn from the best in the business, so you can join their ranks. Welcome back to ACH it out. That's always on. The host, Dr Quary Petty. Today's cohost is Dean Eigenman. Say what's up to? What's up being? That's now you know what Dean's voice sounds like. And today we have augustill and Eric from Cartesi to join US talk about seek a seemingly novel solution to the blockchain space. Welcome guys. Funch, you do the initial introductions and kind of tell us where you came from and what you're done? Sure, so, let me talk a little bit about myself. I am graduated in engineering long time agoal with been working a soft engineer for more than twenty years different verticals, you know, full stack and doing all sorts of things. Had My own agency of soft development more recently before curtesy, and eventually started to get very attracted by, initially cryptocurrency because was playing with you know, as trying to to invest for fun and see where these things went. Then I started to get more into the tech and I started...

...to really appreciate the attack behind Bitcoin, etherium, and I found this this like revolutionary and yeah, and at about that time, I was also looking for something new to do. I wanted to have my own software business, develop something more creative, and I was falling in love with blockchain, and that's when Augusto came in and he was developing these concepts at his reserve research institutes, which is inpen it's going to talk about it later, but just trying to be short here. You know, we rapidly came up with with a company and you got very excited about taking this project forward. The idea evolved a lot from the beginning. We would delve into that, but that's basically my summary of who I am where I come from. Maybe I'll let a goost to sit a little bit, a few words about himself. Thanks for heaviness. So I'm alost. Also graduated here in Brazil, but I'm more of an academic proof file. I did a PhD and became a researcher in mathematics improbability. Well, I've always been interested in computer science and I'm kind of free software radical from the open source community. And when blockchain came and said, okay, this is really a very important piece of dispuzz of you know how to make software free, people more in control of their digital lives, and this mixed with my research and started thinking about it. I'm as a mathematician. I collaborate a lot with sergey pop of, who's one of the founders of ioda, so he was poking me all the time. I Oh, you should do some research project on blockchain. It's so interesting, so close to what I do, and that kept in my had. At some point I came up with an ideal kick, this is an idea that we should publish. I told Sergey and said no, no, you shouldn't publish it yet. You should have a company that is able to implement that and also have something you can publish, because publishes the idea in this dis field have to really have something working. And then I said no, I have no experience on that. So it loo was started looking for people who who could do the engineering part, and that's where Diego came was one of my colleagues, and Eric from the business part. So that's how we got to know each other. So what is what is cartegy? What is it? What does it do? Is it differentiated from other projects in the spaces like a high level, like the like we know, like your yeah, Quick Pitch Im in the details, right, right. So high level is that we we saw that this deficiency around scalability, particularly around scaling computation. Right, because if you want to, if you want to see block chains as a platform to run computation, which is probably the objective of etherium and and another block chain, the sports smart contracts, this platform for computations is quite precarious in different dimensions, right, especially in different dimensions of scalability. Of course, there are many problems that are technic. scalability in different ways, which chose to tackle the problem of how many instructures you can run per units of time. So basically Skilling computation. That's where we started from. But what's really differentiates us from other scaling solutions the fact that we also saw that there was another deficiency that was often overlooked, namely the ability for people to develop smart contracts or the APP logic using mainstream software, stax, or using OS resources. Right. So, if you're going to develop the appening sum, you're going to use solidity, for instance, you're going to have a language that's limiting. You don't have concepts like the file system or anything that I knows would give to you. So what if we could take let's say ans like Linux and run it in a decentralized way, so that that's basically our most, you know, unique proposition here. That why we are a second layer that's scaling computation while preserving the the security guarantees of them the line block chain, like a fium, we are also providing like a...

...full os where people can develop their the APP logic. And how do you do that? Let's like, what's the like? How do you how do you maintain the level one security while providing that scale deility, like it's there seems to be a trade off somewhere, right. So, so basically we are we move from a global consensus to a local consensus. Right. So, imagine that you have a d APP, if you're thinking about the the coin that's running on on the block chain, you need basically the consensus of everybody, because everybody's, if is involved in the ledger. Now, if I let's say, let's take another radical example, if they want to develop a chess game on the block chain, and I'm playing chess with you, right, we are the two only people that are interested in that. In that match, tenzero people running full notes, they are not really caring much about what you're doing. Who is winning that game? So in that case, when you have a local consensus, it's much cheaper to just have the people that are involved or that are interested in the in the computation or in the results of whatever is happening in there to validate or to participate in whatever is going on for that yet. So so what to do is that that basically you allow the blockchain specifies computations that are basically Linux virtual machines. They have a boot file, season root, five season drive, they have an input drive, they have a program that they're going to run and they have an output drive. So this is basically specified on chain. We can explain that for the later on. But these computation runs of chain on the on the VM, and because the VM is fully the terministic, self contained, and it also produces more pro proofs of contents of these states and the mercle tree root hash of its entire state, where you are first able to make sure that all what did they did. Nodes get the same state at the end of the computation and if they happen to disagree right then another validator can can start a dispute and prove to the blockchain that the first validate was involved in a fraud or cheating in the result of the computation. So it's basically an optimistic approach. Valid dators claimer would claim a result for a computation and if someone happens to disagree with that result, they can any chivocally prove to the blockchain that's their honest solution is actually true and inforce it on the blockchain. Are You specifically tied to atherium or just agnostics the blockchain? Well, there's nothing in our spect that that ties us next to any specific blockchain. We basically need some block in supports smart contracts. Our first implementation, what was done only theorem right. So we primarily supports in etherium. We believe that's the most important blockchain for now. So homing focus is there. But with that said, we are also running already on another layer to like manic that's also supports. So that's a cold also running on binance mar chain and we can also port to other block chains later. So if I had like rephrase that slightly, it's you have a deterministic vm that all parties involved and whatever applications being used run locally and then whatever in state that happens at the end of that dam gets gets submitted to the blockchain by some party involved or mobile parties involved, and then those who are those who care, can check to make sure that the their state matches the state that was submitted and if not, there's some dispute resolution on chain. Correct. That's and how do those fraud prooves work? How you how do you prove that someone didn't do a computation right in that you did? So we it's okay, go ahead. Yeah, this is not the reason. That actually exists. Since the S it's called some people called verification game or interactive verification. So two people. You know these all these machines are represented in the blockchain very compact way. It's just a hash and cryptographic Hash. So the blockchain is not actually keeping in in its storage all this huge state in a Linux implementation or anything.

Just a single hash and people just say how this hash evolved. So this is what can be disputed. And from this hash you can also, you know, implement financial operations like, Oh, if this is new Hash, I have some tokens withdraw rate, I can throw this money. But the point of the dispute is like saying no, this is not the Hash, this is not what represents the machine after the computation. And if one raises such a dispute, you can actually resolve it in a very short time on the block chain. But it's an interactive protocol, so us to watch the tradeoff. The tradeoff is that in case of disputes, you have some delay on finality. You're going to have to wait for, you know, one party to insert a little bit of proof, the other part to insert a little bit of proof, and this can take up to no ten steps, so it can delay the all this proofs are very short, so it's not a burden for the blockchain. But those really sit like players of the mercle tree, so they see all nicely with like complexity. Yeah, it's logarithmic on the size of the machine and on the number of computation cycles. So even if you run a super computer on this thing, you can settle the dispute very fast on the block chain. So is the goal of parts he to be a scaling solution, or is it to provide a system to be able to build more complex steps? Or is it both? It's both. It's both we want to to have some enable something like decentralized Linux servers that are well integrated with the with the bays layer, with the blockchain. Okay, yeah, if you take someone coming out of, you know, Computer Science School, they are essentially able to develop a web application or mobile phone application. I mean they may have to Google few things, but they have all the tools and it's very convenient, so they can turn their ideas into products reasonably easy. But on the blockchain space it's, you know, almost require new major just to to get there. This is one of the gaps that is very important for us to feel like we should use the same stack. You learned, you know, you liked python doing your Undergrad it's great to use python. You like the Javascript, so use JAVASCRIP. We want to offer that same environment that people learn. Typically when developing on this don't develop the centralized applications. And it's not only about choosing the language you want to use the code, but also you know, you can use this text. That's you would have on mainstream. Right. It could take a lotry that those zero knowledge for Linux and then just crosscool bio to a recently virtual machine and buds are cold on top of that, for instance. So what this allows for, ostensibly, is to understood, is to have stronger confidence in the fact that those who are participating in any given application that they're using or running the same thing. What it doesn't quite give you, it seems, is the security of the thing you're running. So when you look at like smart contracts, on on, on a theorem one and since solidity and how that then comes down to buy code, you're forcing everyone to write contracts in a very specific form that has been optimized well over time, will be optimized for security in I guess, potentially minimize the fate, like available data or computation, such that people program in a certain way and you hit develops secure best practice to make sure the transfers of transfers of transfer and nothing else and so like. It seems as though by introducing the enormous complexity of the just operating on Linux and if people deloping how they want, there's an introduction for a lot of potential security bugs. Although you can be strong, you can show confidence that no one's like doing something differently, so if there is a bug, everyone's head as the book. Yeah, definitely programming on this will not be or every day Linux development. So what currently people do a lot in some areas of development in the world is what took aback, cause move fast and break things right. So just download all the libraries you need. Doesn't matter if they're buggy. You just have to get something out there for the public and if it's buggets Buggy, you're going to fix it later. That's not how this is going to work on Linux. But there are people who work on Linux on very safe environments, right. There are...

...people who develop, you know, things that go to the rover, to Mars that run linus. Right. So, but this is very different type of development. You're not downloading any library that you think it's you know, it just recently launched in you just feel if it's so import in it. Yeah, and PM study and what you what we give in return is okay, first, the libraries are using. You can select them for with security mind, that's for sure, but also the complexity of the operating system gives you layers of security as well. Right, so to completely. If you think solidity is almost like your programming the assembly. It's very low level and it's very easy to, you know, gi give some pointer to some right to someone to modify your memory in a weird way because it's directly change in the machine. It's like if you're giving the wires and the clippers to the developer. So it's a lot of power but a lot of responsibility. On a Linux environment, if there is a bug, maybe this bug can affect the user space but not the root. So you see, every operating system has ever layers of security. If someone gets you in your browser, maybe they can spy on what you're doing on the browser, but they cannot spy your file system, your files in your folders right or if they can inspire spite those files, they cannot change how the operating system is because they have user access but not root access. So we operating system gives you layers of security that, of course, are computationally expensive to support, but that's also what we offer, the computational skillability. But it these layers of security can counter what to just mentioned. So give me as a few directions. I kind of want to take this. Do you have something for our continue? I'll take your silence as a now. All right, so walking through like as a developer, say I wanted to use CARTESI. What is the like kind of your outsimo use taste for me to develop like where? What should I be using courtesy for, and walking see the process of like building an application and the polling get somewhere. So we wish to to our current development is very well explained in autotorious. It's called the COTS or API use. What we really want is to have our main objective in the end is to have a web to experience for developers, almost to the point that if you did something for the centralized Web, you could just board it to carts in it works. That's, of course, I bit of a stretch, but it's almost to that point. So the feeling would be like you do in developing something like social media application. You have to write something for the server, software for the server, that's what actually is going to work be on the blockchain, and another software for the front end for the person to see on the mobile phone or on the browser. So you have this stick, that quotum is still while the server is going to run on this Cortesi machine. And we're not yet to the point that you can just pre tend you're not developing for blockchain and deploy to the blockchain. That's not what we offer. But that's our main objective. What we currently have requires you to still write some smart contract so if you want to use carts now, you can make do the majority of your logical Linux like the server and stuff, but you still have to write some wiring with smart contracts that we explain how to do. So you still need to know solidity. But we plan with our next release to completely get rid of this requirement. So you'll be able to write a kind of a push server. You know, this typical server that people use for facebook or twitter that receives message from like http requests, is are like package from the Internet, process them and sends them out. So this is your server. You do this logic the way you want with the libraries you want, and you do your front end and they we're going to...

...connect them for you. So this is our objective for the next release. But we're not there yet to need some smart contract development. So just to make a maybe try to make a little bit more clear that to have two stages in the development of the cards is the cards has decay, August was talking about. So the current version that we have, we can look at it as a kind of a computational wakle. Right. So think of any kind of heavy lifting, of competitional heavy lifting. That would be impossible to do. Only theorem. You could do most of your application only film, that you would do normally, but whenever you have a very competitional intensive task, it can offload it to Cretas. Right. So that's what we do now and the next version, as we move forward, to do a variation of roll ups. Then we're going to, you know, go towards what Agus is saying. Like we don't. Developers will need even need to touch smart context to create a the APP on cartasy. They could even settle payments or whatever from the layer to itself. I'm assuming that all of this is run on some what you refer to as Cartesian machine, which requires some developer resources, like hardware resources, for some of to be running these VMS that are attached to some blockchain that report to it. You tell me about like kind of the network of people for running those machines or kind of resources they need and like why they would choose to do so, like who's running is and why? So there is no hardware requirement. You can run it on any laptop or even mobile phone, although it didn't port it to to a phone. Yet it's a it just technically possible. We didn't do it, but currently you can run it on Linux, apple or windows. You can run it on any laptop and this virtual machine. Who runs it for the consensus is currently up for the developer of the the APP to decide. So, for example, if you're doing a poker game, you can say, Oh, Alison bobber playing poker. Alison bobble run their own notes or I room of this five renowned companies that will run our game and if any of them is honest, the truth will be enforced on the main chain. All you can say, no, no, this is open. Everyone can, you know, do the computation. But it's for now up for the de APP developer to decide if it's a very local consensus, a quorum consensus or of an extremely open one. That can lead, of course, to more delays if you have a huge number of people attacking with disputes all the time, but they are solved in reasonably fast time. So how do you prevent those specters in the software? So like the way around it on a Theorem is to have gas and at one point you run out of gas. So you can't write these super complex things that may take a machine forever to process. How do you guys prevent that? In Curtsy? So it's the so the inputs and outputs, these are provided by the main chain. So there is no way to overload the machine with data, because the data is provided by the main chain in the code data in the cheapest possible way that the main chain can provide. This data is called data. So it provides this data like this. So there's no way to to overflow the machine with data. But now, of course, what prevents this machine from entering a crazy loop and just consuming resources like crazy? This is for the the APP developer, to prevent. So it's not like anyone can tell anyone else. Please do this computation for me, like it is in the block chain. Right too, can create a smart contract and all of a sudden all the nodes now have to run this smart contract for you because you us. So in in ours solution, you create a machine and you tell your users. So this is my chest game. Download and run it and trust me, it's not going to enter a crazy loop. So you have to trust that the APP developer that they are not introducing crazy loops into the system. But also there is a way to limit the number of microprocessor instructions that the emulator would run. Right. So that's something that the APP, the Buillibera, can set. Yeah, we can also have a notion of guests, so to...

...say. If we want to limit that, we can also limit that. Right, which the number of cycles? Yeah, some of such a complete senter subject. But like, what's the interaction latency for two parties involved in some in some application? As if I'm seeing this appropriately, I guess it depends on what they're doing. But not for people to interactive an application they have to do something and depending upon what they're doing, it may or may not involve value transfer which they need to verify. So I guess for some actions that require the blockchain, for just for dispute, they would take a long time. But for most interactions of it of a decenttized applications, they'd be the same as it would be for Web to, so dy Wude so the first version. They would be the same as they are for the main chain. So if you're all reacting for all the actions that don't require say, withdrawals or confirmations. So if you just change in the machine and interacting the machine. You can do this. It's the same way as you do with the smart contract. But when you say, okay, now there is some money on side that machine, I I want it, there's a method that belongs to me and I want to throw it now, you have to wait for the challenging period. You cannot withdraw it in the same transaction like request and get the money right away. That would, you know, will be risky. If you could take a state that is not yet confirmed because the challenging period is not over, in just say Oh, in the current state of the machine, I have money, I want to withdraw it, that would be a problem. Right. So if you want to withdraw from it, take a RC twenty tokens, or if you want to take either out, you have to wait for the challenging period to be over. Then you can throw say, for instance, and build anchor social media application. Right, I'm building a clone of twitter. People have done this. But say you're doing the same thing right, and you're doing it on cart as a is it. Does it mean that every time I click a like, I have to interact with a smart contract in somewhere share, perform or is that can that stuff to be handled by a part of the V M that doesn't necessarily need to be pushed or acknowledged by the smart contracts? The two stages, they are right. There is a first stage where we are we implementing roll ups now. So the most basic implementation it would be let for each interaction light software has with our system, they place the the event on chain, right, so it would be subject the same latency that would have on the first layer. But also we have the plan to have aggregators. So that will be basically just like, if you think of the key role ups, are other roll ups techniques that you could have someone aggregating lots of transactions and sending them all at once to call data and then you can have better, better through puts. Okay, I'll let you. Said you're implementing roll ups. Are All working more towards optimistic ones, because they're relatively easy and socially in terms of writing. More general ideas are like zero knowledge roll ups, because it seems as though, since you're only pushing a hash based on the mercle tree of the computation of the VM, it may be more reasonable to something like zr knowledge real ups because the circuit won't be as big. Yeah, but for the disputes you need, you know, the Conte dispute is interactive and the K cannot handle interactive dispute, at least okay, with the current implementations we know. So for now it's kind of an optimistic one. Or maybe I can call it like interactive. That's something you have to marry yourself to at this point. As you as you develop it. Is that like you need to make a decision now and you're kind of stuck with it? Is it's something that can kind of go fluidly, depending upon how these bridges get built and what works. Yeah, we were very modular in this stance. So we take the minimum decisions that we have to take for developing something. So if we have to change, for example, the the consensus Proto call, it's not a huge burden, right to mean have the machine. It's very separate what the machine does to what the consensus does. So if at some point it becomes we are now, for example, discussing the possibility of having the last step of the verification procedure be done in the knowledge. So this is something that, if we decide, is doable. We can implement reasonably fest but the whole consensus a Whi and in the case, not possible yet with the current technology. So it would be overall...

...more aligned with the idea of optimistic rollapse in the sense that we start by assuming people are telling the truth and if someone happens to cheat, then someone has to present the fraud proof on chain, right. So, but this fraud proof is presented in this interactive game that the goose was explaining. For is your is your interactive verification gave similar to true bit, or is it the same? which true bit is exactly the same? Yeah, okay, for those who don't know, we've done episodes with them. So go back and check it or try and find it and put it in the description to kind of have a much part detailed explanation of what that Interactive Verification Game Is. How how to get paid? Like, why do people wow? What's The incentivization? Like it's block chain, so there's all some tide. There's some type of token economics flowing or keeping people align with certain sentences in some wher, shape or form. What is that? In part has a right. So we have we have this token that has a couple of use cases. One of them is to is to be like a Cryptophiel for the side chain that that we are implements. Didn't touch this topic here yet, but we are also implement a side chain to, you know, to scale in terms of data availability, right. So think of this side chain as a blockchain on top of the the main chain and block generators. They are rewarded by CITSI and the users and the APPS. They need to pay cityside to gain the right to insert data into this side chain. So this is one of the use cases. The other, the other one is that we are also creating infrastructure that runs the card nodes on behalf of users, because, of course, not all users want to validate computation or go into the hassle of downloading stuff. We understand that most users of the APPS will just trust some third party or thrust the the some knows that are already validating at least one of them is telling the truth and can enforce truth, truth full results. So we creating a kind of let's say ws or in few for for the cud where where people can run the APPS, can can run their computation on these notes that are provided by the infrastructure, and then this infrastructure is also portable across different organizations. So let's say if you, if you are the APP developer, you want to have at least let's say, for reputable organizations participating in the validation of your the epic. Would choose a few that's like. And also, of course, you might open for the users themselves to just, you know, plug their own machines inviti dates. But you would start already with a few reputable organizations and these organizations would charge for representing users and running competition for the the APPs by collecting a small fee. In citysie, you mentioned data availability blockchain. All roll ups like by definition, use F one for a datavailability layer and the goal of kind of least the current narrative, is to use shards and F two, once that phase hits, to ex drastically expand the databailaity layer for for these, for these roll up implementations. Does that negate your need for the like Cartesi datability block chain, or is it something that you've thought about? So the way we see blockchain development now is a I like to make the analogy that it's like, you know, a PC. If you look at computers back when they feel the whole garage in the university or army. They are very simple machines, although they were big and people started experiment with several different things. So all this special thing will make it faster for to download package, this other thing. You make it faster for running pipelines of constructions much faster. Or graphics. Now, if you look at a computer or mobile device, them actually very complex things that have very specialized hardware. So graphics card, it's very good for this, the CPU is very good for that, the model is very good for this, and this is what we also expect in layer two solutions. So we have, you know, optimistic roll ups doing one thing,...

...seek roll up to doing something else, and the same for datavailability. We expect datavailability to be very specialized in certain things and we are specializing something that is not currently in the radar of the partners solutions in layer one, which is lots of data, big chunks of data that are garbage collected very fast and don't need to be kept by the notes. So we want to have very blodge volume of very disposable data. This is something that is very hard to do in the shards because every the history has to be kept and it's hard to incentivize deletion. You know, all this data can be deleted very fast and I want to pay a little because of this. So this is what we're trying to to do. Do something very specialized in large probs of disposable data and work, of course, in conjunction with with the data availability that the firstly I provide. So in the end I believe that the developers will say, okay, pictures go to cartasis side chain because they're large. Likes go to the shards because they there's more and and it's very relevant for micro transactions. They are very important for us. They have to be more secured, so they go to layer one. But this is not something that is you know, if a glitch happens in a picture, I don't mind, but it's big and I don't pay a lot for it, so it goes to the side chain. It's a very specialized type of data bailability. So who's currently using curtsy like? WHO's building on it? What are the current applications that you're seeing being built? We release the first version of the Ska a few months ago, so of course anyone can go to our documentation borrow, they can download salt and start to build an it. At the beginning of the year we released let's seal first demo, the APP, which was a thor defense game, and that was the only, the only the APP we had implemented. So right now we are running a DAPP incubation program. Now that we finally have a have a product that people can use, we are running we're also running these in conjunction with Gitcoin and we selecting three, up to three good candidates to develop creative stuff using using contest. So we are right now, at the moment, that we are selecting these apple from this applicants and seeing who we're going to subsidize to move forwards. We are basically at the very beginning. Let's be automates, we mentioned. So can you mentioned the partnership have now that we're developing construction? Sure we are. Also, we also have a partnership with medic that's established early this year, may, around May, and now we are, you know, developing a game together. So there there's some cool stuff happening. So body, tens, have some are that are that our audience is listening to you and they're like, I could do that. I can I can apply to that program and trying to become something, to be to build a dapp on yours ecosystem. What can you tell them, like what types of applications are you hope you can get built or like the end in your mind, based on your intuition, another system works. What's a very like optimized application for curtesis that won't work on other systems? It would be very easy to imagine the centralized games. I think it's we can imagine very different kinds of games be implemented with curtasis. So it's the easiest case for us, right. So we actually have two applications that are proposing to develop games on curtasis. Now defies also a good good candidates, let's say, if you want to use some advanced math library or do some complex computation, more sophisticated defy. Of course it's also good use case of curtasis. We also have some some cool feature that to call the the hashing device, that allows people to to introspect into the state of the blockchain and extract data that would be impossible otherwise without the usage of oracles. So, for instance, you can try to get the the gas price, the average gas price over a period of time without using oracles, using our the hashing device frame. So these kind of things could help defy but they also being surprised with interesting projects, you know, proposing things along the lines of Iot. So it's quite cool actually. That's that's the the thing that we that we really want to happen, right, to be surprised and people come with ideas that we had not imagined before.

Definitely interesting to see how kind of creativity people come up with. But it's like it helps when you give them a push of the of the major advantages that you offer outside of like kind of other resources. Like I said, you try to do this on atherium and the suck. But okay, cartesis fixes this. Like what, like what is that thing as? Like you said that it's off chain computation, but productilor doesn't quite understand what that means. Like what part of their application are they optimizing by using card does it so, for instance, we let's let's take the example of the Games that developed creeps. It's a tower defense game. The the the amount of the number of events that that gamers they produce is very small. Right, they put a tower here, they put a tower. There they buy some weapon to put on a tower. So you can actually compact that that kind of a dislog from the users and store on chain and that's not going to be their expense expensive in terms of gas. Now, the amount of competition that's involved in a in a tour defensive game is just impossible to run on chain right comprises bedions or even treatments of instructions for Youtube to play a game from beginning to end, to have simulate all the the crips entering the maze and all things firing crazily, and we are doing that totally total decentralized fashion. That would be impossible to do only theorem, for instance. So that's a good example. So that we are able to do only using the card in its current states as a decentralized computational oracle. Of course, as we add roll ups and we move into a state full curtesy VM and allow the you know, interactions to help along the way, then you allow for different kinds of games like turnbase games, RPG's and other things that you can imagine. So just to give another example that I like, so with the current technology from it in actually from a theorem since it's the beginning, you could try to create some sort of outsourcing of verifiable tasks. Right. You could say, I'll have this task and I need someone to solve it for me and I can deposit some money and whoever solves the task will get the money automatically, no men in the middle necessary, for example. But the kind of tasks that you cannot source in pure solidity a very limiting like, for example, I could do something silly like whoever managed to factor this big number for me will get some money. So put the number there and if you manage to write it as a product of two other numbers, you get the price. That's a task you could completely automated, with no intermediary, no trust. So people can jump ahead and start trying to factor that big number because they know the prize is there and they're going to win. But this is a very silly task. RCT Is this? I mean? Of course factory numbers is important for Cryptography, but it's not something people can actually do. So it's a city task. Now with cartees you can bring this to a level that's actually useful. So, for example, I own a company and I I need some artificial intelligence and neural network whatever that is able to tell if a picture has a address in it, because clients are getting to the store and I want to know who's wearing a dress and who's not, and I want to take the snapshot, take through the AI and tell if it a person has address or not. So it's very mutine question people actually need in the day to day lives of business. So you want to add sourced this. Whoever is able to wield an AI that recognizes people with dress, I will pay them and I don't want a man in the middle. I want direct interaction with the developer. So you could create this task in a descent realized way. And of course the thing is going to be complex and computational intensive. It's not straightforward because you have to, you know, I have to explain how it goes. It has several steps, like first to send all the pictures and encrypt the pictures, but all of these is done in Linux, so it's not it's complex if you think about solidity, but if you think about Linux. Will have a sample of a thousand dresses, pictures. Half of them you encrypt because this is your test data for later. The others you don't think Rypt, so people can see and train...

...their their machines. So it's it's possible to do something like this, for example, take an automated outsourcing and outsourcing that can be automately verified and implement it in a decentralized fashion. Okay, that makes more sense. I like outside of like, I kind of like how far it goes beyond games. Games seems to be like one. It's don't be wrong, I love games and I think they are wonderful, like test bed for innovation and economics and doing blockchain stuff. But I'm always serious about like what other stuff can we do from a computational standpoint if you have a test bet as Baggert Lenox to play around with the stuff. So could be interesting. What else? What else is there? Is there something that I you wish I would have asked you that I didn't. Hmm, I think it's going super well. You don't have any other questions for now? Like a probably ground forever, but we can write out of time. No, no, I honestly don't have any question popping in my mind now, but so you can keep just asking. What's coming up on the roadmap, like what can you expect over the next three months. Right now we are working on our proof of steak to take it to main net. It was on test net. So the proof of stakes, because we're preparing for the side chain that we're talking about before. So the the staking, a lottery system itself is being released earlier, but it's, you know, a milestone toward the the the side chain itself. So so this mechanism is going to be available on main net. Doing you four. So that's a big one. We're also working on something like cartasis explorer, like this Ui, where people can see what's going on in the system, where they can also stake anything that would expect and exploit. Right, you, if you, if you, if machines are, you know, Curtasi is running disputes. You want to see the state of these disputes and so on, so forth. So this is something that we also developing now. And the other big thing in terms of the the core infrastructure, is essentially roll ups. Right we that's that's a very important feature that we are currently focusing. You mentioned other than deeds. Yeah, yes, yeah, we're right. Sorry, none other than these are there are other projects like the one we talked about madic know and so, but these are more it's not the core core development you mentioned store, just the side chain. Why not use other Decentrali storage options like IPFs, swarm se etc. Yeah, the IPFs is is excellent and we actually use it in the background, but it doesn't give you that databailability to guarantee right. I can send you an IPFs link that you check and it's broken. So you you cannot tell. No, none of us can tell the blockchain who was cheating. Did I send you a broken link or did you not even try? So this data bailability guarantees missing there, for example, if you take filecoin for example, it can give you guarantees that people have the data. So this is very important for storage. If you want to, for example, have a kind of decenterized dropbox for yourself, it's very important that you know people are keeping your data in spacetime. But I'm happily they cannot prove the transfer. So if I tell you I'm sending you, you know my neural network that you hired me to do, I'm sending you through file coin, I can prove that you know I have it, the server has it, but I cannot prove to the blockchain that you received it. So the proof of transfer is missing. So this is why we actually need something more like more similar, closer to a blockchain, something where the currence of some event on the block proves that the whole peer to peer network had that data available at least for that moment, for that brief moment, and you can increase the...

...moment by requesting that thing to be present in the storage for longer. Like your theorem. If something happens on a block on the Theorem, you know that it's available forever. We can get something in between. So this was available during this two days period. So you can't complain that you didn't access it. It wasn't the peer to peer network two days. So it's something between, so to say, Ip a fast which has no limit in size but there has no guarantees, and first layer that has a very big limitation side. But the guarantees are gigantic. Of essentially things less forever and for the specially if you're optimizing for like kind of large chunk short term storage, you need kind of their availability, but also like a lot of a lot of room. Yeah, we want to be able to tune this. Well, I think I'm I'm good did you have anything you want to ask? No, I'm good too. I think I've asked everything I wanted to know. A great how do people reach out? How do they how they get in touch other they learn more? Well, we have a community, few communities, on telegram. We have also a channel on this chord for for developers and people that want to learn more technically what's going on, what's supposed to build people we need to help or to build stuff on courtels. So they should go to this cord. Of course you have we are on social media in general, twitter, facebook and so on, so it's very easy to get in contact with us and we really encourage people suit to come and ask questions and get involved. All right, thanks, guys. Thank you very much. Give a pleasure good times for very nice.

In-Stream Audio Search


Search across all episodes within this podcast

Episodes (127)