How to decentralize Intents by Mounir Benchemled

Transcript

good um so I'm and founder of baras swap in portugus uh since recently uh parasa we are a decentralized exchange aggregator running online evm blockchain since 2019 and uh portugus is a new protocol that I'd like to talk to you about uh uh today and I guess you had a great talk from Adrian about intent so I would like also to try on my side to explain intent the simplest possible way uh to also convince everyone that intents are really the future um so when you guys go to a a coffee shop um you can either ask for an espresso and precise the recipe the way

the coffee will be will be made and give all instructions to the Barista which is fine I mean it happens it entails various uh steps and maybe you're an expert in coffee or like a very say coffee Guru um which works but the vast majority of people will just Express an intense they would say just I want a normal coffee and I think even if you are a coffee Guru well uh most of the time you in hurry or just you just want your coffee and you trust that The Bu sty will do the right thing and you just ask for the coffee and the bua will do the magic and providing you the best possible coffee uh that you will that you would enjoy but another example example that I like is uh when going from point A to point B so for instance

uh you want to go to a party in Dubai and there are various options to get to the party so you get a car you need to make sure you had the you take the right Road you respect the speed Li the speed limits and the the traffic traffic lights and all of that which you can do um it's a lot of steps you get more control of course when you have when you are in your car in your car sorry but in fact the your objective is to get to the party and not the way you are getting to the party so why don't you just get an Uber express an intent I'm going to the party to this address and just push a button and it will be done for you the magic will be done by the your driver and the service that's running on the

back on on behind the scenes and you get to enjoy your party you don't need to worry about how the traffic is working and how Dubai traffic is working and the flaw and all of those things um but in trading in Defi and trading in general this is how trading used to work forever even before defi it's just that the infrastructure wasn't there and but if you go to binance or coinbase well you don't tell binance how to execute a microservice or how to run a DB you just push a button and say I'm selling that token for that token or I'm creating the limit order and the trading engine is going to do the work for you and this is now happening in defi as well where you can either go to a

counterparty that's like up today or the vast majority of Dex UI where you need to load your wallet you need to make sure you have enough gas you need to make sure that the gas price hasn't changed and track all of that and or you can just push a button and um Express an intent you're selling token a for token B you're okay with the price sign a message and it's done for you the most optimal way uh we can get obviously we'll get into into into more details but that's the bottom line is you don't need to worry about the magic behind the scenes you just need to worry about your trade being executed and that's I guess for the vast majority of the squ population that they won't be dents and Define natives but also if you're a digion your goal is to really acquire

that token in order to a in into another protocol or speculate on that token your goal is not really the execution detail so I think this works for everyone not just for the I would say the average joke so benefits obviously Ian guess many of you have known and you also the Adrian talk was really amazing but we get at the same time an improved user experience that's what we aim for um to also get VII into the the masses uh we abstract all the complexities like gas uh protection against M and all of that and we get to also simplify new use cases like obviously swaps but also limit orders DCA optimization you name it so now I'd like to talk about

portugus in our take on how we can build intents but not only intent also platforms to run intents in a decentralized way so now if you look at the existing products um you can like summarize them in a user interfaces or apis uh contract or or web2 apis that allow users to express an intent sign a message and bunch of solvers behind the scenes um competing for the user and having the right to winner has the right to execute the trade on behalf of the user problem with this architecture it's working fine is great it processed billions of

dollars so far so it's really working we have a real product Market fits but the problem is that we still rely on a centralized middleman that's like a Google server AWS and all of that it's working um user is still self Sovereign uh we're still in the defi realm at its core but the problem is that it's not Unstoppable it's not non sensorship resistance it still doesn't fit or check all the boxes in decentralization I think we can do better um than that um and this is where we are introducing what we call here a decentralized layer it's one way of keeping an abstraction of not giving it yet an architecture it's kind of an intent if you would an

architecture terms um there are various possibilities like up chains and new layers or like chain link type of of networks but bottom line the idea is that there is no longer a hard Reliance on a centralized server it's a new neutral Network where users can submit their intents and that neutral Network or that what you call DL here will creates auctions and solvers will compete with the DL and um what we get here is well decentralization like this is a neutral decentralized network but also probability and not and we avoid censorship resistance I mean we have sorry censorship resistance unlike if we are running a centralized

centralized party and the platform the pro the the the the the system becomes Unstoppable this is really what you want but obviously as pragmatists we think that uh well we try to improve user experience and to improve resilience but if we introducing let's say a new complex Network that end users have to deal with let's say an up chain where users have to pay gas again we Define the purpose of user experience so this is where we think out of pragmatism we need to still keep those centralized middle men but uh they should be this they will be the same optional so you don't need to use them but you use them out of convenience let's say for instance you are quick swap or Camelot uh you want to provide the best possible

experience to your users so you provide them with your you run your own rayer you abstract all the complexity behind behind the scenes uh you provide them the best possible experience but at the same time they can bypass you if they want to if you are a sophisticated user or a high frequency Trader you can plug in directly to the DL but for the average Jo they will still go to the relayer but what they get to here is auditability and probability on after the facts like if my trade wasn't executed to the best possible way I can still go back in time and check and audit and make sure that well it's indeed happen as expected in the best possible execution really indeed happens so also yes um been running biosoft for

five years um I think uh there is the idealistic world where that we want to see but I think what's uh best we can do for for the ecosystem is to provide iterative improvements like we deliver products that has the core features that users wants and users want nice ux users want protection against MV they want the best possible execution so this is where we are breaking this down into five steps where a step one and two will achieve a similar state to existing product still with a centralized middleman but with an iterative improvement that's not enough of course so this is where we're aiming to launch this DL and breaking down the research

into multiple phases over time doesn't matter how long it take but what Ma what matters is the learnings we get from from end users and the the the product that we put in the hands of users so the first first stage we call dork it's a fancy names came from one of my colleagues who likes the Roman Roman Roman era and that will be uh well we have a product that abstracts meev gas um on layer TW it's faster than than usual product and in stage two where this is where we introduce a network of of solvers and in stage three this is where we introduce a new decentralized layer that's kind of I would say similar to more or less to AA but it's very very

very oriented into trading or transactions instead of a very general purpose which is great also it's a great great concept that we can obviously learn from but then moving forward since it's a decentralized layer uh doesn't matter what its nature but to be decentralized means also an incentive system a self- sustainability so this is where we are working also on new fee models uh there are various of options there are options we can CH pick from traditional Finance finally speaking uh but they can work very well if executing in a transparent way in an efficient way like payment for orderflow where well it's still a platform for trading platform for trading they monetize through fees and that's also uh the path we're taking obviously we're

exploring other things because it's still a new system like subscription models although I tend to think that well fees are the way to go for exchanges and exchange uh platforms or protocols uh but TBD uh also that's done in purpose if let's not do it now let's do it later and let's keep the research search and iteration and feedback especially from users in order to get there and the last step the end game um is to deliver a fully decentralized network uh self-sustaining um using the architecture that's we are looking at right now and there are also various uh questions open questions like privacy for instance and how to achieve privacy and build the best trade-off between privacy and performance which is a very

hard problem problem to solve but the hope is that from here to there is to build what we call here D like decentralized intents um in that in that division that we have uh seen so far anyways yeah thank you that's my talk happy to answer any question [Applause]