Ultimate Application Layer for Ethereum

Transcript

uh hello everyone my name is nikai W I'm Co and founder of gear Technologies um and in this company we basically try to build a perfect application layer uh perfect way to develop applications for three space and uh uh obviously uh last year we deployed V Network which is kind of uh our way of how this should be developed but it's V Network itself is is a network so I won't talk about that what I will be talking is uh about our new product which is g.exe and it's basically the same stuff the same handy way of develop an application but on top ofum so and um okay and um when it kind of

um uh Sol is application platform is basically a computational platform and uh when we try to think out what would be the optimal way of doing a computation in world uh when uh and adopting zes where everything needs to be decentralized everything needs to be trustless and U obviously need to be sensorship resistance what would be the optimal way of uh developing application uh in this environment uh so for example we can uh take uh AUM State at this moment where uh thousands of AAL nodes are doing all the same computations over and over

again uh and then uh every like you send one bux to your body and then thousands of notes despite that this transaction value is very small it still get verified by billions of ethereum stakes for example uh it's it's how the things uh doing now on where we have B montic verification both montic security and on other side of the spectrum we can take uh like uh more uh more optimal solution maybe for in Comm in terms of common sense you can just uh put your web server over there and say okay guys I have this web server just uh send a request to them and and it will

provide a useful word obviously this solution is optimal in terms of uh useful work is it is doing but it's obviously not uh good candidate for uh for a secure and trustless solution um so uh there might be some kind of trick we can have to make all this stuff optimal in terms of Common Sense and still provide a value for web auditor so to speak uh and inde uh zero knowledge proofs can be some kind of way to attack this problem

um uh Zer knowledge proofs themselves can be used for many different six but for once they can be used they basically tell you if I for example take this request put it uh into this set of instruction basically program and then I get this output and uh uh is this correct or not so basically they can answer this simple question without uh you needing to uh do the verification yourself again or even trusting just just cryptographic something cryptographic statement that uh that computation where uh performed in the way they were previously announced and U if you if you get back to the example then obviously we can

just along with executing user requests also provide uh this ZK proof and everybody can be sure that yeah did this what it intended to do um uh still um uh uh still there are some problems with obviously sensorship in this G but but if if the Zer proofs are practical on real time they it can be solution to pretty much any of our problems that's why guys usually refer to this zero knowledge stuff as the K so basically once they super practical super fast um you you don't need any much

else and the zero knowledge proofs can be also recursive so for example if you have like two OB servers then uh one web server can rely on the fact that the first web server is a verification uh right and also includes verification of this stuff inside its own proof so you kind of get two proofs in one uh verified as the one one day um and so um if we kind of think but uh Z much some uh despite providing this framework for optimal execution the uh the v machine you get from this uh proof is not really optimal itself so

execution is not optimal uh instead of doing request that that just that does just one instruction you get millions instructions per every instruction you want to verify or as you want to make a proof of verification as as fast and um so uh if you kind of uh step step a little bit back and try to think about the way uh maybe there is some kind of way that we take this particular value of zero knowledge proofs and uh use something instead uh and uh this uh thing uh that called liquid taking uh indeed can be such a

substitution so liquid staking what is this liquid staking uh what is liquid staking first of all liquid staking is a s for example you secure network by participating staking and then you get instead of looking your toing forever to secure network you get derived token for the fact that you are taking uh back and this token is liquid you can use it anywhere else and so you can also use it for liquid restating and uh uh so so basically use the same tokens that you use to secure but do

additional job uh make it so that it does additional job by securing something else uh you can use half of this token for example to secure one Sy second Sy whatever it has this uh some kind of liquid security where multiple thingss uh but devised by many parts can be secure many things and uh basically we get the same work do so all the operators that uh have the this uh liquid toing rest rest uh they kind of can answer the same questions we would like them to answer like we answered from zero knowledge Prof we asking is this uh

computation correct is this correct and so on and we basically get the same value from that and the good thing also that we don't need actually any new token so basically eum token is already there we just use that and you don't need to spawn your additional token or or whatever everything is already here and there are several solutions to liquid taking one of the F most famous one is uh uh I layer um there is symbiotic um the GU doing great job at kind of Comm commoditizing the the service or serice of liquid so uh let's get back to what actually we

are doing and uh as I mentioned we try to provide the best development experience in space uh without compromising on on anything like uh all our stuff running on on the typical user Hardware so it can be run on your typical laptop uh applications deployed on our Network completely decentralized they they don't need uh some kind of centralized sequences or whatever and the development environment is all is also super friendly with a friendly you can just write couple lines of cod and application just you don't need to care about St about anything and

uh uh considering saying all of I said before uh we are doing it in a special way where we we basically treat every application deployed in gear. EXT as a small roll up of its own and and and all of these rups plus everything that happens inum they they they form some kind of rup swarm where in this thrs form they all connected by a special networking protocol and they they are aware first of all of everything that happens onium so every they can uh you can miss work to any of this applications from eum or from inside this

network and uh they basically also can communicate with each other so for example one application takes the work from from and then it can kind of commission additional work to other application and uh our networking protocols kind of uh roing all these messages uh to different applications and then we also have set of operators this operatives are exactly when we talked about staking they liquid staking operators and then they have some particular State stake on them so they can uh coordinate with each other and make sure that this particular application provided this this particular results and they also randomly uh divided um so so they

assigned to different applications so for example like three operators assigned to this particular application in random and the different uh another group of operators can be assigned to different application at the same time and they communicate with which with each other in real time so that any task that was commissioned to this kind of service uh can be paralyzed not only on Hardware level but also on networking layer and the paralyzation is not only about computation but also about storage access so it kind of shuted and good think about that is also uh now we using liquid staking

mechanism uh and we can replace it with zero knowledge proof just you know just just very fast because the verification mechanisms actually very highly customizable and replaceable uh and the key things about that is um as with Z knowledge proofs uh this this system can also generate recursive Pro of its own and uh so basically every operation that starts

from somewhere on the eum site it will be the original operation signed by the user so it's kind of user starts the interaction and then all the way until it's finished can have this recursive uh proofs uh so every every actor get some kind of stake at the game uh and GE GE the dxc itself um I mean I already mentioned our first product which is V Network and uh gear.

XC basically a successor to this product where we take all the good stuff uh we manage to make uh as a standalone Network and put everything uh like it is uh into inside theum itself so basically uh it's programmable with rust uh so it has a web assembly your machine at at the back end and um it's really very modern way of programming things uh you don't need to kind of take uh design storage or anything you just write regular program and our our service our platform uh just organizes it so in such way that uh uh it just

works forever and um obviously we do it on the regular Hardware any typical user has and um we don't need like a super Computing super Computing sequencers and all this stuff as already mentioned uh has high degree of par Iz ation and uh it basically has a level of performance available from inside theum itself so you can just send the transaction and it will spawn all the computational tasks in it and get back the result to theum uh and uh initially when we do all this thing uh uh if you can of recall

what theum is and uh I was actually started my uh ethereum U my crypto career as a by reimplementing ethereum itself and if you remember that is is referred as a you know World computer so to speak uh I would I would say like eum is World CPUs uh but uh this thing that we are doing is uh some kind of GPU and we had this original uh we have this name uh originally as a AUM GPU but kind of we decided to rename it uh to avoid confusion but the analogy is still there uh so basically you have this uh Central processor being eum and the uh gear. XZ

is basically coprocessor that can do massive parallel computation for this computer uh for the tasks that uh this processor is not really designed to do and has no capabilities to do because it's it's not powerful enough and uh we already have CPU as ethereum and now we have a ethereum plus GE and constitutes a full computer so and G solves many things that eum LS uh first of all it's obviously limited computation onum itself but it solves it not in a way

that mosts do uh that that basically spawns the separated Network where everything happens and uh you know everything need to be breached in R up from roll up and so on uh rather than that uh gear X works in the in in in synchron with with AUM itself so it can take work PS from AUM it can have some work internally as an application platform uh and everything uh and our platform makes sure that everything of that happens in synchrony and this basically leads to lack of liquidity fragmentation that we see with most throwups and um also because we are

modern platform we can use modern techniques uh move faster modern techniques for example we can use prec confirmation where you can where user really has very uh uh very good uh ux UI uh stuff where you can get just press the button and gets its transaction just in the second already confirmed by the network and keep it's a highly responsive experience okay um yeah uh thank you for listening me who stayed this days see all