Unveiling Ethereum's challenges: Our Approach to Innovation - ETHGlobal London 2024

[Music] um Okay cool so this presentation is about the boundaries that n is presenting at the E Global London hackathon uh for context I am swapel I work at nethermind I'm a researcher and I work in some in some products as well uh to begin the boundaries that nether is offering we kind of first need to understand what nethermind is uh and nethermind in in comes in different forms and different ways dep on how you're looking at it but at core it's scaling ethereum and looking for a more fair transparent and open Future uh so another one works in scaling ethereum in all forms in securing ethereum making it faster in making better bringing more people to it and just opening the channel ways for everyone to accept ethereum and kind of bring the E technology to everyone uh and in that endeavor what it means is that we partner with lots of companies that work in the ethereum ecosystem for example we work with eum Foundation we work with igl we work with stocket Flash B Lio worldcoin any other company that you can really imagine is accelerating ethereum values uh we work with them and our bounties are just a representation of our work that we do with these people uh and we're promoting use of these Technologies um and hopefully see some cool novel ideas from you guys right oh first uh little question uh I'm always happy to take questions in the call and or in the in the meeting and wherever you are come to the E come to the hakatan booth and I'll be there helping answer questions cool so one of the technologies that we really believe in is z knowledge uh Z knowledge allows one to prove execution of a program to an a verifier without the verifier having to re-execute the program um I mean blockchains are using ZK everywhere now ZK V evm ZK co-processors there's ZK for privacy um and we have work with quite a few solutions that use ZK uh there's Starks which is the base ZK system of starnet uh there's a lot of technologies that are looking to do proof avigation and we have interest in proof aggregation as well because we believe it acceler is ethereum um and there's privacy Solutions um one could call World coin a privacy solution and it where it bridges that your identity from the people who can verify your identity um and we promote everyone to use um privacy Solutions and just Z technology for exerting ethereum um these are some challenges that we've put out for using Z technology but you're free to choose any solution that you like and there's more resources at the QR code below which links to a more vital discussion of how what you can use Z Tech for but some some easy implementations of Z that you can do is Implement a novel ZK publication there's always new advances in ZK research and there's Nova there's hypernova and there's every day some new stuff is coming out if you have an interest in implementing it come please talk to us we have ZK researchers who will help you understand the the details of the paper help you go go implement it and yeah achieve new novel Solutions um then there's proof agregation which is in line with the vitalic vision of uh Singularity where everything is just making a z proof and everything just settles onto at one point uh but to do that we need to aggregate proofs from everything a rollup my Excel sheet my SmartWatch everything needs to aggregate a proof and everything settles on ethereum uh and we have some ideas on how to implement proof aggregation um so yeah if you have interest in that come talk to us and as always educational tooling uh ZK can be dark magic uh but if no one understands it then there's going to be lack of social trust in it so if you have an um interest in implementing tooling that makes it easier for other people to understand Z we also will be interested in that uh then there is Maxim extractable value uh and maximum extractable value me just is an inherent property of uh blockchains and more so I guess distributed systems um it's this process where you can strategically choose what transactions go in in which order and you can change the properties of the block itself um there's a lot of cool stuff happening in me um there's PBS and MEF boost which is an idea of separating the process of block building from proposing uh which leads to better me uh alignment there's base proms which is Goos to move the role of sequencing rollups onto ethereum itself uh while still having the same properties of like 100 millisecond sequencing you can still I believe this is kind of the next wave of directions of the ethereum aligned rollups and if you have an 8as we have people in the team working on basb confirmations and also on implementation so please come talk to us uh and then the sve uh that is I guess not one way to define sve but in general it it is a a magic sgx box that lets you to Define any uh me strategy and executed trustly um we work with flash BS on this one um so if you're interested in implementing new Noble Solutions like novel auctions for amm or Landing pools please come talk to us with that as well uh there are some of the bounties that we have up for me challenges is like pricing of pre confirmations and block Futures the question really is how do I how do I calculate something that I don't know yet and it's kind of look into the future pricing it back if you have interest in economics or any of those please come talk to us uh you see two QR codes here the one on the right is um for developing swaps which are apps on sve uh and we've collaborating on flash B that one so there are some Inspirations from The Flash B Forum that you might be interested in um and yeah on the left is a more detailed list of what challenges we are putting up for uh for for me then there's data availability uh just yesterday was a big day for data availability Denon went in 444 is live on ethereum now and we all are seeing the amazing effects of it like sub send transactions uh but there's a long way to go we all know that um uh there is valadium and violations which vitalic was talking about yesterday where you s the data all offchain but you just prove the competition onchain and that also has merits where everything now is settling on ethereum but it's scaling and like like a centralized server uh sort of The Best of Both Worlds but some properties that are a bit different there's lots of output called da Solutions I da NDA um but there is no concrete analysis of which one's better or what specific properties they offer uh if you have an interest in like setting the setting some numbers up to it like setting some some comparative analysis or an SDK that lets roll up shoose between different uh da layers that are given time based on which one is cheap that would be a great solution in my opinion for AO protocol uh da Solutions um right uh and also these new DA Solutions are actually lacking a lot of tooling and dashboards where there's not a lot of transparency what's in the blob what's in the commitment and I think there's a long way to go in like educating users of what blobs really do and what these external ders do and how can we make it better for everyone uh then there is user security enhancements and this is kind of in line with ethereum's uh 2.0 Vision where we've been building ethereum and we've been working really hard on making ethereum be able to be the build the next build the platform that we want we want to have a decentralized internet and we're getting close to that and now it's kind of it's of time to make sure that our security is also lined up with that so account abstraction and all these other Technologies help a lot towards this direction um then on on stocket there's we have kyang which is a safe by default language you do pro do do provable computation but also the language is organized in a way where you cannot leave um unsoundness errors in your program um and there's a lot of tooling to be built around this uh I mentioned carang already there's pimo which is an account obstruction SDK um which we can use to build a lot of safe tooling for users uh there's also slashing protection software with with staking and reaking On The Rise I think it's an important uh important measure that we take to protecting solo stakers from slash thing because we do not want to make we we want to make sure that they can participate in this new economy but we do not want to put the risk of reaking and on them so maybe there's some solutions we can do in that direction uh maybe sgx or or there's other directions if you have more interest in this come please talk to me at the booth um yes um there is rip 7614 which is adding stack Trace to each execution which really like enhances user security and also like the user experience uh imagine going to the blockchain and getting a stack R back for a reverting transaction that really changes the Paradigm and we no longer stuck with the with the old tag of oh it's web 3 and like it's just not the same uh same experience of developing we can kind of retain those values even while being decentralized um then there's also modules in wallets where I can extend my wallet that hey only sign if U if the value is less than $10 because this is my safe account I share these keys with my friends but I don't want to like just spend more than $10 on this account right uh blockchain infrastructure tooling um we've been building all these amazing da layers and rollups and I lay avss but all of it is in fact running some some kind of node in some server somewhere and if you don't make it easy to run these nodes and manage these roads we will just have a systemic risk at our hand so we need to kind of make sure that we make that easier um on this effort there is I abss Obel DVT l z dvn all these new networks which really which really accelerate ethereum and really let us push the boundaries of what we can do but all of them require tooling and orchestration and devops and um to make sure that we don't have the systemic risk we need to make sure that we have a stable tooling stable guides and really good uh just infrastructure ready for any solo Staker to participate in this economy um so some some solutions that we're looking for is like some tooling and dashboard maybe some oneclick solutions to deploy de Loy like a layer layer zero dvn Network um maybe a one click tool link to deploy some igen dnos and make sure that they don't can't be equ equivocated um yes so come talk to us at our booth again as I mentioned because we have some good ideas on this um right uh the slide seem to be messed up there um okay cool well there's a kind of struction and stocket is a layer to network which is supposed um a section natively what that means is there are no EA accounts there is no like a legacy behavior on stocket you can only do cool stuff in some sense you can do multi calls where the whole pattern of approving some spend on on an erc20 and then spending it can now go because you can just do that in one transaction and you don't have like like multiple infinite approves left on money apps and which can later be exploited and which we've seen in the past as well um in fact there's some suggestions to even remove the approve uh function from uh erc20 on St because you could you can just merge that into the transfer um there's pimo which provides an SDK for account abstraction and there's some really good ideas what we want to do with account abstraction uh maybe we can do something like hey I only want to do transactions if I send an email people know how to use email and we can now connect the email to their smart contract account um hey I only want to spend $10 through my email and I can now program that and the users like my mom can be safe uh in this welcon ID right this is like a one of the big protocols in Pro of personal area um and if you ever want to build this kind of reputation systems and you know like more social organizing on the blockchain we kind of need some proof of personhood protocols and welcon ID is a great Direction in that in that in that direction uh and there's so many use cases to be built with welon ID like hey I want to sign in with my welcon ID scan my addes no one else can sign into my account now um so some of the IDS oh right sorry that's the wrong audit yeah some other um bounties that we have in the cont action track are visualization of the bundles and solvers there's a whole n network like now doing this intense solving uh but we need some some light in the St Forest I want to see which is there uh solver centralization is there only one solver just footing all of us if so then there's maybe some game theoretic incentive for them to like at one point do a massive attack on all of us um uh infrastructure to make cross chain um control of accounts easier um right now in this world what we have is an account on each rollup or each Network and I have to control them individually through that network but what if I had one account on ethereum uh which control all my accounts on all of the networks and I could just sign once on ethereum and they would just do everything everywhere else U so there's a lot of tooling to be built in that direction with account obstruction um infrastructure to manage cross chain me uh with this idea of this chain abstraction and sve coming around doing cross transactions there is undoubtedly going to be cross me and the users should get protection from that they should be able to maybe get some me back Express how they should be MD um because the space is open and there's no Clear Solutions here so please if you have ideas there come talk to us um and obviously there's a cont section modules where we can have novel pay masters I can now pay in ERC 20s I can pay with new tokens usdc or whatever whatever I like and most importantly I think that I something I'm excited for is recovery mechanisms today if I forget my private key on an EA I've lost it that money is never going to com back we have some ideas of like social recovery where I can get an admin like my myself or my different account but I'm thinking what if I can just do a forget password and recover my account through an email uh that would really accelerate like etherum adoption in my opinion uh and that's an easy challenge uh so if you have any ideas for that come please talk to us metaverse I mean there's a whole company devoted to it now um metaverse is now like the blend of creativity of this this space is are bending we have apple Vision Pro which is just really changing the world we have ai coming along so there's so much cretive creativity in the space to be explored uh and blockchain should really accelerate it we should not like shy of it um some solutions that we have in the space is darket which is a layer two Network which offers like incredible compute you can do so much computation and if you want to build a new future of like metae we should have it on chain and we need loads of comput for that so St is the perfect solution for that uh we also work with improbable which is a gaming company and they're designing like new immersive worlds they're designing new tooling to make um to make metaverse easy to make meterse adoption faster um some ideas that we have is like um creating Dynamic evolving environments using AI or like generative AI but I land into a space and just maybe like based on the number of people there are in the space it changes the space Maybe Bas it's listening to what you're talking to and it's just on real time changing the space um some AR Solutions where I don't know my presentation should probably be in AR and not in this real world um enhancing gaming experience through stockness capeabilities uh good example that I would suggest is Randomness it's really hard to get Randomness but we have some good solutions for Randomness on stocket using vdfs so I would encourage the use of that in some game generation and map Generation Um then there's institutional um and I kind of see this as a the old world and we kind of need to make sure that we bring the old world along and impart our values to the old world so it's kind of I in some sense I see nether mind's uh goal to bridge the between difference between the old world the dra fire and the new world and kind of impart some of our values to the old world and bring them along and not just build a silo of our own um there's so many technologies that that we've built that are so useful for the old world and I can bridge the gap seamlessly um there's stocket again institutions need lots of compute uh they they need it quick they need it fast and they need it cheap stocket with stocket you have some amazing computation available to you maybe some multicolor simulations some derivatives some Financial uh assets that you just can't do on ethereum right now can probably be enabled by stocket again pimo uh it enables people to have like a programmable experience of the blockchain um um for example a traditional Finance asset manager probably wants some multi on what assets they can use maybe they want escalations like oh I see too much of this asset being used some manager should step in and we can enable use those um real enforcing those hierarchies through Primal uh this MEF share where because again as I mentioned Mev is just an emerging property of distributed systems uh but there's a big problem with Mev in the institutional world but maybe we can brid the Gap through MEF share where MEF share shares the me that was extracted from a user back to the users themselves um some other solutions that might be interesting for you is like onchain kyc AML there's like ZK Tech so you can make a privacy preserving as well uh tokenization standards um how can we bring the old assets on chain uh how can we mix them cross chain how can we we enable like people to trade them with the ZK kyc uh proof of personal solutions for institutions how can we make sure that the people who trading these assets are safe and not like going to use them for for wrong reasons maybe some proof of personal solutions would help us there um and uses of suave because now we can give sve uses sgx or theed execution environment which makes sure that we can have a a quick um correctness of computation and institutions love that uh actually all of us love that so we can build some solutions of auctions and some Fair auctions using this thrusted Hardware um cool then the popular one uh staking and reaking and I mean there's so much happening in raking world and but also there's so many risks that raking beings um I mean the partners that we work here in the section are like aan and Lio um and I'll focus on some the solutions that maybe we can look in the staking reaking World um I mean there's always the first one which is new reaking use cases there's ever evolving speed of extending eum security to a new network to a new use case um and I don't think we've actually seen all of them and I expect that I will see some cool ones at this hackathon um but also dashboards for IG lay over or DVT um what I mean by that is IG lay brings a lot of risks and we need to make that risk apparent to everyone so we should probably have some public stash boards showing like a overlap of validators across I abss showing like um uh liquid staking centralization on abss for example if there is a new liquid staking protocol that is restak a lot but there's some systemic risk with the liquid staking protocol if if that's slash then the whole system collapses uh and one of the ways of catching this on is for seeing this making sure this is in front of everyone's eyes and dashboards and tools like that in my opinion really help there um and also there's reaking related eips uh what we mean by that is reaking on by itself brings a lot of risk to ethereum uh um but and we can make we can minimize that if we uh change etherum to specifically be safer to that for example when you get slash on IG layer through Beacon chain uh as in when your beacon chain stake gets slash on IG lay ethereum Beacon chain is not even aware of the slashing it just thinks that you're running perfectly fine because your slashing is applied when you withdraw from the beacon chain so maybe if you can inform the beacon chain here you will slash through icon layer and the beacon chain can react right it maybe like it rases the rewards your proposal a fewer times um and probably there's probably some other changes that can be done to the beacon chain design to support reaking and staking use cases um so if any interest in research areas please come talk to us we have a whole Suite of research team that are loving to accelerate ethereum in the research space then there's Ai and blockchain um I mean it's emerging like we just can't stop it so we have to make sure that we align it uh so I think there good time that we do some ethereum AI alignment and make sure that AI works for us and not against us uh some sol Solutions in this space is Giza Tech uh Giza Tech is provable AI models where you run the AI in a zikas snark or zika Stark in this case it's a stark and you can prove the execution of an AI model like for example right now when I ask open AI hey what is what is the solution to this prompt it just gives me an answer back but I don't know if open a is actually running gp4 they could just be running gpt3 or gpt2 or it could just be hardcoded strings for all I know but if they run it in a in a z environment then I could be really that they're actually running this code and I can consume this this string or this prompt with confidence um and I really believe that open source llms are the future I mean open LMS kind of just happen out a mistake of like some toorent leak but they've been exerting a page which is much faster than we had expected uh there's olama and there's so many other models really specializing in different TKS and really accelerating how the open Community can accelerate and participate in the in the llm world so some of the challenges we ask of you to look into is uh explore how St and ZK provs can like do Integrity of ml models how maybe can we can we train models on chain I mean there's the amount of computer slowly getting unlocked maybe some small models are possible to do already uh can we use AI to improve uh blockchain on boarding right now I go to let's say metamask and it's really hard to actually navigate it can the AI figur out what language I use in localized instructions to how I use blockchain maybe even redo the entire UI again and to fit my way um any way is used to Ai and trading I what I mean by that is like trading is a lot of patterns and pattern recognition and AI is really good at doing that can we make sure that the AI can do a m share or even start doing auctions can AI do Dynamic auctions based on changing market conditions um and how can Dows and and onch contracts specifi specify their risks or like their wishes in AI let's say I want my Dow to do have a constitution and I want it to be fast right now Dows are slow because they have to vote for each for each proposal but if the Dow can express their their their demands and their their constitution in an AI model that goes all the time and maybe with some sub checks that stop that there if it goes wrong I think that would really accelerate the the adoption of dows because Dows right now are kind of slow in action and lastly there is the star track um star is a layer to scaling solution that uses Stark for Zer knowledge and privacy uh it enables just a new level of computation uh in the ethereum space uh you can do multical simulations you canun like run AI models and whatnot so the some of the solutions that are avable in this in the St ecosystem is Sky Lang which is their zika St language enables unlimited amount of computation uh sharp is a shared Pro service so you can write your program in a nice high level language and send it to the shop and it will give you back proof for it which enables just a oh sorry I forgot to add cheap proof so it and also it verifies on ethereum for you so you can do some offchain ZK Solutions using Kyo U sharp just today uh and then there's Voyager which is an API service and a stocket RPC service for stocket um yeah enables you to get seamless data quick access to stocket data itself uh some of the ideas that we would like to see is like just pushing the limits of startet computation how can I do some kind of financial derivative that really just pushes the bounds of stocket um computation capacity using stocket native accounts abstraction protocol for for enable enabling the next generation of um users to onboard to stocket and also building no no novel tools using Kyo zero and sharp um Kyo zero is a lowlevel ZK proof language so the proof that you're getting from kyozo are the most efficient ZK proofs that you can get in a nice nice UI and nice efficient language so I think that's a gold M of tooling to be built there uh and lastly yeah uh you can follow us anywhere you can come join to our Discord come talk to us at our booth um we're happy to answer questions if anyone has questions right now I'm also happy to chat right now or outside y thank you all thank you for the presentation uh quick question is about Starck I'm curious about the when you say uh you can do computation for uh like like monteal simulation so my understanding is you can you should use Cairo is correct as a language and you write the program which you would uh yes uh so the question is if I I said in one of my examples if start can you you can do multicolor simulations of.net and the question is do we have to use kaido for programming that uh and the answer is yes K kaido one which is the new version is very drust likee um in fact actually I tried programming uh Kido one with Claude and it works really well so you don't even have to do that much you just have to ask Cloud to write the program for you but yeah uh you can now program your ZK proofs in a rust like language on stocket okay and where do you sorry the last where do you store I mean can you use the results some uh how do you use the result or you store the result or yeah yeah so as an example if you're running the the monal simulation fully on chain then you could compute the result and just save it um in a storage slot or you could just emit it in event and then read the event offchain um yeah but you can also send a Kyo Z program straight to Sharp where it gives you back a proof you can consume the proof and the output uh so there's various ways to use the output of a car program cool if that's all the questions then I'm happy to chat outside uh come see us at the booth at the nethermind booth uh or on the Discord and I'm happy to answer questions thank you [Applause] [Music]