Chain Abstraction for the Modular Era

Transcript

cool all right all right we'll be sharing that link around um feel free to to share it as well and yeah guys welcome in um glad to see a lot of familiar faces and some new faces so why don't we start with some brief intros for for those of you who are new to the chain abstraction landscape in our community as well as here uh let's start with you and kit and then move to Brit and Griffin and then keep going down the line welcome hi hello everyone uh my name is anit I have been in the crypto blockchain space since 2017 uh before that I was an aerospace engineer I did my masters my PhD in

aerospace engineering found crypto in 2017 uh got very excited F down the rabbit hole never came back and um did a bunch of stuff in between 2017 and here uh ended up in coinbase for almost 3 years was doing a lot of backend smart contracts NPC wallet staking and then left coinbase almost one and a half years ago and joined Frontier I lead research over here and what we basically do is basically uh analyze the transaction supply chain so from the moment a user signs a transaction till the moment that transaction lands on chain whatever poops the transaction passes through in between uh we analyze that we write

about it we figure out what incentives are what are the different players and we talk about it within the community that's that's my intro awesome welcome in uh awesome uh I'm Brit I am with a cross and I've been in this space since about 2019 I think um I came in through the community front door uh to the risk laabs team I've been working on both um and across since then um and for anybody who might not know across is is a token bridge between chains and our our little stick is that we the cheapest and fastest bridge out there awesome Griffin

hey everybody uh pleasure to be here H quick background on myself so I actually I got into crypto back in 2017 as well I got in through the lens of crypto gaming actually I built a pretty simple onchain game leveraging uh the four square API and some simple smart contracts on ethereum um then for for my background beyond that so I studied computer science at Stanford my academic research was actually embedded operating system so worked a little bit at the hardware software interface building bare metal OS for like arm arm chips essentially uh during my time I also did get into cryptography and did zero knowledge

protocol research first before getting into the chain abstraction landscape so I actually worked at the intersection of Dow treasuries and sort of proof mechanisms um how do you keep pulled money private was essentially the problem for those who remember Constitution down getting out bid narrowly by a hedge fund billionaire um that was one of the problems I tackled while in school um and then founded the company actually and left early and up been working on holiday for over two years sweet and Tabasco and then Arun and we can kick off yeah for sure sounds good um yeah so quick background from my side uh my name is Tabasco uh although my real name is

Ethan and I do develop relations at partical Network so I hand all of our developer content all of our kind of externally facing stuff um and I guess like a little bit of background myself and then I can kind of quickly intro particle so I've been in the space now for roughly four years or so I kind of started as an independent developer where I built a few different applications and kind of things for the community that ended up getting a decent amount of traction um which ended up leading me into doing develop relations at an RPC provider and doing that for roughly a year and a half then I joined particle Network at this point about a year ago um so at particle we've kind of always been focused on ux Solutions so things that kind of improve core crypto ux and kind of allow users to onboard in the

space and start using applications in a way that's just a little bit easier so our history has kind of been rooted in Social logins and NPC based wallets and stuff like this um and then in like last year we shifted the focus quite a bit to account of struction and we became like one of the largest account of straction providers in the ecem just by like volume of user operations and stuff like this which we're pretty happy about and then this year we wanted to kind of take this technology kind of what we believe to be the next step of like all these different ux facing things which is account of which is chain of struction we're doing this with the means of kind of what we call account level chain of struction so we kind of take a unique approach to it but I assume we'll kind of go deeper into

this but yeah this is kind of what we do at particle brief intro and um looking forward to this conversation awesome um I'm arjin I'm one of the founders of con um I got my start in this space back in 2016 um I previously I was studying physics um and interested in a kind of broader range of topics related to like theoretical physics um decided that I wanted to work on something that would be you know approachable to more people I don't know how I ended up here um I uh uh after shortly after drawing this space became really interested in like solving a lot of problems associated with like infrastructure and user experience um I was like a big believer in the ability to build like public

goods on top of blockchains and um so started connect in 2017 with a very simple mission um figure out how to we were actually one of the first projects to say this but now it's kind of cringy but basically figure out how to how to bring ethereum to a billion users um figure out how to actually make this ecosystem work um and and I think pretty quickly that led us to to becoming very deeply involved in scalability research um we built one of the first l2s on ethereum in 2018 um and then uh as kind of time went on we saw that there was this like we saw the the kind of expansion of the RO of thesis and and we saw that there was this like shared problem that all R ofes had around interop and so in 2020 we made the kind of jump to to focusing full-time on

interoperability and going from being this like largely research oriented organization to becoming like an actual protocol um we were really really fortunate that we we ended up building one of the first like intent based interoperability protocols back in 2020 um and then like that kind of expanded to include um any kind of like arbitrary cross-chain intent um uh at the time we didn't really have the language to even describe this uh and so it's it's really cool to see the way that the space has evolved um and the direction that we're all headed now um and then since then our Focus has been on figuring out well you know so we sort of saw like a a host of problems associated with with intense um and and scaling intense uh starting from 2021 and so like our Focus has been on figuring out how can we solve those

problems to get to the point where like anyone can run a solver um to the point where you can run intent based protocols in any chain for any ecosystem for any application um and that's kind of where we are today and like like Ethan I think we we have some some exciting stuff there that we're working on that I'm sure we'll get into the course of this discussion let's get into it awesome well thanks all for coming um I think to start off uh you know everyone here is part of the cake working group uh and kit you guys at Frontier research have put together this cake idea to sort of illustrate the landscape of uh chain abstraction nowadays could you just kind of take us at a high level through what cake is um and then we can dive into some of its

components yeah um I can like give a brief overview of like like I can talk about like three main things is water chain abstraction why is it necessary and then what is cake and what is what are we trying to p and so chain abstraction I think for the some of the first instances I saw the word chain abstraction used was in Connex docs and the idea is basically uh users should not care what chain they are uh on and it should just be seamless um similar to how when you use dev2 you do not care if the application that you are interacting with does it live on Google cloud or AWS or Azure you just

you just interact with the application on the UI you the users should not care about like if if the application that they're interacting with is on um is on optimism or theum or Cosmos or whatever um so that idea has been around for for some time I think this is the right time to basically explore that idea because at this point of time uh what you have is you have like this abundance of block space you can literally one click one click deploy you can go to conduit or you can go to other rer service providers you can literally one click deploy launch a new rollup and so block space is basically abundant the usage of block space is what is scares now and so what you need to do right now is

basically resolve the ux problem that is what the bottleneck is and right now what you're doing uh what all the users are doing is basically they're holding in their mind like these different where they hold these different states and they are trying to basically before they do any operation they try to collect all these states send them to a particular chain which a particular destination and then do whatever they want to do it's a very buggy experience you basically are asking the user to run like some process in their head uh do do all the operations wait for 10 minutes spend tens of dollars before you can actually do the thing that you want to do with the blockchain you still have a lot of activity on all these different rollups because this just shows like how

powerful Financial incentives are but the ux is very very buggy today and so if you want to bring the next set next 1 billion set of users you need to smooth all that away from the users maybe the current set of users care very very deeply about uh what are the individual properties of all the different rup that they're interacting with but uh chain abstraction basically believes that uh it's an app it's an app decision the users should not care about it and so when you bring in the next set of users maybe they do not care about uh those properties that much and so that all that should be abstracted away and so this is this is what chain abstraction is all about this is what why is it necessary um when we were basically writing the cake we saw like there are

many different ways people are trying to solve the cross chain problem uh but at the end of the day there are basically four layers that uh all these applications are interacting with or are at some location or the other in the in the frame in the cake framework so you have the application layer these are things like unisoft 1 in where the user interacts with them on the front end V some front end um you have the permissioning layer this is the wallet and this is what the users gives permission to uh do something VI initiating transactions you have the solver layer uh this is like a new layer that has emerged due to the Me properties due to like uh use uh it being very painful to basically get your

transaction included and all the Searchers evolved all the market makers evolved and so you have the solver layer which basically tries to solve the routing problem of somehow finding the best route and making sure that the transaction lands on chain and then finally you have the settlement layer which is uh like all these different chains all these different Oracles which do not care about R routing or risk so much they care more about like settlement and finality and making sure that that uh things are fraud proofed so these are the four different layers each of these layers have different tradeoffs that they need to care about so in permission layer you care about like ux versus agency so users might want more agency so like I personally read the the hash that my Hardware wallet generates

but there are other users who just use telegram BS and they click and so there's a wide range of users uh in crypto and so that's what you're trying to do on the permission layer on the solver layer you're trying to find the perfect route based on what the user wants to do and on the settle ler you basically trying to settle things uh in a prop proof manner as quickly as possible I commend you for reading the transaction hash on your wallet every time you submit a transaction it's Quite a feat sometimes um sometimes cool um yeah very very helpful explanation and layout of the chain abstraction space and so the main takeaways that I have here uh I'm just going to keep it very high level which

is that uh we have these four components that make up the chain abstraction space we have the application layer this is the Unis swap the one inch that users interact with we have the permissioning layer this is the wallet or the account layer that uh users again interact with but to instruct the applications then we have the solvers that are looking to fulfill each of those uh intents submitted by the user and then settlement where where it all sort of lands um and so I just want to go around the horn again and give everyone a chance to sort of uh give like their position positioning here you know where are you guys are you at the application Level or are you at the account level or solver or settlement um and then sort of like you know why you guys have chosen

to participate in this cake working group to ultimately bring about the best chain abstraction user experience uh to onboard a billion users uh Brit do you wanna do you want to start with across and maybe touch on the uh intent standard that you guys have going on with Unis swap um yeah sure happy to so in terms of like where we sit in that stack um we don't fit neatly into just one layer um we're up to stuff in in in each of those layers a little bit not so much on the like the wallet layer but uh across V3 when we launched that we released um a cross plus which allows daps to build on top of a cross so we're like building dap tools to help with the chain abstraction

uh we also released across settlements which is um the that base layer that's settlement layer uh the mechanism that we have there uh and then that kind of relies on our our network of Rel layers or solvers whatever you want to call it there's a hundred names for that but we we call them Rel layers um and then we of course like maintain our Bridge dap so we're we're at the DAP layer as well um and in terms of the standard yeah we recently just put out a a standard on Cross chain intense with Unis Swap and the idea behind that is is if we can get adapts that are building these like RFQ mechanisms to adopt this standard it

would allow a network of solvers to plug into all of these different daps more easily uh without you know the excessive burden on uh making big adaptations for each different use case they would all be kind of like reading the same uh transaction type or they messaging basically so um yeah pretty excited about that arj you have anything to add there um nothing to add specifically I mean I can I can go if if that's helpful but or we can keep going down line yeah go ahead I just thought that um you know you may have something to

add there with the with with the cross chain standards I'm just kind of picking on you because of xc20 but but but but feel free as diplomatic as you like I mean I no need for diplomacy like I I am a huge supporter of the across and unisoft work I think there's a big need for it um I think that if we can find better ways to standardize around what these interfaces look like we can promote better competition from the get-go and avoid the Chucho that was like the cross chain ecosystem um and and so yeah I'm I'm like strongly in favor of that as well as any other standards that help like maintain decentralization help maintain like openness um but I would say there's a challenge with like the across new swap standard and and like I've spoken to both Hart and then members of the new

swap team about it which is that it is like purpose built for like trading in limit orders and like to figuring out how to expand that out like if we want to get the most bang for its Buck then ideally we should try to figure out how to expand that out to like other kinds of intents as well but I think that's like a that's something that is like worth that is just part of like an ongoing discussion that that we'll all have about it yeah and I just really want quick want to say like no shade on xc20 um a fan of what they're doing with Connex as well so want to keep that friendly Vibe and both of these things are needed no matter what like there's no there's no way that intents are going to work for every longtail asset and like there's no

way that xc20 is not going to work for like eth or usdc or any other major asset yeah it's different use cases right like there's no perfect solution for everything you just have to figure out the best combination to fulfill the largest set of needs out there Love It Good Vibes all around okay so this is on uh on the bridging and uh sort of like the intent application solver uh touching a few different areas there uh Tabasco you guys are working on the account level more of like the wallet layer um could you kind of like explain how you guys are approaching chain abstraction and uh you know how how you guys might work together with like across and the Unis swap standard

as well as xc20 yeah definitely um you know I think maybe for some context we've been thinking at particle a lot about like how you can I guess understand like you know logically the chain obstruction landscape because I think one of the the big problems recently within like the whole chain obstruction vertical is that you know we've had the cake framework which I think is like one of the I mean really one of the main like foundations that really kicked off a lot of the recent activity within the chain struction landscape and a lot of projects that have popped up in terms of like inspiration from the the cake framework and of course you know huge props to connect for kind of kicking off with the the term of chain abstraction um so I think like there's been a lot of growth in this vertical recently and

particle we've been trying to find ways to like better I guess articulate the landscape and the projects that are built within it um so like based on the cake framework and kind of the granularity of the cake framework we actually recently published an article with agoric where we kind of proposed a way of looking at it that kind of looks at the landscape from a maybe higher level um so from a broader perspective and specifically at those that are within the chst struction landscape so those that are specifically building for chain obstruction so we kind of look at this as chain obstruction solutions kind of existing at three uh relatively simple levels like one being the blockchain level so you can think like uh polygon agair or even like op optimism super chain to some degree

things that sort of create these chain collectives with like shared security uh shared bridging standard St like this um uh then you have the application Level which is you know solutions that are specifically targeting application developers developers that are building applications that can leverage solutions to actually create what you might consider like a chain abstracted experience within the confines of a given application we consider this to be like a gor even skip to some degree and I think connect kind of Falls within that to some degree um which is like kind of providing developers with the tooling to actually build applications that are basically uh you know removed from the confines of a single chain and have the ability to you know Agate balances across chains and stuff like this then where we set is kind of on the account level there's been a lot of work

on the account level recently so you have us you have near is also kind of generally considered on the account level at least from our perspective um even like one balance which is kind of something that Frontier research recently announced we also consider this to kind of be also on the account level which is chain obstruction but applied to um basically you know uh well in this case account so on our side we kind of take this perspective from the uh smart account side so we uh have this standard called or not standard but like our this technology called Universal accounts which essentially exists to kind of provide users with a very easy like universally reproducible mechanism of having one balance that can span Every Chain um so I know like a lot of the CH instuction work that happens today is like how do you give users a balance

they can use anywhere um so we think some of this can be applied at the account level um to a somewhat broader degree so like basically just giving users access to their funds across all chains is one thing um so we kind of see train obstruction as a whole as generally being a combination of at its foundation an account level solution so something like our Universal accounts or even your train signatures or something of this nature and also an application Level solution so we think that like what is a like true chain of structured experience is something that really does break the kind of confines of chains and isn't really restricted to uh for example what developers can interact with on the application Level and also how users can interact with the entire ecosystem from like a more standardized perspectives like having my account I

can go to 100 different applications and use the same account use the same balance without the developer necessarily needing to integrate a specific SDK that's important and then also applications need to have the ability to take this to a more granular degree and actually create Advanced workflows across chains um in a way that also can kind of take advantage of solutions like this like I think um you know for example we've been doing a lot of work with like aor registration I've doing a lot of research in D conect recently um among a few other solutions that have kind of been centered around this application Level and we believe that this future where you have an account level solution that kind of exists it's pretty widespread whether it's like some sort of standardized mechanism or for example Universal accounts from particle or something like this um in combination with applications

that kind of B's philosophy of ch raction and adopt solutions that are geared around improving the experience on the application Level and actually breaking them from these these kind of confines or I guess breaking them from these chains um we believe that this is kind of the future that we're headed towards and hopefully particle are trying to build towards us at the account level so we kind of see ourselves at a very specific position in the space which is kind of contributing with universal accounts um which what we want to see is this adopted alongside a number of other chain of sttion solutions uh specifically those that exist at the application Level and even potentially at the blockchain level um so this is kind of like the perspective that we take on the on the train instruction space and how we're kind of positioning ourselves within

it awesome guys um yeah Arun please um well I I think Griffin also has to go but um I can give a quick summary of where we're at as well because I think one of the challenges in the space is that it's evolving so quickly that like definitionally in the past it was like it was less of a chain abstraction stack and it was more like oh you were building a vertically integrated chain abstraction product um and so that's why you know for example today like we and across compete um because we have competing Bridges and we have competing intent networks but does that necessarily need to be the case forever I don't I don't know I don't think so um for for connect um today I think very similar to across we have you know we don't neatly fit into one part of the stack we have like the the bridge

UI that people use to bridge with we have like an underlying developer platform that people build applications with and then we have this like intent Network directly that we distribute and then underneath the hood we have like a mechanism for like settling intents like that um for us what we've realized is there's actually one piece of this puzzle so the main reason we have all these pieces is because to be honest until now nothing really existed right like we needed to make sure like you have you kind of have to build the whole stack yourself if there's not anything else but there's there's one part of this one problem here that we've been really very like obsessively focused on now for years um which is in 2021 we saw this like really troubling thing when we when we had a bunch of I think we had

like 30 to 50 different intent solvers in our in our ecosystem about 30 to 50 million in tvl we saw this really troubling thing where we found that of all of the 30 to 50 intense solvers maybe 10 five to 10% were actually like us actually doing anything um and and we double clicked on why we realized that it's because when you're when you're kind of uh doing intense across chains your liquidity ends up kind of moving around and uh and if you don't actively manage that liquid if you're not actively like rebalancing and managing your inventory um it ends up usually on chains where it's not really going to get utilized um and uh and so we found that there's this there's this like critical underlying problem associated with with intents and with chain abstraction um that where unless we're

able to kind of solve for how uh solvers can like be more passive how solvers can not have to be like very explicitly uh like you know like agents with explicit expertise like folks like winter M who have hundreds of millions of dollars of capital and like can build in-house aggregators run operational teams like um how can we Mak solving open to more people aside from that um that's been the core problem that Conex has been really has really been prioritizing over the course the last few years and we I think we're about we're maybe about like 10 to 10 to 14 days away from doing some announcements Associated to that uh but I would say like the direction that we're headed is focusing on that piece of the stack so everything that happens I guess in the cake framework everything

that happens between the solver layer and the settlement layer feel the alpha coming um awesome yeah stoked to uh to hear those announcements uh Griffin uh last but not least please cap us off and then you know for for the rest of this you know everyone feel free to just chime in and kind of uh you know know go back and forth without any prompting questions we want to make it like more of a workshop but before we get into that Griffin uh please yeah yeah thank you uh happy to talk about where we sit in the stack yeah we're not neatly positioned to any of the sort of horizontal layers we s we sit at the intersection of two um

specifically we are at the account and solver layer and I can sort of speak to why before saying exactly what we do uh like when I think about the problem of chain abstraction right like what what is driving the need for this infrastructure in the industry it broadly in my mind is a function of sharding for sharting out execution onto multiple State machines right you know we wouldn't be having this discussion if we were all on ethereum L1 and now we have a bunch of different execution threads on L2 on L3 and I think the reason this industry exists is what do we lose when that happens well you lose Atomic composability so I think

everybody here in some sense is working towards solving the problem of bringing composability back to the blockchain but there's one really big problem which is how do you actually bring composability to a system that is no longer synchronous um you're not on the same sort of absolutely ordered blockchain um and then this is sort of where my operating systems background comes in I'm like well asynchronous coordination at a fundamental level is a mut texing problem this is a classic sort of distributed systems problem space and this becomes how do you actually coordinate a set of different computers that are no longer sharing the same you

know clock and ordering um to execute on a common outcome and I think there are few approaches to this and there's a lot of interesting work in Industry right I think some people um are working towards solving this problem at the Block Builder level right we have protocols like swap um and that sort of how do you build a block across many shards so that you can coordinate on the outcome across these virtual machines our opinion is that you should do this at the intra Block Level that you can coordinate not building a block but at the sort of application layer so at the Smart contract layer uh and we've done this at the Smart account layer like why would

you want to do this sort of coordination at the Smart account layer because that is the sort of Nexus point of interaction across many shards and many protocols and we call this orchestration like at the end of the day these are every phrasing of what we're talking about is really boiling down to coordination and orchestration of many components to fulfill user outcomes that are necessary and helpful like pulling your balances from many different places hitting one of many bridges waiting synchronizing locking uh we think about that internally as an orchestration problem so we' built um an orchestration

system that is sort of directly embedded into an account that couples with a network so basically like when you orchestrate you delegate some authority to a solver Network that can then run these tasks you almost like task delegation to to hit your outcome um and we've done this for a while in production uh with some large clients um and this is sort of where we want to take the space is let's embed directly into the account layer um orchestration tooling one so that you know every app developer doesn't need a build at kubernetes equivalent because that's just absurd if every protocol is building orchestration themselves so we think that just should just be a common

Service uh for the industry sweet and anything to add just a just a quick U overview on what is cake working grope I think I missed that that when I was when I was doing like the info I think so our idea of cake working group was we have done two like realtime IRL workshops as well where people we bring in people from different parts in the cake work uh cake working group who are in different layers of the stack and then just uh make them talk about the problems that they are trying to solve and what are the tradeoff points and where they make the tradeoffs in these different design decisions and then we also have a telegram uh Channel where a lot of these teams come in and

then they then they discuss debate share memes and share good cake and so uh what we are trying to do what we are trying to do with uh the working group is basically at least come up at one common set of language or definitions about how we speak about what we are trying to do because even right now when you talk about solvers there are 10 different ways people call solvers and so uh at least come up with a single definition about what these different things are and then uh once you have the definitions on once you can talk about uh how uh you define different properties of the system you can come up to some sort of a standard uh and so that is uh what we want to do next and so people have been doing work on that

so acoss has been doing work on that with respect to the standard they have published uh for cross and limit orders Conex has done xcc2 but I think there's a lot more stuff to be done wherever there are these different layers that interact with each other and so that is uh so cake working group is a venue uh to basically do that yeah and I mean you raised a couple good points there which I think that that uh yeah for the second half of this call would like to kind of more so focus on where where you you kind of speak about how these all kind of come together in the terminologies but also in kind of practical implementations and um you know coming from the perspective that Tabasco brought with the you know three

kind of layers of chain abstraction as we know it today I think that's a common common thread that people are still like associating chain abstraction is just one thing like it's just like one idea and not really breaking it down into like each each part so I guess I'll po like kind of an open question which is like like in a world where there was multiple of these parts put together like what would this look like from like a user perspective or like how would this kind of actually like like go like integrate like how do these solutions kind of come together guess um s go go yeah I was just going to kind of quickly I guess try on my side from this uh you know I think that this is a quite an interesting problem you know at least

when we started kind of getting the transction space um which you know at this point was um I guess you know around five six months ago at this point um one thing that we noticed was that a lot of sentiment like you mentioned was that chain obstruction was sort of a a monolith almost that it was actually like one thing that somebody would do someday and that this would be like the end all of chain obstruction but like honestly one of the main points of rhetoric that we've been like trying to hammer in through all of our content and everything that we've been producing is like chain interaction is actually the opposite of this I mean it's a stack of products I think the cake farmer did a very good job of kind of like starting to shift the sentiment around this and we've been trying to reinforce a lot

through our content and that like what really will at least from our perspective achieve chain abstraction is ultimately a number of interruptable solutions that exist different points in a stack that will collectively achieve this kind of Ideal that is chain abstraction and actually what we're doing right now is that we even believe that like chain obstruction as a term is too broad and too uh there's not enough social consensus around chain obstruction like what what does it actually mean because one of the biggest like I guess not arguments but conversations on Twitter recently have been like what are the relationship between intense and chain abstraction is like and like how do these things actually uh like is is intense like a means of achieving chain obstruction is chain obstruction a product of intent

like so like the semantics of everything have been very unclear up to this point um so like we believe that like I mean first and foremost there needs to be an actual formal definition of what chain obstruction is um and of course how it relates to these different Technologies which I think starting from the cake framework and kind of some other mental models that have been produced recently we trying to kind of you know ease the industry into at least from our perspective um I think that this kind of goes pretty hand inand with this Vision that like at least from a develop a developer perspective what I personally see is that like chain obstruction will kind of be a uh an application specific type of imp impation and like of course there's probably going to be some standards I think there's going to be

some some bridging standards some cross CH standards some things that will be quite important towards making this uh more streamline but I think ultimately the actual solutions that users will use are going to be relatively modular especially on the application Level so for example what type of SDK API does a developer want to use for um I guess like the way that we see the term like orchestration is usually kind of like a developer actually facilitating Crossing transactions or something of this nature so like how would they actually achieve this on the application Level think this will continue to remain uh like it's up to the developer to choose this there's not going to be one solution that will kind of own this entire vertical I think there will be standards that will kind of streamline this but I think ultimately like the way in which chain

obstruction will be achieved even when you look at it as a collective of of solutions won't necessarily be okay here's four or five solutions that work well together and like now we've done chain obstruction I think it will continue to be like different solutions that are building different parts of the stock they improving different parts of the stock um that developers will be able to choose from in sort of like a more modular sense what do they want to use to enable CH structure within their application and what does the ecosystem want to adopt on an account level on a user level to to achieve this as well um so I think that a very good point was made here a moment ago which is that uh you know that the the semantics chain abstraction aren't very well defined and ultimately there needs to be a better

understanding collectively that chain abstraction is not a monolith that it's actually instead a very collaborative and uh I mean quite fundamentally ambitious approach in terms of what we want to achieve as an ecem I mean right now we're talking about removing all bear streen chains to where user is you know in some cases even unaware of what chain they're using so this is a more ambitious thing that it might seem on the surface which is why it's so important to have such a collective of solutions so at least this is how we see it and we've been pushing very hard on this point over the past few months yeah um yeah I appreciate that and I and I think you're you're totally right that there needs to be like better Clarity around the term and around like the technology um I think that the lack

of clarity comes from maybe just like the way in which I mean look like when we introduced a term we were we were introducing it less of as like here is a specific prescriptive way to do things and more of like here's a vision for what the future should look like and that's maybe how I personally Define chain abstraction at the moment is that it's it's a thesis right it's like the modu of thesis it's a vision for how we like the the end state that we want to achieve and that Visions are public fundamentally public goods right they're not they're they're just ideas they're just this just represents a way to think about where we should end up and and I think that way to think about it is specifically very specifically users should not have to care what chain they're on that's that's basically it

right like if we can achieve an outcome where users don't have to care what chain they're on then we've achieved chain abstraction now how we go about doing that and the specific technology that's used is is definitely an open question and I and I agree like I I really appreciate the cake framework specifically because it is it is it is sort of brought to like the fact that hey there's actually a whole ecosystem of different things that are going to be involved here that that we'll have to work together to make this happen um this was kind of this was actually like the big problem so maybe maybe just to take a step back like why did we start talking about chain abstraction last year it was because we ran into this issue um we were uh we were probably like the first ones to be doing uh

intense with something on the target chain like bridging across chains and then doing executing something on the target chain which you kind of introduced back in like 2021 and we we were trying to find pmf with this idea with this product um all of all through like the bare Market from like 22 2022 onwards um but we were struggling and the reason we were struggling wasn't because people users didn't want the thing that we were that we had they were like yeah I actually I see the end state that you want to end up at but the problem that we were having was that there was just so many other pieces that were totally tangential to the work that we were doing associated with okay well like how are you going to abstract away the need to like switch you know chains and metamask like how are you going to

get rid of gas how are you going to get rid of all these other things and we realized okay this is this is just way bigger than us like there's no way that we can solve these all of these problems and so what we started to do was like we we started experimenting with this idea of like well what if we just go and pay ux researchers to go like create public goods like just put out public goods content and like do public good like user research to figure out what is the right way to build this stuff and I think that idea kind of turned into this like notion of like well hopefully if we can Inspire other people to coales around the same idea then maybe we can get more people building each like individual pieces of this solution so that you know there doesn't have to be

one project building everything yeah I think I just want to I guess like highlight something that you just said about the way that you define chain abstraction I think this is something that we I mean this week uh we spent hours thinking about because we're about to produce something that kind of like tries to provide a little bit more of like a a structure as to the definition we've been thinking a lot about like how do you actually Define chain obstruction like what is what are the actual semantics around this and I think that it's which you mentioned this kind of which I believe is very important is kind of like chain obstruction is almost it's not like a specific like prescriptive process or technology it's experience it's kind of like an ideal that is like you mentioned sort of this uh it's an idea it's a public good it's

not like something specific that will yield uh it's an intent yeah exactly it's it's not something that will like yield an exact technology or something that it'll be uh it'll be filled or solved by a collective techn of companies that are building different technology so I think that like yeah this is very important and this is that we've been thinking a lot about as well is like um you know how do you how do you fit within this definition and I think like we there doesn't necessarily need to be like a super rigid like definition that you know everybody needs to follow to a te it's going to coales in something you know that's more open and will change over time I assume as kind of like you know different Technologies kind of uh innovate different parts of the stack but we definitely do kind of agree that the the

future will look like um ideally just to kind of provide a vision maybe to to bounce off of is like the future will look like basically us are going to an application and ideally the only kind of a experience that should be here is between the user and the application that they're using there should be no uh understanding or or kind of being mindful of the chain that they're using or anything other than their account and the application which I think this is actually very interesting uh point that um Frontier research recently made in their paper about one balance which is this transition from a chain Centric landscape to an account Centric landscape and I really like this terminology because I think it represents some of the philosophical shifts pretty well U anyways this is kind of the the we agree very much with um kind of this idea that chain OB

production is more of an experience will evolve over time rather than like a specific definition yeah so to answer Andy's initial question I think like I agree that chain abstraction is a vision it's a intent it's an it's like a it's how users should use uh crypto in a multi chain World in a multi rollup world where where you're launching one new rollup every week and so um it is necessary I also think it is inevitable because uh there is enough there is enough friction that is caused by hopping multiple chains and saving G like having gas in multiple chains that it that it is like app developers are seeing like Drop in usage uh because like there is some friction around it

and so it is inevitable both users and app developers want something like this and so what underlying Technologies actually end up uh materializing uh is a game that will play out it is it will be a very messy game but we know that we cannot just solve that that problem on our own there are multiple people several teams people much more smarter than us who are thinking about that problem this problem day and night and so we just want to bring everyone in the same room and just talk about it and and challenge us and and correct us if I if we are wrong so which is why we have we have taken the approach of being more open just publishing what whatever are uh whatever we are thinking about and just engaging much more actively with different players in in the Ki so are we thinking that users really

care about like what chain they're on and their trust assumptions that are with given chain because I feel like I would like use pretty much any ethereum rollup that settles to ethereum maybe question something if it has a smaller validator set and uses like a modular data availability layer like Celestia Avail ETA maybe be like okay I know that but like I feel like outside of that I I'm not sure that I would really personally care like blast I think is a really good example of like people just full Deen over there without really like you know it was it's effect it was a very clearly a multi- with like like small owners and rollups are effectively still there and we can there a whole another live stream about that but I think like is there is there are you

guys building for users who care about where they are or are you more so just building for like people that like the future that you think of like you just put all these chains together in like one cluster and and just assume that people don't care or is there kind of parameters that you can kind of play with just on how how you guys are seeing the user demand from your uh individual products uh that's a good question I think uh one thing to note is like users preferences are a spectrum around like how comfortable they are with giving permissions giving more like free form permissions to certain set of applications and so you have users which use telegram Bots which

generate like 20 30% of Dex order flow at this day and it's it's uh users just fully trust the application and you have other set of users who use Hardware wallets and so there are there are there's a spectrum of user preferences there's also a spectrum of chain security as you said like you are much more comfortable being in the evm uh roll up Landscape and maybe a little bit near it but if it looks like a centralized system then then maybe not and so who makes that decision is is an important thing to parameterize about uh the use like the beauty of crypto is it is uh you have this sensorship resistance property of

exit so you can exit from any mechanism and if you retain that property you give users the ability to choose whatever type of system that they are comfortable with wherever they lie in this ux versus agency spectrum and uh I think there will be a multitude of different variations of chain abstraction stack uh for each and every type of user and so it it at the end of the day depends on the user what they want to use and then um there's like this decision on what chain to actually deploy your application on is also like a application specific decision to to some to some point and so if a user like today if a user is using unop they are trusting unisoft smart contracts they are trusting unisoft

front end they're making they are trusting unof front endend to make sure thinking that okay this this this front endend is not hacked or they're trusting unof smart contract that it is enough audited so you user still trusting uh the applications uh at some level uh similarly if the uh the chain abstraction applications are taking one more assumption which is what chain we want to let the users interact with it's it's not exactly what you want but it's like a spectrum of things and it should be parameterizable yeah I mean I honestly think like so I think there's like I think there there's obviously like multiple categories of people here right I K it's tot right there's there's like there there is a class of people in this space and you know 100% power to

them there's a reason why the space exists but that there's a class of people in the space that are going to be like yeah you have to run your own nodes for everything and you have to like you have to do everything locally you basically run you like Fork the unisoft interface run the interface locally to actually interact with unisa because you can't trust the hosted code there's no way to verify it in in in the browser and like yeah that's true right like if you the moment you're using any any application in the browser like you could change the application code and like no user whatever know the difference um so there's a range um however I would say like I think people seem to really Mis understand the level of additional trust that is involved in the chain exraction thesis versus where we are today and I think the reason for

this is that people people like I mean you pointed this out yourself like blast is a really good example right like we we sort of have this thesis where we're like okay well users will not use applications in places where they think they could their funds could get stolen but then people go and AP into stuff like BL all the time um in practice if you're looking at like a given chain security model right there's actually if you're like within a single chain there's not really any way for for your funds to get stolen um you know if the chain produces an invalid block you can just just like join the minority fork and then like continue on with the chain right there's not there's not actually a way to like lose money there you can always Fork out through social consensus and so I think like the the cases in

which those security risks actually affect users are more specific spefically associated with how like funds are being like how these transactions are being completed across chains and how funds get settled across chains and in that sense the like chain abstraction intense model is strictly Superior and it's strictly Superior because from the P perspective of the user their interaction is done in seconds right like there like the shorter amount of time that they're they're kind of waiting around for something to happen the more likely they are at risk um and with this with the chain abstraction model you're like offloading all of that security risk all that complexity all of the like overhead of needing to think Think About You Know How likely is it that this is this is potentially going to get hacked or something while I'm using it you're offloading that all to a like

sophisticated third party where they are fronting their own money to do it and like you as a user are completely safe as soon as they've as soon as they've taken that action of fronting their own Capital yeah I think I can add a little color on this as well like I I think it it will be a spectrum at the end of the day because I think different users of different applications will have a different profile of preferences um and I I think a good example of how this has been handled historically when it comes to sort of web server security is actually the browser um like if you look at you know how is https handled how our certificates handled in Chrome for the average Chrome user you go to a website you click through and you're good um

however if you really want to know you can click at that you know you can look at the certificate you can look at the RSA signature you can take that down run verify yourself go all the way up the sort of to the root of trust look at your certificate bundle in your browser if you want to all of that information is public and you have full access to that so if you're really hardcore by all means like verify my TLS certificate on my website um and then there's also like if you go to a website that's not https secure you get a big warning on your browser that says hey stop um so I imagine we probably go into something like this like if you're running a chain and you're sending significant value to

a network that you know has like a one out of one consensus on like a side chain I I imagine most like I think most wallets at some point will say hey uh warning are you sure about this um and then the last thing I'll say is from the direction Industries going I think by the time we have billions and billions of users I think there's going to be an enormous pressure and ALS sort of technological pressure for everything to be um a ZK rollup at this point like I I I I think very soon this will be all these systems will be proof based not very probably within a couple years um and that sort of alleviates some of these risks Beyond liveness risk which like who can actually produce the proof

um and then then I think you rely on sort of UI level warnings uh to so the user is actually aware I think this actually brings up a very interesting uh part of the conversation when it comes to chain obstruction which I've also spent a lot of time thinking about recently which is like I think the chain obstruction you know like like was mentioned will actually be a you know variety of solutions which exist on the Spectrum to I think the Spectrum would most likely and this you know conversation be defined by the degree to which they express information to the user um I think like to one part of the spectrum there's value in like being as simple and as easy to understand as possible like the user doesn't even know what chain that they have their balances on

um I think like this is you know one scenario that we'll see quite often especially in more consumer facing use cases like games or social applications I mean users aren't going to Care necessarily you know where their funds are or like what trains are interacting with which makes sense but also there's some users that that will and I think that this will do two things number one you know different applications are going to use different solutions and different stacks of solutions and this goes back to what I was mentioning earlier which is that ultimately chain abstraction will be a modular stack of account level um and application Level Solutions which themselves will have different levels of user perception as to what they're actually doing and where their balances are etc etc that is ultimately up to the application developer to want to go ahead and

integrate right so do they want an account level solution that is uh you know a little bit broader maybe works with different uh wallets that kind of you know tries to abstract away as much as possible do they want like a specific wallet that maybe you know has like native chain of truction enabled there's some wallets they integrating this with different solutions so I mean like the specifics here are going to deviate based upon use case and I think this is very important that chain of instruction will be implemented um to a modular degree and I think the other part of this conversation that comes up that's quite interesting is security and how chain of Tru can actually affect security I believe I was actually before this I saw a tweet where somebody was mentioning how they think that chain obstruction will uh make security harder

because of course you're doing multiple actions for end users rather than a user just signing one action and I think this also goes down into intense now maybe it was socket that quote retweeted this or somebody did um where they said they made a good point which is that chain instruction actually to some degree makes security easier and it actually reduces uh security risks because no longer is security necessarily being handled by un users specifically it's being handled in some some part by solvers which are economically incentivized to make more secure decisions so this is uh what I believe to be a very important point I think security is another one of the major major problems that like from a user perspective we haven't necessarily solved but I actually think that chain abstraction and to a large Point most chain obstruction Solutions will

actually contribute to the betterment and the resilience of security um rather than the the detriment of it so I think this is also quite an important Point um you know from the usage perception like where they see things and also kind of how Security will be handled so uh I think there's a lot of problems being solved here um but this is yet just one point I wanted to make love that take yeah I think I I tend to agree that's uh very important as well um Arjun I saw you kind of pointing your finger up in the air as like plus one um and in the these last few minutes just want to go was my tweet oh nice cool well uh well yeah I mean just to go around and kind of as as we wrap up get some closing thoughts uh

this has been an awesome panel and so uh Aran since that was your Tweet if you want to start uh with your closing thoughts and we'll we'll go around and hear from everyone um yeah I mean I think um um I guess from my side I'm I'm super excited by where the space is right now um I think uh you know there there's a lot of like questions right now around like how does chain of structure differ from the cross chain thesis I've done a little bit of tweeting around that and I want to like work to to within the cake uh working group to like consolidate around like what the train abstraction thesis specifically is and why it's different from crossin um and I think that there's I think a big part of what why it's different is uh or a big part of what

feels different is also that there's there's much better delineation of responsibility and there's a much better like clearer answ as to like how you can get a bunch of like a broad ecosystem of different projects to work on a shared set of problems to create something that's just a lot bigger than any individual project um for me that's what I find most exciting about this is that like I think there is this is sort of like this is such a big open problem that like any project that kind of gets involved at this stage um to kind of help like like Mark out what this ecosystem will look like is going to be enormously successful and is also going to be contributing to be building like like the future of the internet like I think this is going to have like actual historical significance um so yeah I'm

I'm really happy and and like grateful to be on both this call and also building alongside a bunch of people that are that are doing some really cool stuff here um aside from that I would say the only other closing thought on my side is as I mentioned earlier in this call like Connex is heading uh in a new and really exciting Direction here where we want to try to like work with everybody that's operating the chain of Attraction space and help this space accelerate and succeed as best as possible and in particular do that by decentralizing solvers um which we think is like perhaps one of the biggest open problems remaining um and so uh if you are a listener if you're on this call um I want to talk to you about it and uh and if you're if you're a listener um I

definitely follow follow like the connect Twitter um over the course of next couple of weeks so you can you can hear more about it awesome yeah decentralization and permissionless of that sofware Network so that winter mute and the rest of the market ERS don't just uh monopolize it all um and as far as the the uh you know working together with the people on the call either we'll we'll organize a group chat or maybe an kit and the working group can expose a telegram invite link and we'll get more people into that group if you aren't already in there um and so yeah just to uh just to close out anyone else closing thoughts um this has been a tremendous space so open floor for anyone uh who's got something to add

yeah I I've got some some thought in this like I when I think about chain abstraction um the reason I think this is a very good signal for ethereum in particular is that to me the problems we are facing are an indicator of scale um like the in in in so and and not too many words like the problems we're solving have largely been solved before in the context of other computer systems of other distributed systems uh and I think we're coming to like the conclusion we've come to as an industry right is you can't run the world computer on a single machine right a blockchain is a single-threaded computer that's about as powerful as a calculator

from the 80s or 90s right if you expect to get a billion people on that you know good luck uh and you know like Google right they're running billions of containers and highly optimized data centers on like an x86 instruction set you know ethereum is is not an efficient Virtual Machine by any means so we're we've recognized you need concurrency you need l2s you need orbits you need Super chain hyper chain subnets like every major Foundation has said we need to scale out as horizontal as we can possibly go and the second you do that like we're no longer in vertically integrated land we're in distributed systems land and that's a whole class of problems and that's a very rich and deep

part of computer science really um but you only get to like tackle those problems if you have scale and like it would have been premature to do this three years ago like there there would have been absolutely no need to think about these abstractions like orchestration coordination intent based bridging but now we're there and that's a sign of maturity in the industry and that that's why I think that's that's like the fact that ethereum is really like one of the only ecosystems to be here too it is a very important thing to realize too that this is a mature ecosystem in the market to be thinking about this class of computing problems because this was only really thought about when web 2 companies really scale Beyond running like a server on one instance and they

started having their own server farms and clusters yeah I think that aligns with the general idea of the modular thesis as well um Tobasco yeah yeah I guess just a you know a quick closing thought from my side first of all I think what Griffin just said is absolutely true and I think that was uh could take especially when it comes to the kind of maturity of the ecosystem and the fact that you know we're now thinking about these Technologies I think is a good indicator of that um you know particle we' kind of the whole reason why we got into chain of struction this year and why we wanted to really focus all of our engineering resources on on achieving this at the account level through Universal accounts um is because you know we've been

building we've been building economic traction stuff and and social log into the space for uh you know multiple years now and what we realize is that like the tooling is here now like you can create a pretty killer experience within the silo of one application um I even know that like you know for example holiday with some of their existing products um has like done some pretty cool work in this Regard in terms of uh you know different games and so on uh and there's a lot of other kind of products that have built some you know uh some pretty wild experiences in the stle of one application now I think really what we found was that once you move outside of the styo of one application and start doing things across the ecosystem which really I mean the goal is that users will be able to leverage their accounts

across across chains across applications this is really where the experience starts breaking down and I think that we've seen onboarding in in isolation um but really I I hate using the word mass adoption or like mass onboarding or anything like this but I think like you know getting retail users into the space will happen on an ecosystem level and this is really where chain of exraction comes in this is why we've been so focused on this uh as of this year because we believe this to be like really one of the core missing points on the ux level um for web 3 to to kind of move past its uh prerequisite of having some sort of uh you know technical inclination actually use use the ecosystem um which today is very much true with the need to manually bridge and kind of manage balances across

chains and stuff like this so this is uh you know quite a big uh quite a big problem that we've been trying to address with universal accounts and you know with kind of providing this smart account framework to give users one balance they can leverage across you know any application on any chain um so yeah I think this conversation was great and I think that the the feuture of the space uh with chain abstraction is um very exciting and definitely looking forward to everybody here on the call I think that everybody is building very solid solutions that of course we hope to to kind of be directly compatible with and integrate pretty closely with um so yeah I'm excited for this and thanks for hosting Andy and Robbie of course Brit uh we'll go to you and then to an kit and and wrap up awesome thanks guys um I couldn't

have said it better I mean uh following up from you three you guys had a lot of good stuff to say on closing thoughts it's it's good to be building in a space where even like competing narratives sometimes can be working together which is an interesting thing and like um yeah just it's a good sign of of things to come and we're pretty excited to be working with different pieces of the or different layers of the cake if you will um our big Focus right now is on uh providing like dap tools and the settlement layer so again if any listeners out there are interested in any of those tools please reach out to me um either on Twitter or Telegram and uh yeah just just really excited about everybody who's here and and what we're

all Building Together awesome thanks thanks a lot Robbie thanks a lot Robbie and Andy for having me thanks for having this conversation uh if you if anyone who's listening wants to join the cake working group hit me up uh we'll we'll get you added and um please join share your thoughts correct us it's only together do we do we increase the size of the pie and uh we will be releasing more details about one balance soon so so look out for tweets from Frontier research awesome thank you guys let's abstract the chains thanks