{"id":"EYzYAokCVgM","title":"Cross-app communication (Ethereum Serenity, NEAR, Polkadot, Cosmos)","description":"In the world of chain interoperability and sharded solutions, interaction of applications must be redesigned compared to single chain ecosystem. In this panel each project discusses the tradeoffs and benefits of app chain interaction that they are aiming for.\n\nModerated by James Prestwich, Founder, Summa.\n\nPanelists:\nIllia Polosukhin, Co-Founder, NEAR Protocol\n\nJustin Drake, Researcher, Ethereum Foundation\n\nChristopher Goes, protocol developer & researcher at Tendermint / Cosmos\n\nAlistair Stewart, Web3 Foundation research scientist & Polkadot\n\nThis event was recorded in Osaka on October 12th, 2019 as part of Devcon5.\n\n--\nFor more information on NEAR Protocol: https://nearprotocol.com\nDiscord: https://near.ai/discord\nTwitter: https://near.ai/twitter\nGitHub: https://near.ai/github","channel_name":"NEAR Protocol","published_at":"2020-01-03T00:00:00.000Z","thumbnail_url":"https://img.youtube.com/vi/EYzYAokCVgM/hqdefault.jpg","transcript_text":"we have four different projects represented up here today we have Justin Drake from the etherium foundation Chris goes from cosmos Allister from web three and Ilya from near do you guys want to each take a couple minutes and introduce yourself Justin mm-hmm I'm Justin Drake from different foundation mostly researching if - Christopher goes and the lead author on the interlocking communication specs for cosmos AIA last issue it lead scientists whether that means the web 3 foundation i designing the protocol for polkadot alistair would you mind speaking slightly closer to mike say oh there we go yeah perfect hi I'm Alistair from the web tree foundation and I should have work on protocol design for polka dot hi my name is Julia I'm co-founder near protocol perfect alright so starting off we're talking about cross app communication what even is that does anyone have like a quick definition of what we're trying to do here I think most of us familiar with the etherium smart contract model you know Justin can you put cross app communication in the context of smart contracts for us I mean on f1 right now you have the idea of synchronous communication which happens instantly I think possibly for the purposes of this topic it's actually we actually want a synchronous communication and I think we also have a context where we have two apps which are separated by a boundary this boundary is relatively expensive and one way that we cross this boundary is using the stateless model so the stateless model is where there's a some sort of light clients mechanism to communicate a route across the boundary which could be a sharp boundary or blocked a boundary or whatever and then passed messages which are authenticated against that routes using medical pass for example okay so we have smart contracts they're separated by some major boundary and there's so many different ways to cross that boundary I think each project here except near and the eath to have different names for the boundary I think near we call it sharding can we compare for a second charting to para chains to zones shouting impaired chains are actually very similar we call para chains para chains because they were all different whereas Charlotte system should have been all the same thing but it's sort of moving to the case where shards are actually bit different so chars and para chains are actually fairly similar ideas I'm not sure we're really talking about anything different here but zones are a fair bit different so but in all cases we're basically dealing with a chain that does its own thing but it's also able to communicate and you know we so we want it we have their own block chain with its own network of peers and nodes and may be different so in our case different parrot heads will be rowing different code but they're all part of a bigger system which coordinate them okay I mean I think you've we're part about sharding is that first of all it's the same code it everywhere it's kind of the same convene to the same validators book producers etc versus like whisper change it may be a different coat the like blood producers are different there's maybe kind of a lot of differences that like people can go and change consensus right and that is pretty different from how people think of like sharding and kind of regular I mean both like regular database world as well as in kind of blockchain I think the primary difference between sharding pair chains and my understanding of like execution environments and e2 and IBC zones and cosmos is that all of those three assume a global topological view at least held by some route chain or some master chain and that view informs the security constructions of the whole system and it has to be like agreed on as some singular value at some point in time whereas zones do not so the security model of IBC is we build it in cosmos only talks about two modules or two applications on two different chains talking to one another and what kinds of authentication and like message delivery are needed to realize that which is sort of a weaker assumption which means that you don't need to reason about the global topology it can be changing you don't know one no singular entity needs to even know what it is but you can't make as many assumptions about what like incentives might exist for other actors in your model okay so you're saying that we're charting is homogeneous chains with a shared security model and para chains are heterogeneous chains with a shared security model zones are heterogeneous chains with heterogeneous security models that is right okay or sort of there must exist a minimal like mutual comprehensibility for consensus authentication and for encoding of particular messages but we try to keep that like requisite agreed interface as minimal as possible okay so there has to be some way for them to talk to each other but you you don't necessarily know that the message coming in now it is good or is a trustworthy I mean you know what it means for it to be authenticated so you have some like like client algorithm of the other chain and IBC and you have a way to use that like client algorithm to track the consensus and track whether a particular message was sent to received but even in in an obviously we separate understanding of a consensus authentication and just sort of basic message delivery from understanding of what the message itself means so that means that if there are two like smart contract key chains communicating over IVC the base layers of those chains which might just be execution environments don't understand what the messages mean smart contracts can send different messages one chain to other smart contracts or their apps on another chain and it just needs to be the case assuming you want a correct application that the juice of our contracts on each end or two applications on each end of a channel understand the same message format okay so we're getting pretty deep into the weeds I think but generally speaking all of these are you know different chains that are communicating sending messages back and forth through some shared interface right notably all of these are asynchronous communication which i think is necessary when things are on different chains in like almost every case so is one of the like defining features of eath one is smart contract composability and like instant cross contract communication why is the entire industry moving away from that because we have no choice so it just doesn't work it doesn't scale doesn't scale so we're kind of all unanimously agree that there is no scalable system with synchronous communication no if you want to out-brake the contracts on ten different chains it works with any combination of ten if you're a hundred thousand yards no I mean the the alternative way is pretty much putting like a human like clusters of computers that have a trusted setup between each other and pretty much executed everything in peril was like kind of some type of merge conflict resolution kind of thing and so good I mean I would qualify a little bit I would still expect there to be some synchronous communication occurring I mean right now in our implicit model of apps I think we're thinking of apps the synchronous that may not be true either in like a single app may have parts of its code on different shards and itself executed asynchronously or be running in different threads on a like validated machine I would expect there to be synchronous execution environments where you know some apps that interact with each other all the time end up living and then sort of the more expensive communication boundaries have to be asynchronous and will be used more infrequently I think that's a huge departure from like the current model of it of a smart contract right right isn't that fair to say yeah and I'm not even sure cross apps the right word going back to the original point because often it will actually be the same app running on multiple chains that's talking to itself like we have one token on 100 chains and only one of them is allowed to mint it something like that but it's that then one like logical application well I think it's very difficult to say what an application yes you don't want to get into like the semantics some services like we do in kind of micro service you know web to world where there's a ton of services it may be one big application and they all communicate asynchronously I guess one one way to potentially escape des synchrony is to go with optimistic strategies where you have some sort of entity that's more powerful than a single chain and can observe multiple chains at the same time and can place bets as to how they will all behave in together so you're kind of talking about the credit markets there if there's some entity that's observing multiple chains and you know seeing message emitted on chain a and taking action on chain B there kind of issuing the message the sender on chain a credit right there taking action on their behalf and expecting a future payment for it okay interesting so how would a system like that handle error cases or unrolls like we we we still have no idea like how state progresses between multiple chain systems in practice how does like a shortage smart contract handle the error case where the message on sharda never like shows up on chart B or it's excluded from the chain and the mother history goes forward I mean this is where love complexity creeps in so you know you're trading of potentially good things like low latency and and like resemblance of synchrony for you know lots of complexity and in the tail risk where you have to start rewinding things and going back so yeah so we try to design our system so that this this doesn't really make sense because messages should be able to get passed from one block to the next and should and reverting one thing or that all the messages yeah pretty much like you still want the synchrony at some point and actually it's a faster you can do the synchrony is a faster you can kind of progress the whole system and then if there's any invalid as a state transition or in something got missing you revert the whole kind of chain so I mean I think like in this sense yeah polka dot and how we're designing is pretty similar where you pretty much like from perspective of the users they need to observe the relay chain or in our case just to chain and then if there's something wrong with one of the sub chains or shards you just kind of just reverse up to the point and recover from there I said like from this perspective is the same as like observing helium red so generally speaking if there's an error we reverse that shard to its last finale points no man you three vertical system the whole system yeah that's what happens it poke it up we reverse the whole system okay every pair of chain every pair of chain when you revert the relay chain you revert every power chip okay because I mean otherwise somebody mentor you know a billion tokens on one chain and centers everywhere else it's a much more synchronous model than than cosmos so or to synchronize as fast as possible otherwise kind of and very fast so when we say synchronize were talking about achieving system-wide finality as fast no no okay no no no no so the idea is that we can look you Munich eight faster than finality hmm so each it's a cesarean polka dot basically every religion block contains headers for some parroting blocks that these blocks happen coming to the point of view that relay cane sort of fries synchrony but we can revert the entire relay chain in order before it's finalized and we really want to do this because we can't check we're not sure if entirely sure if any other power chamber locks are actually correct the first time we put them in right because you can't have the relay chain validating every block of every pair of chain well it the relay chain validators do but only one hundreds of them it is a hundred power checks this is a second chain validator is in eighth two right is the only subset of only a subset of the validators validate each shard yes but it's a statistically representative subset so you don't really have that much of a security reduction it's almost equal in security I would challenge that but what if there's optimistic bribery or what if there's like someone bribing the set of validate who happens to get elected I think you have much lower security well yeah basically we have assumptions on you know how adaptive yes hacker is just for clarity how many like well like like glands well there's to expect on beacon shape from the how many validators will they be on the beacon chain and how many validators will there be validating an individual shard so on the beacon changes everyone partly because having everyone participates helps with the finality gadget and then on individual shards you have much smaller committees on this on the order of 256 pages let's say ok and then a lot of this discussion I know doesn't apply to cosmos because there is no nothing blocks we never block blocking sounds hard it does give you nice properties though but when you don't know the topology you can't block right there is no beacon chain or relay chain or master zone so to be honest we don't have a good answer for this question yet but one thing that might happen and that does seem to offer some nice properties is that the risk of reversions will just get priced into assets so that there will be some like a very fast confirmation version of a peg token and there will be some like very slow confirmation but more secure version of the peg token with the challenge period another mechanics and there will exist a flow mark between the two tokens at like 0.9 and 9 to 1 so if you want faster transfer you can pay somebody else to take the risk okay so you expect the free market to solve a lot of these problems for you I would say the question that I would be interested in the answer to because it will determine like which these models works well is sort of what are the costs of dealing with this in system versus dealing with it like at the application layer in a market and in dealing with it in system imposes more costs on the system design more constraints on the system design but it makes it arguably easier for application level like security properties to be reason about ok so how much of the like complexity of implementation are you passing off to the developers of applications rather than the developers of the system yeah I mean that's pretty much general question like how can we make sure that developers who are now building a medium can actually like transition to this new system so is you know totally like so anyway everything they learned but instead like applying some of those principles I mean one thing that should be noted is that even if the consensus layer designers try to make it as easy as possible for the app developers that's usually in the best case or the average case or the optimistic case in the worst case these systems just fell pretty badly in the similar way and so I guess you also want to try and hmm well I guess one possible consequence of that is that as as the industry matures we're gonna have all the complexity from the market anyway as a kind of second layer of insurance as it were I was just gonna say one sort of another question that I'm interested in the answer is you know I don't know it for some of these systems but in IBC we're spending a lot of time thinking about fault containment so if some part of the graph of interconnected applications fails and so like prints token say that's a very easy case what happens like where what other parts the system does it impact where can you contain it and so we're trying to put in a fair amount of containment at the application layer like tracking basically we know you have this kind of dividing line between chains or consensus algorithms or even applications you know tracking acid flow or tracking value flow at that boundary and then you can contain faults if you have some off priori assumptions about what the acid flow could look like but this is pretty much like the whole thing why we at least from me like the benefit of blockchain is that it's handled for a for developer on the blockchain well right you're very much saying that oh now a application developer need to actually handle this kind of cases themself as as like pretty much building up their own like consensus or their market around incentive structure around this right building up some kind of insurance structure in this versus like if you think of like hey I'm building the game or I'm building a defy application I can't really focus on the game design or the economic design of my application budgeting in this complexity becomes like a you know ten there's more work from like perspective building applications yeah I mean one application developer has to reason about it other people can use that application so just like one token developer the poison oak you know I think Justin pointed out a little earlier is and the happy path we get rid of a lot of that complexity for the developer but many of like many attacks on smart contracts and many like issues with them and issues with layer two security models come from the applications developers not understanding the chains assumptions right not understanding the consensus layer well enough so do you really think that we can remove that complexity from app developers or do they need to understand chain internals to be an effective blockchain app developer we'd really like them not have to do that right this is no good up we're trying to go with a simple system where like every chain has the same security and ideally when you're communicating you'll have the same code on both sides so instead of having a standard for message passing you'll have the same code on both sides and you want we trying to get a mechanism where you can wear with your upgrade this code is simultaneous so in this case we basically have something that acts like one app across many chains it'll be the same code but each each chain will have its local state and you might use this say if you have an app on one chain you provide it happy in every other chain that will not that will abstract away the the communication across the chains so how does it work if like one chain is Dax and as a one is like a game so the like they're not a smart contract classrooms which is like literally special case like customized consensus chains here but what's the communication between them I want to trade my game asset on the decks then you probably be using some sort of a fee on your own chain on the game chain which will be using code that was written by the the Dex check so the advantage of unified security model says that it's easier for developers to reason about I agree with that but one concern I have is that it will lead to in some cases security models being mismatched with Apple Asians like I think this is the case for some larger aetherium tops at the moment like like say maker if you use maker you're paying for the security of aetherium you're getting the security of mkr and if like the Utes usage of MKR or the the market cap you know some relative value of that exceeds aetherium it starts introducing really really weird incentives I'm gonna like get one step more controversial because I like having opinions you're when you're using maker you're getting the security of the media neiser not MKR like that's the bound on your security is how strong is their centralized Oracle so would you say in cosmos like app developers get the ability to choose their own security model whereas in other like multi chain ecosystems they're they get the chain security model I think that can be true to be honest I think these protocols especially cosmos and any of these other three are very complementary IBC will allow near to bridge to polkadot to bridge to eath do to bridge to the cosmos hub that's fantastic you know maybe there will be some some applications which need their own zones I think maker is a good example and I think expect there will be many that don't so on that note IBC will allow everything to bridge to everything else right why is that because we cheat and we define an interface instead of defining the implementation so IBC says there exists a like client and then we say that everyone has to write their own like client and we assume that it exists so if you have a light client you can use FEC I mean I'm really interested about this idiom light client that you guys are obviously writing right now right no no yeah because that's like talking to James about this recently yeah this is something that I work on so we write Bitcoin like clients typically but we're also looking in eath 1.0 light client and cash is a mess and is very difficult to do inside of a consensus process and there's expectations I'm assuming for ease to play now the aggregation of public keys will be pretty expensive probably for us too it really depends we can get into like the mechanics of light clients another time I know we only have like 10 minutes left are there any because I know we're all super interested in my clients that's what we all spend all of our time doing but you know let's take it back to like the average app developer for a second right now one of my main issues with writing aetherium smart contracts is fees and metering and I feel like this is going to grow exponentially more complex in a multi-chain ecosystem so for example if my app on shard a sends a message to parrot Ain be transiting through zone C using IV C of course because we're in the nice like future world where everything works who pays fees at each step and how presumably each of these is going to have a different fee asset because this is a completely heterogenous ecosystem do we have any idea how developers are going to handle that this is hard this is hard I mean I can say like for homogeneous environments the like the idea is to kind of maybe under price or over price sometimes but make sure that developers can kind of buy in gas or whatever fees on the entrance and then do not need to kind of handle or care about as it progresses through the system because if you think of it and you have like 20 steps of asynchronous communication between apps the complexity of like trading and like having different prices of different shards etc we will kind of kill everybody just building an application where like some other application is to buy in some more gas because it's like right out of gas halfway that's kind of a very complex problem yeah something about polka dot is we don't necessarily have meeting internally different chains might not have metering some chains will have different metering which makes it more complicated but hopefully I don't know if polka door cosmos will end up standardizing on thoughts and atoms for paying for things but then you have to no definitely not because to go back and answer the question I think fees will be relatively easy if the protocols consider fees before they set their you know code and stone and extremely hard if they don't so in the case of the cosmos hub the cosmos have intentionally intends to accept many feet tokens and kind of offload the work of the user paying the transaction and multiple fees to validators converting V tokens and batch later which makes much more sense but how do you price it yeah that's the Hobbit and yeah it's ardent but it can be a key and still be done by a validator set or like in protocol of a chain instead of by every transaction author every single time they send a transaction but I mean as a developer right I need to write some code that in the front end will decide how much of which talking to attach to transaction right so I need to actually know the exact procedures that validators use or or use the kind of API switch now become like a points of centralization like gas station right now where you kind of like can be now attacked but you know actually mispricing it or even like just jacking up the prices and spending walks away oh good I mean yeah I mean you're right the complex's reckoning of cash markets are gonna increase I mean one thing that was mentioned is that each chain will have its own gas market but then also from a UX perspective do you really want to have tokens and every of these chains and like how do you deal with changed for example that's also one of the strategies that we're taking in these two is to push away this logic at the application layer so you know we in the way we can't get it wrong at the consensus layer commentation and standardization and having it evolve you know this room for even guess abstraction where if you paint making a transferring a token maybe you can also pay the gas in that token there's also ideas around prepaying gas ahead of time so that would you know hedge against the volatility in the future on these markets how is gas token doing these days I don't know so I guess token is an etherium one project that prepaid gas fees I mean people mind a ton when pressure low and how's it using it yep okay so I kind of like Chris and Justin's answers because my as an engineer my favorite kind of problems are somebody else's problems also someone else's problem yeah so I think it's really interesting that we still don't have clear pictures of how fees are going to work in these systems we know that it's going to be easier on the homogeneous systems on the sharded systems like NIR and like eath - and we have various ideas on how to tackle it but you know it is it fair to say that there's no clear picture of fees in any of these ecosystems we have clear picture it's implemented on github right now okay so we can all go check out what in here is going to do yeah I mean I think the other part to consider is that there are different actors in play here so the kinds of users sending transactions like between you know different tokens or a decks are very different than the like you know custody providers who might be transferring large amounts of funds across chains according to security model evaluations like we're solving different problems for different people and I think it will be the case that the cost of these barriers between chains which ranges on some continuum from zero in a synchronous system to low in a homogeneous Chardon system to medium and heterogeneous charted system to as high as possible and IBC you know may lower by good like clients will affect what kinds of transactions take place across those boundaries what kinds of amortization of like transfers being scheduled and matched occurs that sort of thing yeah that's really interesting to me because it implies that developers need to understand the limitations of the system that they're building on when they decide where to build their application Justin I know you have to get going pretty soon anything you want to say and kind of wrap up and then one quick comment on this Fred okay like if we have lots of complexity at the the gas market and we have abstraction and we have lots of different standards being used and we might see we might make it very difficult actually for the process of block block proposal to happen kind of efficiently so if when I want as a validator when I create a block I want to try and choose the transactions in such a way that I maximize my fees but in order to make them by my fees you know I need a whole team of PhDs to you know also include things like you know front running on Dex's and whatnot and one potential way this could evolve into is basically a relayer market so you have real layers which creates really optimal blocks and then they basically sell those blocks to validators who will include them on the blockchain and that's that's potentially a centralization point out sourced block production yeah interesting okay and so that wraps up the fee discussion we're gonna wrap up the panel anything you want to say like in conclusion about eath 2 or your work we're getting there eventually you know it's we've been working on this for years but it's we can see the light up at the end of the tunnel you know now starting to work on on phase one and phase two we have this exciting you know potential significant simplification of phase one and I guess for Phase two we might be in a luxurious position where other chains such as near are basically launching before us and you know taking on the the risk of experimentation so I'm looking forward to learning from them awesome crystal any last words I would emphasize what I mentioned in my talk at Def Con that it bears reinforce izing that most of these protocols were designing I think our non-contentious they can become benevolent monopolies they're not like tokens there isn't a finite amount of like protocols that can exist and once something like some analogous protocol to the internet protocol becomes standardized it just makes everyone's lives better so we should work together phocion they were a bit scared that there is a finite amount of security out there interesting so that's why yeah we gave the shared state shared code she had everything possible any other closing notes about polka dot in 1/3 that was it thank you","transcript_synced_at":"2026-11-04T04:26:34.252Z","segments":[{"start_seconds":3,"text":"we have four different projects represented up here today we have Justin Drake from the etherium foundation Chris goes from cosmos Allister from web three","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=3s"},{"start_seconds":12,"text":"and Ilya from near do you guys want to each take a couple minutes and introduce yourself Justin mm-hmm I'm Justin Drake from","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=12s"},{"start_seconds":21,"text":"different foundation mostly researching if - Christopher goes and the lead author on the interlocking communication","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=21s"},{"start_seconds":29,"text":"specs for cosmos AIA last issue it lead scientists whether that means the web 3","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=29s"},{"start_seconds":35,"text":"foundation i designing the protocol for polkadot alistair would you mind speaking slightly closer to mike say oh","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=35s"},{"start_seconds":43,"text":"there we go yeah perfect hi I'm Alistair from the web tree foundation and I should have work","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=43s"},{"start_seconds":52,"text":"on protocol design for polka dot hi my name is Julia I'm co-founder near protocol perfect alright so starting off","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=52s"},{"start_seconds":62,"text":"we're talking about cross app communication what even is that does anyone have like a quick definition","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=62s"},{"start_seconds":68,"text":"of what we're trying to do here I think most of us familiar with the etherium smart contract model you know Justin can","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=68s"},{"start_seconds":76,"text":"you put cross app communication in the context of smart contracts for us I mean on f1 right now you have the idea of","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=76s"},{"start_seconds":85,"text":"synchronous communication which happens instantly I think possibly for the purposes of this topic it's actually we","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=85s"},{"start_seconds":94,"text":"actually want a synchronous communication and I think we also have a context where we have two apps which are","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=94s"},{"start_seconds":104,"text":"separated by a boundary this boundary is relatively expensive","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=104s"},{"start_seconds":110,"text":"and one way that we cross this boundary is using the stateless model so the","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=110s"},{"start_seconds":119,"text":"stateless model is where there's a some sort of light clients mechanism to communicate a route across the boundary","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=119s"},{"start_seconds":128,"text":"which could be a sharp boundary or blocked a boundary or whatever and then passed messages which are authenticated","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=128s"},{"start_seconds":137,"text":"against that routes using medical pass for example okay so we have smart contracts they're separated by some","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=137s"},{"start_seconds":145,"text":"major boundary and there's so many different ways to cross that boundary I think each project here except near and","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=145s"},{"start_seconds":153,"text":"the eath to have different names for the boundary I think near we call it","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=153s"},{"start_seconds":159,"text":"sharding can we compare for a second charting to para chains to zones","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=159s"},{"start_seconds":170,"text":"shouting impaired chains are actually very similar we call para chains para chains because they were all different","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=170s"},{"start_seconds":178,"text":"whereas Charlotte system should have been all the same thing but it's sort of moving to the case where shards are","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=178s"},{"start_seconds":184,"text":"actually bit different so chars and para chains are actually fairly similar ideas","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=184s"},{"start_seconds":193,"text":"I'm not sure we're really talking about anything different here but zones are a fair bit different so but in all cases","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=193s"},{"start_seconds":203,"text":"we're basically dealing with a chain that does its own thing but it's also","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=203s"},{"start_seconds":211,"text":"able to communicate and you know we so we want it we have their own block chain","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=211s"},{"start_seconds":217,"text":"with its own network of peers and nodes and may be different so in our case","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=217s"},{"start_seconds":227,"text":"different parrot heads will be rowing different code but they're all part of a bigger system which coordinate them okay","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=227s"},{"start_seconds":235,"text":"I mean I think you've we're part about sharding is that first of all it's the same code it everywhere it's kind of the same convene","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=235s"},{"start_seconds":243,"text":"to the same validators book producers etc versus like whisper change it may be a different coat the like blood","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=243s"},{"start_seconds":251,"text":"producers are different there's maybe kind of a lot of differences that like people can go and change consensus right","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=251s"},{"start_seconds":258,"text":"and that is pretty different from how people think of like sharding and kind of regular I mean both like regular","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=258s"},{"start_seconds":265,"text":"database world as well as in kind of blockchain I think the primary difference between sharding pair chains and my understanding of like execution","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=265s"},{"start_seconds":275,"text":"environments and e2 and IBC zones and cosmos is that all of those three assume a global topological view at least held","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=275s"},{"start_seconds":284,"text":"by some route chain or some master chain and that view informs the security constructions of the whole system and it","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=284s"},{"start_seconds":292,"text":"has to be like agreed on as some singular value at some point in time whereas zones do not so the security model of IBC is we build it in cosmos","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=292s"},{"start_seconds":301,"text":"only talks about two modules or two applications on two different chains talking to one another and what kinds of","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=301s"},{"start_seconds":307,"text":"authentication and like message delivery are needed to realize that which is sort of a weaker assumption which means that","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=307s"},{"start_seconds":316,"text":"you don't need to reason about the global topology it can be changing you don't know one no singular entity needs to even know what it is but you can't","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=316s"},{"start_seconds":324,"text":"make as many assumptions about what like incentives might exist for other actors in your model okay so you're saying that","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=324s"},{"start_seconds":331,"text":"we're charting is homogeneous chains with a shared security model and para chains are heterogeneous chains with a","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=331s"},{"start_seconds":338,"text":"shared security model zones are heterogeneous chains with heterogeneous security models that is right okay or","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=338s"},{"start_seconds":347,"text":"sort of there must exist a minimal like mutual comprehensibility for consensus","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=347s"},{"start_seconds":355,"text":"authentication and for encoding of particular messages but we try to keep that like requisite agreed interface as","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=355s"},{"start_seconds":364,"text":"minimal as possible okay so there has to be some way for them to talk to each other but you you don't necessarily know that the message coming in now it is","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=364s"},{"start_seconds":372,"text":"good or is a trustworthy I mean you know what it means for it to be authenticated so you have some like","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=372s"},{"start_seconds":380,"text":"like client algorithm of the other chain and IBC and you have a way to use that like client algorithm to track the","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=380s"},{"start_seconds":388,"text":"consensus and track whether a particular message was sent to received but even in in an obviously we separate","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=388s"},{"start_seconds":395,"text":"understanding of a consensus authentication and just sort of basic message delivery from understanding of","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=395s"},{"start_seconds":403,"text":"what the message itself means so that means that if there are two like smart contract key chains communicating over","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=403s"},{"start_seconds":411,"text":"IVC the base layers of those chains which might just be execution environments don't understand what the messages mean smart contracts can send","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=411s"},{"start_seconds":419,"text":"different messages one chain to other smart contracts or their apps on another chain and it just needs to be the case assuming you want a correct application","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=419s"},{"start_seconds":428,"text":"that the juice of our contracts on each end or two applications on each end of a channel understand the same message format okay so we're getting pretty deep","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=428s"},{"start_seconds":438,"text":"into the weeds I think but generally speaking all of these are you know different chains that","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=438s"},{"start_seconds":445,"text":"are communicating sending messages back and forth through some shared interface right notably all of these are","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=445s"},{"start_seconds":454,"text":"asynchronous communication which i think is necessary when things are on different chains in like almost every","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=454s"},{"start_seconds":462,"text":"case so is one of the like defining features of eath one is smart contract","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=462s"},{"start_seconds":470,"text":"composability and like instant cross contract communication why is the entire","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=470s"},{"start_seconds":476,"text":"industry moving away from that because we have no choice so it just doesn't","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=476s"},{"start_seconds":483,"text":"work it doesn't scale doesn't scale so we're kind of all unanimously agree that there is no scalable system with","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=483s"},{"start_seconds":492,"text":"synchronous communication no if you want to out-brake the contracts on ten different chains it works with any combination of ten if you're a hundred","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=492s"},{"start_seconds":501,"text":"thousand yards no I mean the the alternative way is pretty much putting like a human","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=501s"},{"start_seconds":508,"text":"like clusters of computers that have a trusted setup between each other and pretty much executed everything in peril","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=508s"},{"start_seconds":515,"text":"was like kind of some type of merge conflict resolution kind of thing and so good I mean I would qualify a little bit","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=515s"},{"start_seconds":524,"text":"I would still expect there to be some synchronous communication occurring I mean right now in our implicit model of apps I think we're thinking of apps the synchronous that may not be true either","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=524s"},{"start_seconds":532,"text":"in like a single app may have parts of its code on different shards and itself executed asynchronously or be running in","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=532s"},{"start_seconds":539,"text":"different threads on a like validated machine I would expect there to be synchronous execution environments where","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=539s"},{"start_seconds":547,"text":"you know some apps that interact with each other all the time end up living and then sort of the more expensive communication boundaries have","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=547s"},{"start_seconds":555,"text":"to be asynchronous and will be used more infrequently I think that's a huge departure from like the current model of","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=555s"},{"start_seconds":564,"text":"it of a smart contract right right isn't that fair to say yeah and I'm not even","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=564s"},{"start_seconds":571,"text":"sure cross apps the right word going back to the original point because often it will actually be the same app running on multiple chains that's talking to","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=571s"},{"start_seconds":579,"text":"itself like we have one token on 100 chains and only one of them is allowed to mint it something like that but it's that then one like logical application","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=579s"},{"start_seconds":588,"text":"well I think it's very difficult to say what an application yes you don't want to get into like the semantics some","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=588s"},{"start_seconds":596,"text":"services like we do in kind of micro service you know web to world where there's a ton of services it may be one big application and they all communicate","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=596s"},{"start_seconds":605,"text":"asynchronously I guess one one way to potentially escape des synchrony is to go with","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=605s"},{"start_seconds":614,"text":"optimistic strategies where you have some sort of entity that's more powerful than a single chain and can observe","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=614s"},{"start_seconds":622,"text":"multiple chains at the same time and can place bets as to how they will all behave in together so you're kind of","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=622s"},{"start_seconds":631,"text":"talking about the credit markets there if there's some entity that's observing multiple chains and you know seeing message emitted on chain a and","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=631s"},{"start_seconds":640,"text":"taking action on chain B there kind of issuing the message the sender on chain","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=640s"},{"start_seconds":647,"text":"a credit right there taking action on their behalf and expecting a future payment for it okay interesting so how","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=647s"},{"start_seconds":659,"text":"would a system like that handle error cases or unrolls like we we we still","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=659s"},{"start_seconds":666,"text":"have no idea like how state progresses between multiple chain systems in practice how does like a shortage smart","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=666s"},{"start_seconds":673,"text":"contract handle the error case where the message on sharda never like shows up on chart B or it's excluded from the chain","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=673s"},{"start_seconds":683,"text":"and the mother history goes forward I mean this is where love complexity creeps in so you know you're trading of","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=683s"},{"start_seconds":691,"text":"potentially good things like low latency and and like resemblance of synchrony for you know lots of complexity and in","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=691s"},{"start_seconds":700,"text":"the tail risk where you have to start rewinding things and going back so yeah so we try to design our system so that","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=700s"},{"start_seconds":708,"text":"this this doesn't really make sense because messages should be able to get passed from one block to the next and","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=708s"},{"start_seconds":715,"text":"should and reverting one thing or that all the messages yeah pretty much like you still want the synchrony at some","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=715s"},{"start_seconds":723,"text":"point and actually it's a faster you can do the synchrony is a faster you can kind of progress the whole system and","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=723s"},{"start_seconds":730,"text":"then if there's any invalid as a state transition or in something got missing you revert the whole kind of chain so I","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=730s"},{"start_seconds":739,"text":"mean I think like in this sense yeah polka dot and how we're designing is pretty similar where you pretty much like from perspective of the users they","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=739s"},{"start_seconds":747,"text":"need to observe the relay chain or in our case just to chain and then if there's something wrong with one of the","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=747s"},{"start_seconds":754,"text":"sub chains or shards you just kind of just reverse up to the point and recover from there I said like from this","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=754s"},{"start_seconds":763,"text":"perspective is the same as like observing helium red so generally speaking if there's an error we reverse that shard to its last finale","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=763s"},{"start_seconds":771,"text":"points no man you three vertical system the whole system yeah that's what happens it poke it up we reverse the whole system okay every pair of chain","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=771s"},{"start_seconds":779,"text":"every pair of chain when you revert the relay chain you revert every power chip okay because I mean otherwise somebody mentor you know a billion tokens on one","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=779s"},{"start_seconds":788,"text":"chain and centers everywhere else it's a much more synchronous model than than cosmos so or to synchronize as fast as","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=788s"},{"start_seconds":797,"text":"possible otherwise kind of and very fast so when we say synchronize were talking about achieving system-wide finality as","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=797s"},{"start_seconds":806,"text":"fast no no okay no no no no so the idea is that we can look you Munich eight faster than finality hmm","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=806s"},{"start_seconds":814,"text":"so each it's a cesarean polka dot basically every religion block contains","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=814s"},{"start_seconds":822,"text":"headers for some parroting blocks that these blocks happen coming to the point of view that relay cane sort of fries synchrony but we can revert the entire","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=822s"},{"start_seconds":830,"text":"relay chain in order before it's finalized and we really want to do this","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=830s"},{"start_seconds":837,"text":"because we can't check we're not sure if entirely sure if any other power chamber locks are actually correct the first","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=837s"},{"start_seconds":846,"text":"time we put them in right because you can't have the relay chain validating every block of every pair of chain well","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=846s"},{"start_seconds":853,"text":"it the relay chain validators do but only one hundreds of them it is a hundred power checks this is a second chain validator is in eighth two right","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=853s"},{"start_seconds":861,"text":"is the only subset of only a subset of the validators validate each shard yes but it's a statistically representative","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=861s"},{"start_seconds":870,"text":"subset so you don't really have that much of a security reduction it's almost equal in security I would challenge that","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=870s"},{"start_seconds":879,"text":"but what if there's optimistic bribery or what if there's like someone bribing the set of validate who happens to get","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=879s"},{"start_seconds":887,"text":"elected I think you have much lower security well yeah basically we have assumptions on you know how adaptive yes hacker is","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=887s"},{"start_seconds":896,"text":"just for clarity how many like well like like glands well there's to expect on beacon shape","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=896s"},{"start_seconds":905,"text":"from the how many validators will they be on the beacon chain and how many validators will there be validating an individual shard so on the beacon","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=905s"},{"start_seconds":913,"text":"changes everyone partly because having everyone participates helps with the finality gadget and then on individual shards you","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=913s"},{"start_seconds":922,"text":"have much smaller committees on this on the order of 256 pages let's say ok and then a lot of this discussion I know","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=922s"},{"start_seconds":931,"text":"doesn't apply to cosmos because there is no nothing blocks we never block blocking sounds hard it does give you nice properties though but when you","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=931s"},{"start_seconds":940,"text":"don't know the topology you can't block right there is no beacon chain or relay chain or master zone so to be honest we","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=940s"},{"start_seconds":949,"text":"don't have a good answer for this question yet but one thing that might happen and that does seem to offer some nice properties is that the risk of","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=949s"},{"start_seconds":957,"text":"reversions will just get priced into assets so that there will be some like a very fast confirmation version of a peg","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=957s"},{"start_seconds":965,"text":"token and there will be some like very slow confirmation but more secure version of the peg token with the challenge period another mechanics and","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=965s"},{"start_seconds":973,"text":"there will exist a flow mark between the two tokens at like 0.9 and 9 to 1 so if you want faster transfer you can pay","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=973s"},{"start_seconds":980,"text":"somebody else to take the risk okay so you expect the free market to solve a","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=980s"},{"start_seconds":986,"text":"lot of these problems for you I would","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=986s"},{"start_seconds":993,"text":"say the question that I would be interested in the answer to because it will determine like which these models works well is sort of what are the costs","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=993s"},{"start_seconds":1000,"text":"of dealing with this in system versus dealing with it like at the application layer in a market and in dealing with it","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1000s"},{"start_seconds":1009,"text":"in system imposes more costs on the system design more constraints on the system design but it makes it arguably easier for application level like","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1009s"},{"start_seconds":1017,"text":"security properties to be reason about ok so how much of the like complexity of implementation are you passing off to the developers of applications rather","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1017s"},{"start_seconds":1025,"text":"than the developers of the system yeah I mean that's pretty much general question like how can we make sure that developers who are now building a medium","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1025s"},{"start_seconds":1033,"text":"can actually like transition to this new system so is you know totally like so anyway everything they learned but instead like","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1033s"},{"start_seconds":1041,"text":"applying some of those principles I mean one thing that should be noted is that even if the consensus layer designers","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1041s"},{"start_seconds":1049,"text":"try to make it as easy as possible for the app developers that's usually in the best case or the average case or the","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1049s"},{"start_seconds":1057,"text":"optimistic case in the worst case these systems just fell pretty badly in the similar way and so I guess you also want","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1057s"},{"start_seconds":1065,"text":"to try and hmm well I guess one possible consequence of that is that as as the industry matures we're gonna have all","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1065s"},{"start_seconds":1073,"text":"the complexity from the market anyway as a kind of second layer of insurance as it were I was just gonna say one sort of","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1073s"},{"start_seconds":1083,"text":"another question that I'm interested in the answer is you know I don't know it for some of these systems but in IBC we're spending a lot of time thinking","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1083s"},{"start_seconds":1089,"text":"about fault containment so if some part of the graph of interconnected applications fails and so like prints","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1089s"},{"start_seconds":1096,"text":"token say that's a very easy case what happens like where what other parts the system does it impact where can you contain it and so we're trying to put in","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1096s"},{"start_seconds":1105,"text":"a fair amount of containment at the application layer like tracking basically we know you have this kind of dividing line between chains or","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1105s"},{"start_seconds":1113,"text":"consensus algorithms or even applications you know tracking acid flow or tracking value flow at that boundary and then you can contain faults if you","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1113s"},{"start_seconds":1122,"text":"have some off priori assumptions about what the acid flow could look like but this is pretty much like the whole thing why we at least from me like the benefit","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1122s"},{"start_seconds":1131,"text":"of blockchain is that it's handled for a for developer on the blockchain well right you're very much saying that oh now a application developer need to","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1131s"},{"start_seconds":1140,"text":"actually handle this kind of cases themself as as like pretty much building up their own like consensus or their","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1140s"},{"start_seconds":1148,"text":"market around incentive structure around this right building up some kind of insurance structure in this versus like if you think of like hey I'm building","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1148s"},{"start_seconds":1156,"text":"the game or I'm building a defy application I can't really focus on the game design or the economic design of my","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1156s"},{"start_seconds":1164,"text":"application budgeting in this complexity becomes like a you know ten there's more work from like perspective","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1164s"},{"start_seconds":1171,"text":"building applications yeah I mean one application developer has to reason about it other people can use that application so just like one token","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1171s"},{"start_seconds":1178,"text":"developer the poison oak you know I think Justin pointed out a little earlier is and the happy path we get rid of a lot of that complexity for the","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1178s"},{"start_seconds":1186,"text":"developer but many of like many attacks on smart contracts and many like issues","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1186s"},{"start_seconds":1193,"text":"with them and issues with layer two security models come from the applications developers not understanding the chains assumptions right not understanding the consensus","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1193s"},{"start_seconds":1202,"text":"layer well enough so do you really think that we can remove that complexity from app developers or do they need to","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1202s"},{"start_seconds":1209,"text":"understand chain internals to be an effective blockchain app developer we'd really like them not have to do that","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1209s"},{"start_seconds":1217,"text":"right this is no good up we're trying to go with a simple system where like every chain has the same security and ideally","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1217s"},{"start_seconds":1224,"text":"when you're communicating you'll have the same code on both sides so instead of having a standard for message passing you'll have the same code on both sides and you want we trying to get a","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1224s"},{"start_seconds":1233,"text":"mechanism where you can wear with your upgrade this code is simultaneous so in this case we basically have something","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1233s"},{"start_seconds":1241,"text":"that acts like one app across many chains it'll be the same code but each each chain will have its local state and you might use this say if you have an","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1241s"},{"start_seconds":1249,"text":"app on one chain you provide it happy in every other chain that will not that will abstract away the the communication","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1249s"},{"start_seconds":1255,"text":"across the chains so how does it work if like one chain is Dax and as a one is like a game so the like they're not a","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1255s"},{"start_seconds":1265,"text":"smart contract classrooms which is like literally special case like customized consensus chains here but what's the communication between them I want to","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1265s"},{"start_seconds":1274,"text":"trade my game asset on the decks then you probably be using some sort of a fee on your own chain on the game chain","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1274s"},{"start_seconds":1282,"text":"which will be using code that was written by the the Dex check so the advantage of unified security model says","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1282s"},{"start_seconds":1291,"text":"that it's easier for developers to reason about I agree with that but one concern I have is that it will lead to in some cases security models being","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1291s"},{"start_seconds":1299,"text":"mismatched with Apple Asians like I think this is the case for some larger aetherium tops at the moment like like say maker if you use maker","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1299s"},{"start_seconds":1307,"text":"you're paying for the security of aetherium you're getting the security of mkr and if like the Utes usage of MKR or","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1307s"},{"start_seconds":1315,"text":"the the market cap you know some relative value of that exceeds aetherium it starts introducing really really","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1315s"},{"start_seconds":1322,"text":"weird incentives I'm gonna like get one step more controversial because I like having opinions you're when you're using","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1322s"},{"start_seconds":1331,"text":"maker you're getting the security of the media neiser not MKR like that's the bound on your security is how strong is","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1331s"},{"start_seconds":1338,"text":"their centralized Oracle so would you say in cosmos like app developers get","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1338s"},{"start_seconds":1345,"text":"the ability to choose their own security model whereas in other like multi chain ecosystems they're they get the chain","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1345s"},{"start_seconds":1355,"text":"security model I think that can be true to be honest I think these protocols especially cosmos and any of these other","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1355s"},{"start_seconds":1364,"text":"three are very complementary IBC will allow near to bridge to polkadot to bridge to eath do to bridge to the","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1364s"},{"start_seconds":1371,"text":"cosmos hub that's fantastic you know maybe there will be some some applications which need their own zones","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1371s"},{"start_seconds":1379,"text":"I think maker is a good example and I think expect there will be many that don't so on that note IBC will allow","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1379s"},{"start_seconds":1387,"text":"everything to bridge to everything else right why is that because we cheat and","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1387s"},{"start_seconds":1394,"text":"we define an interface instead of defining the implementation so IBC says there exists a like client and then we say that everyone has to write their own","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1394s"},{"start_seconds":1402,"text":"like client and we assume that it exists so if you have a light client you can use FEC I mean I'm really interested","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1402s"},{"start_seconds":1409,"text":"about this idiom light client that you guys are obviously writing right now right no no yeah because that's like","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1409s"},{"start_seconds":1420,"text":"talking to James about this recently yeah this is something that I work on so","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1420s"},{"start_seconds":1427,"text":"we write Bitcoin like clients typically but we're also looking in eath 1.0 light client and","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1427s"},{"start_seconds":1433,"text":"cash is a mess and is very difficult to do inside of a consensus process and","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1433s"},{"start_seconds":1440,"text":"there's expectations I'm assuming for ease to play now the aggregation of public keys will be pretty expensive probably for us too","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1440s"},{"start_seconds":1449,"text":"it really depends we can get into like the mechanics of light clients another time I know we only have like 10 minutes","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1449s"},{"start_seconds":1457,"text":"left are there any because I know we're all super interested in my clients","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1457s"},{"start_seconds":1465,"text":"that's what we all spend all of our time doing but you know let's take it back to like the average app developer for a","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1465s"},{"start_seconds":1473,"text":"second right now one of my main issues with writing aetherium smart contracts","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1473s"},{"start_seconds":1479,"text":"is fees and metering and I feel like this is going to grow exponentially more","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1479s"},{"start_seconds":1485,"text":"complex in a multi-chain ecosystem so for example if my app on shard a sends a","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1485s"},{"start_seconds":1494,"text":"message to parrot Ain be transiting through zone C using IV C of course because we're in the nice like future","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1494s"},{"start_seconds":1502,"text":"world where everything works who pays fees at each step and how presumably each of these is going to","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1502s"},{"start_seconds":1509,"text":"have a different fee asset because this is a completely heterogenous ecosystem do we have any idea how developers are","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1509s"},{"start_seconds":1517,"text":"going to handle that this is hard this is hard I mean I can say like for","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1517s"},{"start_seconds":1524,"text":"homogeneous environments the like the idea is to kind of maybe under price or","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1524s"},{"start_seconds":1531,"text":"over price sometimes but make sure that developers can kind of buy in gas or whatever fees on the entrance and then","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1531s"},{"start_seconds":1540,"text":"do not need to kind of handle or care about as it progresses through the system because if you think of it and you have like 20 steps of asynchronous","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1540s"},{"start_seconds":1549,"text":"communication between apps the complexity of like trading and like having different prices of different shards etc we will kind of kill","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1549s"},{"start_seconds":1558,"text":"everybody just building an application where like some other application is to buy in some more gas because it's like","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1558s"},{"start_seconds":1566,"text":"right out of gas halfway that's kind of a very complex problem yeah something about polka dot is we don't necessarily have meeting","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1566s"},{"start_seconds":1574,"text":"internally different chains might not have metering some chains will have different metering which makes it more complicated but hopefully I don't know","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1574s"},{"start_seconds":1583,"text":"if polka door cosmos will end up standardizing on thoughts and atoms for paying for things but then you have to","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1583s"},{"start_seconds":1589,"text":"no definitely not because to go back and answer the question I think fees will be","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1589s"},{"start_seconds":1598,"text":"relatively easy if the protocols consider fees before they set their you know code and stone and extremely hard","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1598s"},{"start_seconds":1605,"text":"if they don't so in the case of the cosmos hub the cosmos have intentionally intends to accept many feet tokens and","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1605s"},{"start_seconds":1614,"text":"kind of offload the work of the user paying the transaction and multiple fees to validators converting V tokens and","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1614s"},{"start_seconds":1623,"text":"batch later which makes much more sense but how do you price it yeah that's the Hobbit and yeah it's ardent but it can","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1623s"},{"start_seconds":1630,"text":"be a key and still be done by a validator set or like in protocol of a chain instead of by every transaction","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1630s"},{"start_seconds":1638,"text":"author every single time they send a transaction but I mean as a developer right I need to write some code that in","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1638s"},{"start_seconds":1647,"text":"the front end will decide how much of which talking to attach to transaction right so I need to actually know the","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1647s"},{"start_seconds":1654,"text":"exact procedures that validators use or or use the kind of API switch now become like a points of centralization like gas","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1654s"},{"start_seconds":1662,"text":"station right now where you kind of like can be now attacked but you know actually mispricing it or even like just","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1662s"},{"start_seconds":1670,"text":"jacking up the prices and spending walks away oh good I mean yeah I mean you're right the complex's reckoning of cash","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1670s"},{"start_seconds":1679,"text":"markets are gonna increase I mean one thing that was mentioned is that each chain will have its own gas market but then also from a UX perspective do you","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1679s"},{"start_seconds":1688,"text":"really want to have tokens and every of these chains and like how do you deal with changed for example that's also one of the strategies that we're taking","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1688s"},{"start_seconds":1697,"text":"in these two is to push away this logic at the application layer so you know we in the way we can't get it wrong at the","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1697s"},{"start_seconds":1705,"text":"consensus layer commentation and standardization and having it evolve you","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1705s"},{"start_seconds":1713,"text":"know this room for even guess abstraction where if you paint making a transferring a token maybe you can also","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1713s"},{"start_seconds":1720,"text":"pay the gas in that token there's also ideas around prepaying gas ahead of time","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1720s"},{"start_seconds":1728,"text":"so that would you know hedge against the volatility in the future on these markets how is gas token doing these","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1728s"},{"start_seconds":1737,"text":"days I don't know so I guess token is an etherium one project that prepaid gas","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1737s"},{"start_seconds":1745,"text":"fees I mean people mind a ton when pressure low and how's it using it yep","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1745s"},{"start_seconds":1752,"text":"okay so I kind of like Chris and Justin's answers because my as an engineer my favorite kind of problems","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1752s"},{"start_seconds":1760,"text":"are somebody else's problems also someone else's problem yeah so I","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1760s"},{"start_seconds":1771,"text":"think it's really interesting that we still don't have clear pictures of how fees are going to work in these systems","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1771s"},{"start_seconds":1778,"text":"we know that it's going to be easier on the homogeneous systems on the sharded systems like NIR and like eath - and we","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1778s"},{"start_seconds":1787,"text":"have various ideas on how to tackle it but you know it is it fair to say that there's no clear picture of fees in any","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1787s"},{"start_seconds":1794,"text":"of these ecosystems we have clear picture it's implemented on github right now okay so we can all go check out what in here is going to do yeah I mean I","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1794s"},{"start_seconds":1802,"text":"think the other part to consider is that there are different actors in play here so the kinds of users sending","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1802s"},{"start_seconds":1809,"text":"transactions like between you know different tokens or a decks are very different than the like you know custody","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1809s"},{"start_seconds":1817,"text":"providers who might be transferring large amounts of funds across chains according to security model evaluations like we're solving different problems for different people and","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1817s"},{"start_seconds":1825,"text":"I think it will be the case that the cost of these barriers between chains which ranges on some continuum from zero","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1825s"},{"start_seconds":1833,"text":"in a synchronous system to low in a homogeneous Chardon system to medium and heterogeneous charted system to as high","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1833s"},{"start_seconds":1840,"text":"as possible and IBC you know may lower by good like clients will affect what kinds of transactions take place across","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1840s"},{"start_seconds":1847,"text":"those boundaries what kinds of amortization of like transfers being scheduled and matched occurs that sort of thing yeah that's really interesting to me because it implies that developers","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1847s"},{"start_seconds":1856,"text":"need to understand the limitations of the system that they're building on when they decide where to build their application Justin I know you have to","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1856s"},{"start_seconds":1865,"text":"get going pretty soon anything you want to say and kind of wrap up and then one quick comment on this Fred okay like if","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1865s"},{"start_seconds":1874,"text":"we have lots of complexity at the the gas market and we have abstraction and we have lots of different standards","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1874s"},{"start_seconds":1880,"text":"being used and we might see we might make it very difficult actually for the","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1880s"},{"start_seconds":1887,"text":"process of block block proposal to happen kind of efficiently so if when I","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1887s"},{"start_seconds":1896,"text":"want as a validator when I create a block I want to try and choose the transactions in such a way that I maximize my fees but in order to make them by my fees you know I need a whole","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1896s"},{"start_seconds":1903,"text":"team of PhDs to you know also include things like you know front","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1903s"},{"start_seconds":1910,"text":"running on Dex's and whatnot and one potential way this could evolve into is","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1910s"},{"start_seconds":1917,"text":"basically a relayer market so you have real layers which creates really optimal","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1917s"},{"start_seconds":1925,"text":"blocks and then they basically sell those blocks to validators who will include them on the blockchain and that's that's potentially a","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1925s"},{"start_seconds":1933,"text":"centralization point out sourced block production yeah interesting okay and so","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1933s"},{"start_seconds":1942,"text":"that wraps up the fee discussion we're gonna wrap up the panel anything you want to say like in conclusion about eath 2 or your work","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1942s"},{"start_seconds":1956,"text":"we're getting there eventually you know it's we've been working on this for years but it's we can see the light up at the end of the tunnel you know now","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1956s"},{"start_seconds":1965,"text":"starting to work on on phase one and phase two we have this exciting you know potential significant simplification of","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1965s"},{"start_seconds":1973,"text":"phase one and I guess for Phase two we might be in a luxurious position","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1973s"},{"start_seconds":1980,"text":"where other chains such as near are basically launching before us and you know taking on the the risk of","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1980s"},{"start_seconds":1988,"text":"experimentation so I'm looking forward to learning from them awesome crystal any last words I would emphasize","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1988s"},{"start_seconds":1997,"text":"what I mentioned in my talk at Def Con that it bears reinforce izing that most of these protocols were designing I think our","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=1997s"},{"start_seconds":2004,"text":"non-contentious they can become benevolent monopolies they're not like tokens there isn't a finite amount of like protocols that can exist and once","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=2004s"},{"start_seconds":2012,"text":"something like some analogous protocol to the internet protocol becomes standardized it just makes everyone's lives better so we should work together","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=2012s"},{"start_seconds":2021,"text":"phocion they were a bit scared that there is a finite amount of security out","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=2021s"},{"start_seconds":2028,"text":"there interesting so that's why yeah we gave the shared state shared code she had everything possible any other closing notes about polka dot in 1/3","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=2028s"},{"start_seconds":2037,"text":"that was it thank you","url":"https://www.youtube.com/watch?v=EYzYAokCVgM&t=2037s"}],"youtube_url":"https://youtube.com/watch?v=EYzYAokCVgM","url":"https://near-pedia.xyz/videos/EYzYAokCVgM"}