Eva Beylin ~ Working With Layer 2 Panel

foreign [Music] so hi everyone welcome to the layer 2 scalability panel um i'm giorgios i work with loom network and i build the side chains and plasma and the main focus my main focus currently is on the variant called plasma cache hi everyone i'm ben i also work on plasma stuff i work on a team called plasma group and we have a plasma cache variant that we work on we would describe plasma as non-custodial side chains oh and our moderator's here we're going to finish our introductions and then give it over to ava um so my name is liam i work at l4 we're building a framework called counterfactual which is designed to make it easy for developers to build applications that use state channels which is one of the kind of two primary scaling layer two scaling techniques hi i'm jack o'halleran with scale labs we do side chains so very focused on smart contract scaling uh hi i'm mo from seller and we do layer two scaling combining multiple technologies including state channels and side chains uh specifically it's plasma semantic side chains and uh also we are uh proposing this uh crypto economics to ensure safety of the layer two platform and i'll also incentivize liquidity uh on there hi jason tuitch from truebit we do large computations on the blockchain so if you're running out of gas you can call the true bit and keep on going hi my name is eva i used to work at omisego now i'm supporting the ethereum foundation um apologies for being late i'm just finishing another panel but i am excited to dive in um so my first question is does scaling matter does it really matter and what are users dying to to scale for basically yes i think that's why we're we're here um yeah i mean i think they're that we already see like a lot of they're they're what you could do with uh like i could talk from the tribute perspective and i'll let you guys talk state channels so so um [Music] we're you know if you want to check a proof of work inside of a smart contract you might need to there's all kinds of or i guess what i want to say one i guess this is maybe something that people are not familiar with in the in the in the layer 2 scaling space but what truebit has built also is a file system which is kind of useful for getting the data it's it's combined inside of the i guess incentives for for the solvers and verifiers but a sort of a file system that allows you to bring the data from the blockchain or from ipfs or things like that into the system but um certainly you know um there's uh there's applications for that uh cool so to create some controversy i would say no scaling doesn't matter or at least this is like the wrong thinking pattern to think scaling matters um the actual i think the correct thinking pattern should be uh okay we're building this blockchain application we're trying to push it to mass adoption we're trying to get a user to use it and does this application actually able to create a value for the for the developers and for the users of these kind of decentralized applications and to create that value to each scaling absolutely needed uh yeah then in that case yes like you know uh if we actually want to reach to mass adoption for the scale for the decentralized applications we do need to cut down the cost that's the first thing right so we cannot let user to use this application with extremely high cost and we also need to improve the user experiences especially on the front of user interactions the user interaction on blockchain is horrible these days there there's absolutely no interactions on blockchain now that's the part that scaling can actually solve so you know in that front we think that scaling technology is fulfilling the user requirement or the user need of opera or the product need of interaction on top of blockchain uh so that's kind of like my perspective on that and you know specifically speaking state channel is very good for that front like you know it's kind of a you know enabling interactions so but yeah all right i'm going to jump right on the controversy there and agree with you so i think it's easy to think of scalability as a technology problem and a technology issue and something that a bunch of researchers can figure out who really understand cryptographic techniques and scaling techniques but it is equally a user experience problem so we are trying at the end of the day to get users on board this platform i saw somebody at a metric yesterday 8 500 people have used an ethereum depth thus far this year we need to have 10 percent month over month growth to make this thing what it can be and we'll get there but user experience matters both for developers and for the end user if the end user has to completely change the way they use an application we're never going to get like you know out of this you know echo chamber we live in where we're creating products just for each other where ultimately people are just trading crypto right so we have bigger visions we need to make sure user experience is also a part of scalability so for me i think that you need to always have a use case so generally blockchains and scalability are usually looking for problems to solve when they have the solution already so layer 2 solutions they allow you to basically cut off all the mental costs from the gas costs the transaction fees and the finality requirements so for example if i have a game and i want to play a move for example i want to summon a monster or something i don't want to wait 15 seconds for that so the main problem currently is that even though we all have implementations i believe of schemes that work they all have uh certain requirements such as that you need to watch for example like for example for plasma and state channels you need to come online every week and watch and check that there has been no dispute or no invalid state transition so the main problem with that is that again if you want to have proper security of your funds you still need to interact with the blockchain and this is a huge pain so i believe like it is very important to figure out like a good use case which for me the best use case is payments so get payments done right get a huge network like for example like lightning or some sort of like huge plasma chain and or searching and use that so my main issue with layer 2 that we have too many solutions that cannot interoperate with each other so i think maybe a cool way to frame sort of the question that you've asked is like it almost feels like a chicken and egg problem right you have like on the one hand adoption where you want people to be like actually using your thing and on the other hand you need to be able to support that usage and so we've seen like cryptokitties which kind of did like push the usage limits of ethereum and that that sort of made us think oh crap we really need to worry about scaling and then on the other hand now we have lots of blockchain games on ethereum that like could have more usage and still be like within the black the gas limit and we don't see that so i don't think it's like either or it's like we're gonna see each of these things like nudge each other along as we go so like for example um the plasma implementation that my team works on if you guys have heard about plasma you might have heard of like this rsa accumulator magic prime stuff right one something that we realized is literally that only matters a year after you have like a huge number of users and so we realized we don't even need to worry about that right now but once we get users we will need to worry about that right so it's like very much incremental like one interacts with the other yeah i think i probably agree most with georgios here where it's it's kind of a matter of trying to figure out what what are actually going to get people to use these things and when you're actually looking at you know how we build towards getting more adoption trying to interoperate a bit with our different solutions the problem right now is that we all generally are saying scaling matters and there's like side chains and ways you can get rid of the necessity for having to deal with gas costs or multiple different state channel implementations and we're kind of just going in some direction without it without caring for like why is it that we're doing this in the first place what is some example use cases we could like not just kind of get generally working but like work to the level of like a silicon valley level like product that's executed really well and then when that thing starts getting traction then growing it like if the kitties was a good example there's a bit of like nuance there from a product point of view because there was a lot of speculation going on and it wasn't just pure play like a fun dap but i think probably what scale does matter such that when we do find that useful product that we can start to grow the feedback loop on that it doesn't like fail halfway through but it's not the blm doll it needs to be a focus and actually getting that product out um and and again to georgia is a point we need to be able to build interoperable or at least standardized solutions so that we can get these things deployed faster so that when we do it that is that the next product we can scale with it so on this point of usage um you know layer two is really one of the more hyped topics in ethereum what do you feel about the balance between the hype and the true interest and usage for these scaling solutions today we can start from okay my chance to be controversial yeah i don't know i mean i think i think one i think kind of the same point one of the problems is there's a lot of hype because these are like intellectually stimulating things to work on perhaps and a lot of us enjoy that but it's not necessarily the case though once these things are all like shipped and deployed then immediately they're going to get all this usage and the world would be great and you know the price will go to the moon it's that's not where we are i think it's more like this it's probably overhyped in that sense but it is the case that like i said a second ago if we can get these things to work it'll make it much more likely that when we start figuring out useful products for this technology to to create that will actually hit something of use so i would i would say probably over hyped if i had to pick one or the other yeah i mean maybe taco that i feel like a lot of us here are like at east denver and so many of us have gone through the like the colloquial like rabbit hole dive and then oh my god we're gonna replace all forms of governance with you know block chains right um so for us for those people it feels very hype because we know that to do that on the massive scale like we need to have like scalable like tooling but the rest of the world hasn't that hasn't made that click like definitely doesn't see that i don't think there's any hype the hype existed a year ago when layer 2 solutions were being promoted as a one solution of its all for payments for smart contracts trees of plasma chains and whatever else in the original plasma paper like i really feel that was irresponsible over 2017 how layer 2 was pitched as a solution when it still has multiple open problems such as watchtowers for state channels or if you want to have like multiple parties participating that they all need to sign or for plasma that we need to have like some sort of compaction of the history and i really really do not think it is hyped i think it is a hard topic that simply people do not care about because somebody else will do it for them so i it really pains me almost that not enough people are like working more on plasma or side chains and so on because there are tough problems that were hyped you were being told that hey it will solve all your problems zero fee payments forever but here we are there's still no perfect solution yeah i agree with georgios i think uh it's absolutely not overhyped it's talk to any developer who's building a decentralized application this is first and foremost in their mind right now success equals failure in the sense that if you're too successful with transaction throughput you're going to lose money okay and and that's not a good dynamic for anyone so most people i talk just in terms of cost that obviously depends on your financial model but um you know do the math on at 10 cents a transaction and if you do one per second a day right it's incredibly expensive so we need to solve these problems and there's probably you know it may be only overhyped in the sense that there's a couple other very critical problems that need to be solved for but you know talk to talk to people building we need to do this and and and also i think of these as building blocks all of us are working on different kind of ways to skin the cat these building blocks are going to be used as part of large ecosystem so we talk to people building enterprise great decentralized applications they are putting together very sophisticated architecture maps and and so i love the efforts here and i just wish there even more people here today i don't we need more hype in the audience i'm curious what were like the motivations behind so you're building scale which is a solution for this kind of thing right what are the motivations that led to like some specific applications that you think are the ones that people are trying to build that if they were to use this they would be able to benefit from like what are the specific enterprise applications or whatever might be yeah so great question so we're we're very focused on computation so we have a side chain we don't have the same defense measures in terms of of exits we're going to you know over time grow and grow security so people are looking at coming to scale about scaling computation and they're also thinking about using that in conjunction with state channels using that in conjunction with plasma and uh you know things like music things like gaming solar so thinking about how you have smaller high speed lower monetary value transactions that are more computation oriented with smart contracts on scale and then leveraging for peer transactions other systems that can help get real certifiable scalability so that's that's what i'm seeing and you know talking to people almost every day are there anything that are trying to use ethereum right now but they just that are stuck like they actually can't so they're coming to you because they need an alternative place to go to launch their business like what is there like a tangible like we can't use ethereum because x therefore we use player two system another great question so we're we're just trying to help people that build for ethereum so you write you write something in solidity you want solidity to run faster you go put it on scale so we're not necessarily trying to displace anything it's just take your smart contract and in 30 seconds you can have it up on a side chain so that's you know i think i think it's just smart architects out there are saying we need layer two and they look at all of these different options and they go try to figure out how to solve the problem i really think we should avoid specifically talking about our project and focus on the technology stack because we have had evm based chains for a while and the most interesting part in this so for example we had like the tender main base dvm a few years ago from cosmos and the poa network has had a side chain that everybody used to buy their food so i think it would be more interesting like so that we can focus on the actual technical differentiations from our solutions because honestly i've had also an evm-based site in that i can deploy my evm solid the contracts for months now so let's pick it up okay well let's keep the questions then in alignment with that okay yeah i just i wanted to oh i jumped over you oh it's okay i just reacting against this comment but i think that uh i agree i think that the the problem with the hype is that it it blurs the distinction so i think the more that we can focus on saying here here are the specific applications that we can solve and here's how these solutions are different you know the better off it's it's it's uh it's it's gonna be so i think you know for when i when i think about i think it's also about us getting back to the reason you know why did we start these projects in the first place right i mean and i and i look we're you know we're upstairs in the maker space you know building art dolls because that's that's what true i mean art dao is is just for people who are not feeling just talk about a specific application you know a machine that makes art and then sells that art and use the revenue to you know improve its own circuits and and self-sustain living on the blockchain i mean that's really what i think i mean for me personally that's what what makes exciting and that's why you know we built truebit the way that we have this big focusing on autonomous machines but you know really what and i think there's an important distinguish between sort of what's trustless and decentralized right so those are two different things this is an object that just only survives on the blockchain and it doesn't have access to a laptop or anything like that so it's not a corporation right so i guess that's uh what yeah i mean let me revisit the question the question is the high process usage right so for us cetera network very little people know about us uh you know there's no hype basically the balance between us is that we have very little hype but we have real users that is already existing our network and in terms of hype i really don't care uh what i care about is that first thing can i make my users spend less using blockchain and having better user experience second is that can we onboard developer and let our developer have a realistic monetization path building blockchain applications and we have both so uh that's what really matters to us and you know uh to answer vm your question that is there a specific use case let me just take one example here that is not possible building on ethereum and also have high cost in traditional industry but can be enabled by layer two solutions let me tell uh you know the audience a very simple example that is skill based real cash game it's fully regulated in the u.s and you can write and you know in the blockchain you have very very low overheading in terms of payment lower overhead and adding on stage channel we have all the interactiveness you need to build this business model so basically developer get higher cut user get lower cost and the experience is exactly the same as you know no more like you know centralized games so that's what we are shooting for and this is a getting real usage just in four days in our test net app cedar x we see about one thousand user onboarding and people are actively playing these games on our test net so this is like what i think matters and you know i really don't care whether it's a high performance yeah cool i would agree with georgios let's not mention project names as much as possible um the next question i want to ask is um some people argue that layer two solutions piggybacking on layer one makes the root chain less secure um what is your perspective on that and if so is that bad so what makes what makes the chains most secure um so the question was some people argue that layer two solutions make layer one less secure by piggybacking on its security what is your perspective and is that a bad thing yeah fault well there is a there is like a it's definitely true that maybe that's not accurate but there is like a core nugget here which like is very important and like very valid and basically what it is is that even though you can like move transactions off chain uh with layer two solutions because the resolution to that like make those transactions secure because those occur on chain like the ability to censor those transactions and cause invalid outcomes because like the right layer two disputes are censored from the chain because you can use that to steal people's money it does increase the total value of like the assets that are being transacted so it's not that the security of layer one goes down it's that the security of layer one's compromisation has a higher reward because it means that you can compromise what's going on at layer two and break the security assumptions of layer two so it's not that the security of layer one goes down it's that it needs to be high enough to support like the worst case scenario if everything goes wrong basically layer 2 security for i think not for side chains but for plasma and state channels the case is that your value in this in the plasma chain cannot ex in my opinion cannot exceed the security the hash power that cannot exceed a minor attack on the parent chain because if i submit the challenge and that challenge gets this sensor thought it gets reverted within and with every organization um basically i'm not able to get my dispute in um one thing so the original question was that some people argue that it makes it less secure if you commit to it so i will not say less secure but there are certain other layer two approaches which used the parent chain for data availability so for not for state channels not for side chains but for plasma we have this data availability problem that if i have make some off change state transition in order to later make my exit my withdrawal request to the chain i need to submit like some medical proof or something that proves that i own the coins this data cannot be assumed to made available at any point and this is uh i think in our opinion what makes plasma plasma is that you never assume that you have some sort of data availability oracle which gives you the data so there are other variants mainly i believe this is based on the starkdex that was uh presented recently at stanford and also on the ignis site in plasma chain which is which basically what they do is that they commit a miracle route to the parent chain but they also put all the data um required for the users on chain uh in the call data for example of the transactions and the data are not stored in a smart contract but they're placed somewhere in the chain so users can access them the thing is that when you store all this data on chain effectively if i want to say to save let's say some kilobytes of data or some megabytes like it will fill up the block so it doesn't reduce the security of the chain but it reduces the usability of the chain because essentially you're committing the data that you're committing to the chain somehow it's equal to the data that you have on the on the side chain so if you commit like megabytes you still take up the capacity of this of the parent chain and so for example if you take up the whole block and nobody can use it so i think that uh layer two can actually improve the security of layer one like ethereum isn't designed to hold as many tokens as it is uh the more we can you build side chains and uh bridges and spread all the tokens out over many different places the more decentralized we are the more secure we are and you know there's there's that's uh you know that that's i think side the main difference between like in my opinion layer two solutions and side chains are not exactly the same because side chains do not have shared security model like it a sidechain has a totally different security model from the fine chain and you happen to interact with the with it's not even the pattern chain it's like it's on the same level for me you just you interoperate in a way but like our schemes they require the base chain to be secured your schemes do not which is an interesting differentiation i think um okay um so i i guess another question around layer one versus layer two solutions do you think layer two solutions are an interim um before each 2.0 and layer one scaling is solved for i think we're running out of time okay all right last question i'll repeat it uh all right okay let's wrap it up you think layer two solutions are interim solutions before each 2.0 and layer one scaling is solved for yes or no and why uh no strictly no they reduce the capacity you simply bloat the chain less like instead of doing everything on the parent chain you simply take all the state updates of chain and you just commit whatever needs to be done it's in no way outshined by if 2.0 or any kind of scalability on the base layer you just offload computation and that's always good yeah i think there's like a really deadly simple way of thinking about this which is there's two key variables there's like time to finality and then there's the fees that you pay so how soon after i do something is it confirmed to have been done and when i do it how much does it cost no matter how fast the main chain gets layer 2 provides net better or net improvements state channels make it so that the once you're in a channel the fees are zero and the time definitely is instant which is the best possible case of both and plasma makes it be the case that the fees can optionally be zero depending on how the operator works and uh it takes a little it's still not instant but the fees are optionally zero which is net improvement so no matter what happens in the layer one chain even if it gets you know incredibly sharded and really really fast layer two is still strictly necessary and a strict improvement so it's a very simple like yes for me yeah and and just to uh adding on that i think another thing to note is that even for layer one scalability the other one scalability has a bond right so the the the direction people are going is the sharding and shortening is essentially a parallel computing scheme and for any parallel computing scheme as long as there is a small portion of computation that is not parallel and the benefit of parallelization quickly diminishes so in that front like layer one scaling does have a bond it's the curve is going to look like this and there are you need to have very very complicated ways to try to make the system as parallel as possible because if you do it naively the more shard you have the the higher percentage the unparalleled communication would be so you know this is definitely something that also to keep in mind that the layer one scalability has a bond i think it depends on your layer one because almost all these things could be integrated into the base layer so it's it's sort of uh i mean like you can so so yeah i guess that it it really can it's um but i they're needed but you know there are different ways that one could could implement them so you need the whole toolbox to to build all the things that you want to build all right last comment um i think in five years from now if we still have 10 000 users a month layer two there won't be a need for it if we have 200 million users a month layer two will absolutely be a critical part of this and i think we all believe in the growth so we're playing an important role here i want to thank the panelists and thank you guys for joining us