# Keynote: The REAL state of L2s by Bartek Kiepuszewski | Devcon SEA

- Speakers: [Bartek Kiepuszewski](https://streameth.org/speakers/bartek-kiepuszewski)
- Channel: [Devcon](https://streameth.org/devcon)
- Date: 2025-10-09
- Duration: 25:39
- Topics: blockchain, smart contracts, distributed ledger, ethereum, Science & Technology
- Watch: https://streameth.org/watch/yt-ik2JxmHDmyw
- YouTube: https://www.youtube.com/watch?v=ik2JxmHDmyw

## Description

The evolution of Layer 2 solutions has been pivotal in scaling blockchain technologies. This talk, led by L2BEAT founder Bartek Kiepuszewski, delves into the current landscape, recent advancements, and future potential of L2 ecosystems. It will try to address some myths and current challenges of the space. Some important changes to L2BEAT risk framework will also be announced.

Speaker(s): Bartek Kiepuszewski
Skill level: Intermediate
Track: Layer 2
Keywords: Architecture, Layer 2s, Best Practices, myths, reality

Follow us: https://twitter.com/efdevcon, https://twitter.com/ethereum, https://warpcast.com/devcon
Learn more about devcon: https://www.devcon.org/
Learn more about ethereum: https://ethereum.org/ 

Visit the https://archive.devcon.org/ to gain access to the entire library of Devcon talks with the ease of filtering, playlists, personalized suggestions, decentralized access on Swarm, IPFS and more.

Devcon is the Ethereum conference for developers, researchers, thinkers, and makers. 
Devcon SEA was held in Bangkok, Thailand on Nov 12 - Nov 15, 2024.
Devcon is organized and presented by the Ethereum Foundation. To find out more, please visit https://ethereum.foundation/

## Transcript

[Music] can you hear me all right um so um being a Founder uh of el2 beat uh I do have a unique opportunity to work with uh essentially every single L2 uh Team out there and we've been working with them since uh very beginning of this space which is like four years ago that's incredible uh time flies by very quickly um we've worked with the early Pioneers in ZK Tech uh we've worked with the uh plasma group that renamed to optimism we've worked with um arbitrum and uh from this stage I do want to acknowledge uh all these teams because without your support we certainly wouldn't be here and uh as L2 beat we are extremely extremely thankful uh for all the support that the community gives us so uh let's have a look what we've managed to achieve well uh first of all we've managed to launch almost 50 rollups and over 50 LDA l2s that's incredible achievement it's a lot of systems and we've got almost as many to launch in the coming months uh now we've put a lot of value a lot of TVs on these systems as well and if you look at the charts they all going up only um the thing is that you know we've all that growth vers also a lot of responsibility so we do need to like uh understand that end users who are using these systems they are putting money into these rollups because they do believe that rollups inherit the security of ethereum and with that kind of realization uh my opinion uh we need to start getting serious about security uh We've also managed to uh scale ethereum I mean um ethereum was charmming along quite fine but it started to get really really slow for the demand and the fees were uh getting really really high so surely we are scaling these numbers also go up uh this is incredible um however uh there is but you know guys there's always but right and uh I'm just here uh to be honest with all of you and I really want this space to get serious and this is my opportunity to uh uh to call uh for your support uh to make sure that we don't fail we don't fail the expectations of the community that we're starting to uh be really serious about the security of what we're building because you know we like using training wheels for too long if you're an adult using training wheels and I repeat uh it's been 4 years uh then you kind of like are really immature right so it's okay to use training wheels if you're a child it's not okay to use training wheels if you're an adult and I think it's time for all of us really to stop being shy about it we really should all speak out and we shouldn't really suffer from uh emperis uh uh clo syndrome so what is this big butt well first of all most l2s today do not have a proof system which is kind of surprising because early Pioneers like starnet like uh ZK sync like Aztec four years ago when they were launching their first application specific rollups they did have a proof systems so yes you can launch today an L2 with one click of a button however is that really an L2 is that really an rollup you know you're doing is you're launching something which is secured by the multisig I don't think that's good enough and the state state of the ecosystem today is kind of like this on this diagram so on the left you can see the current l2s with the proof system on the right you can see uh current to to's current without the proof system and I would bet that the vast majority of upcoming l2s will not have a proof system that would include essentially every single op stack chain except except for op Manet and Bas and kudos to them by the way you know they like Champions um however every single op stack chain simply does not have a proof system that chart uh on your uh right uh will also include all the orbit Stacks which uh they do have a proof system however it's actually behind often very very short permissioned wh list and sometimes this Whit list is just one actor it's the same as uh State proposal it's essentially ver is a state proposal and it's only them who can challenge themselves like what DF like seriously now most elt do not use security councils what do we mean by the security Council Security Council essentially is a multisig that consists of at least eight participants and require 75% consensus threshold so you can think of it as a big MSE but it's not just about the size I mean size may be important but you know it's not about that right it's about the fact that we want the participants to be geographically decentralized you might have heard yesterday you know an amazing field d uh presentation about you know the need for uh Geo diversification that's what we want uh from these structures and essentially we want the participants uh most importantly to come from different companies and different jurisdictions uh that's super important I'm going to show you some examples like why uh however think about the U security councils as those supercharged multi6 right there's a very very important social lar uh behind them so this is the current state of things and again you know it's very very bad we only have Security councils in arbitrum optimism uh polygon ZK sync and I know that Stark net uh scroll uh interestingly fuel uh is launching with uh a security Council everybody else is essentially a very small uh internal often uh multis and and frankly it is extremely hard to tell the difference between uh these multi and and simple EAS so uh the third big item that we I think did wrong is that most all theas meaning non rollups uh are set up with abysmal da trust assumptions and I use the word abysmal a because I like it and B because it's really really bad uh like look at these guys on the left there are few examples again arbitrum ver ver immutable X however almost everybody else is like literally posting da to their server in the basement or whatever I mean we have no idea we have literally no idea point is they're really bad and they don't seem to care so much so maybe users don't care we don't know but you know we need to really uh look at these that data and tell everybody hey uh that's not data availability committee data availability committee originally was like created by uh in Champion by star quare for the uh starx implementations and by arbitrum but that was not the point that you can like say hey I've got a one survey in my basement I can like call it data availity committee that's not what was uh the point of all that exercise so all together I'm sorry to say that but at the moment uh in most l2s permissioned operators can steal or freeze your funds and we're here really to hey make you all aware of that sorry to say that but we need to change the tud right so uh why should we care about proof systems there are at least three good reasons in our opinion why should we all have a working proof system one is that that actually allows you for permissionless exit in case all the operators are down and they may be down for whatever reason we had uh quite recently a case of dydx going down they have warned the users uh a lot of users didn't exit however if you do have a uh proof system you can make the system so that in a permissionless way someone's going to take over or you can build an escape hatch users will be a able to get their funds out that's super important uh without the proof system you just simply cannot do that it's impossible then uh the second one is that you can actually improve on the trust assumptions of the security Council that's assuming of course you have one and the reason for that this is quite Nuance uh however what uh you can do now is that instead of situation when the malicious proposer and this is like the diagram showing the the vanilla optimistic up without the proof system which you can see in a lot of op stxs today there's a let's say very strong multis that can override the state root and there's a proposer that proposes State rot so if that proposal is malicious all they need to do is to bribe to members of the security Council to look away not to do anything malicious to just simply don't do anything in which case the malicious proposer will actually go through and they will steal the fund right now once you introduce the proof system situation is much much harder for the proposer because now they need to bribe the majority of the security Council not only they have to bribe majority they have to actually make them do something malicious and not just simply look away that's a very very different proposition to make someone look away is to say hey if I give you you know $10 million you just you know lose your keys right or go for a long international flight if you want to make someone do something malicious that's an entirely different proposition so we think that this changes especially with the public Security Council uh trust assumptions fundamentally finally proof systems uh uh if you're stage two allow you to remove any intermediaries whatsoever you don't need a security Council uh or if you do have it you know it's only for like emergency situations if you like right so that may have actually profound regularity uh regulatory implications so you may actually want to launch your uh L2 as a stage two system from the very start that's possible uh but of course you know you do need a uh to have proof system ideally you might want to have more than one and there are already some announcements of systems uh doing that so this is one of the uh the latest announcement from the ne mine team you know they will be building a a roll which is intended to be a stage two at launch so uh again uh that was about the proof system now why security councils and not just simple multi sees and the the reason is that don't believe multi are multi that's the reason unless there's a social layer that actually can convince you that these are uh fundamentally Diversified we we've had uh several uh big uh events uh in our history we had multi-chain uh that uh claimed you know they were very uh decentralized and out that uh no they were not and this is the claim that you cannot really verify independently yourself uh huge uh um attack or inside job or rug we are not sure then we had a situation with Oasis where uh they were approached by UK court and they had to actually use the multi uh to extract some funds from the protocol uh it would be impossible to do that if you had a geopolitically diversified Security Council to do that because there's no court order that can actually reach uh everybody um finally quite recently you know we had a essentially attack on the mtis you know don't think for a second that multi6 can be attacked eventually we had to uh we have to get rid of all of them so to sum up we think uh this is the correct way to think about these systems and they security uh if you have a stage zero rollup with no security consult essentially uh malicious operator can do whatever they want with your funds uh if you are stage zero rollup with a security Council then uh you can try to bribe uh Security Council minority that's uh maybe a hard thing to do but you know much much easier than bribing the security Council majority which you will need to do if you're malicious if your rollup is uh has aov system and finally no one can steal your funds if you're stage two that's the promise of getting to stage two so the question is do we have the right incentives you know for for for uh projects to actually care and the problem is that the only thing that we can do uh we as L2 beat and we as ethereum Community is to apply social pressure uh so this is vitalic saying that you know starting next year he will plan only publicly mention l2s that are stage one he uh previously even said that he's not going to call the system rollups if they're not stage one and this is what you can see today on l2b you know it's very hard really to say uh what is the security of your system right so we were wondering what can you do at the moment we have a stages for rollups we don't have these stages for validum optimium and you know we were wondering for a long time hey you know what maybe we could like introduce stage zero plus for proof systems uh like rollups which are not stage one however they are like better than stage zero but after months of discussion we decided no guys it is time to grow up essentially and uh we we really need to break [Applause] this so uh what we're proposing to the community and this is going to go on the Forum and uh you know with the uh um Community feedback uh that's what we want to do first of all we want to sort of create a separate category for systems kind of similar today the main difference is that you will have to have a proof system uh to be stage zero so for example uh Stark net today will be stage zero under this classification uh all the op stack chains don't have a proof system except based on optimism they will not fall into this category right uh and of course you know we will give time for uh uh for uh systems to adjust so that's like the main category and that should be like a super league right of systems then you have another category of systems that are not using ethereum da so they do use additional uh trust assumptions that come with da uh we call them all da but they would include validium Optimum and whatever you know hybrid construction you may actually uh create uh however they have to give you reasonable de guarantees that cannot be your basement that has to be you know reasonably sized data availability committee or if you're using Celestia if you're using a veil you need to use uh the bridge and you will have to use the bridge because you need to have a proof system so what about the others uh they will still be able to uh we will put them into the third category which we call and now I'm wait wai waiting for the community feedback you know how do we call these systems so our working name is others uh you may find whatever name the point is they are secured by the multis right and we will expose these multis for what they are that's what we want to do in our UI so essentially uh the UI is going to look roughly like this uh you will see this breakdown you will see rollups you will see validium and optimums and you will see others and the default sorting will be by security not by the tvl right let's not Chase tvl with the bad security that's not the point that's going to end really really [Applause] badly so um we will promote stage one and stage two projects and we will look at stage zero projects as contenders for the others we're happy to list them you know we will be extremely uh liberal uh you just need to be uh essentially aligned with ethereum uh you will need to obviously have a bridge you know that would allow you to like move funds however we will look at the trust assumptions we'll look at the multi6 and we hope that slowly but surely you know we will move from others uh to either validum andum or to rollups right but this is what we kind of think the add this category would look like uh this is the real data right now the real systems that may fall into the others category if they don't uh introduce the proof system you will see exactly you know who's the proposer who's the Challenger and who's the upgrader funny thing is you can see that today on l2b is just that you know this information is so uh hidden deeply uh in The Details page that you know I bet that only researchers uh and enthusiasts check out the data it's all available today however we do want to expose the data to end users we want end users to be uh truly aware of what's going on so that we are all accountable for the system that we're building so is it enough to just say hey I've got a proof system you know am I in this first category well it's not uh our pledge to the community as L2 beat is that the next year we're going to put significant resources into actually looking super hard and very deep into these proof systems to make sure that they are sound they are uh uh complete uh we will analyze both ZK and optimistic uh you may have seen uh some of the talks from the morning you know how we approach uh looking at the uh um fraud proof systems we will do essentially the same for the ZK uh we will go into the source code we will look at how you uh created your trusted setup we will look at your circuits and see what exactly is being verified on chain right we want to make everything super uh transparent so that trust assumptions uh they must be clearly communicated and more importantly your proof system cannot be hidden uh behind unreasonably small white list right so uh we're hiring researchers uh we will do all that work and uh this is our pledge uh for the next year to come I hope the next year is going to be uh the year of uh l2s and rollups however it's not about I can launch my uh rollup with one click of a button that's not the point the point is you want to launch and you want to be able to launch a system with uh good security uh ideally you want to in inherit as much security as possible from ethereum uh and there's a lot of work to do uh for all of us to actually reach that but if we don't do that then all we're doing is uh essentially creating uh thousands of secure uh side chains we don't want that uh I think as a community so with that thank you very much and uh join us for the FL let's give him a big hands wow thank you thank you I think we are eum Community really appreciate your work and yes you just cook them on the stage I love it okay let's do the Q&A is it important that Ros have decentralized that sequencer or other safety or Med mechanism are sufficient so this is a very very uh good and important question uh I uh think that are different designs that uh we will uh see I don't think decentralizing sequester is like super important for the security of users funds but it may be important for the uh realtime censorship resistance in certain situations uh vitalic said uh during his opening keynote that the future might be that you know we might see rollups going based uh leveraging ethereum uh infrastructure to uh combat realtime uh sensorship resistance While others like say mega e they might actually have very centralized sequence and they may rely only on the escape hatch uh we might see hybrid constructions I think the design space is huge uh and uh right now you know we like observers SL SL2 beat right you know we really want to see what's going to happen and you know how that's going to play out all right thank you uh next one will te based proof system be considered stage two even if they imply trust in Hardware manufacturer um so the short answer is no uh because um with the constructions that we see today uh essentially uh if you're ising as GX Intel could submit a proof uh and they could potentially uh block steel freeze whatever they want uh without anyone really noticing and without ethereum noticing uh however with all that work that's being put forward you know to create the trustless permissionless tees maybe uh I'm being told that you know this is actually work that's extremely exciting however short answer today no okay why is optimism classified as St one wow based on the evaluation they're not as the foundation control proposal Pro is in it entirely well I mean they uh they essentially uh meet all the uh criteria so it's not really about the proposal process is about you know who's controlling the funds and you can have a centralized um proposal uh however there's a fallback so you know if they go down uh then the whole system becomes more of a permission l so I think it's important to recognize that also what is the role of the security councel and what not right we want stage1 systems to allow you to exit if the centralized proposer stops so for example with the idx you know there was The Proposal was super centralized however when they stopped you know people could exit right so it's not about whether you centralized or decentralized is about whether you can actually exit in the permissionless way um they met all the uh criteria I mean we were refining by the way I mean criteria is not something that's set in stone because all these systems are evolving right so we need to evolve uh with these systems so the criteria might be changing a little bit and we are looking very very closely and both optimism and arbitrum because clearly they like the two leaders there are a lot of nuances that I don't have time to go into but uh uh but yeah I mean it's not like you can you have this stage uh designation forever right so if if there's something like a new information or whatever that we might have skipped or missed you know it's quite possible that you may lose this designation all right okay what are the core reason projects don't build towards stage one uh complexity time cost Talent it's surprisingly costly and like I said the Pioneers uh four years ago they were essentially building for like dydx was literally the first one of the first if not the first uh ZK rollup it was application specific but still it was the first and if not for small nuances it would be stage two really uh it's the governance process that we require you know uh for stage two that's failing but you know like for all intents and purposes is a stage two system it was built four years ago so it's not like it's impossible I think what makes it you know super hard today for for all the Roll-Ups to actually do that frankly is that the majority of the rollups are not built by the teams you know they launched by the rollup as a service providers and we need to incentivize them to actually like do better right so uh and it's hard yeah I mean no one said that it would be easy all right thank you and we running off time thank you Mr Bic and yes let's give him big hands yeah I'm sure we appreciate your hard work and your researching team thank you thank you all right it's lunchtime and