Building a Post-Merge World

foreign [Music] thank you so much all of you folks for coming and especially thank you so much for uh our panelists that you made it here and I'm very excited to have many interesting uh people on the stage here today so um yeah let's get into it I would say let's start with a round of introductions uh so please uh tell us what project are you from what you work on within the project and we are at Defcon so mention maybe your favorite cute animal or something like that favorite cute animal of yours sure uh my name is Paul hauna co-founder of Sigma Prime which is the company that maintains Lighthouse also I guess a co-founder of Lighthouse I've been working on ethereum staking since about 2018 was involved in founding Lighthouse of written a good chunk of the code on that um and now in the day-to-day I'm um kind of I guess overseeing project managing still engineering as much as I can my favorite animal is a quokka Australian animal I love cooker so much yeah uh okay hi I'm Chris I'm co-founder of company called L2 beat and the two stands not for uh ethereum two uh but for Layer Two so uh we're we're working on uh researching and providing analytics for different layer 2 scaling Solutions um and we're also running a bunch of notes including if 2.0 node uh our main goal is contribute to the decentralization of of the film Network by providing uh some kind of images for for uh these kind of devices like uh uh Raspberry Pi or rock 5 and turn them into a fully film node pretty much automatically in this way I make it easy for for uh regular users from users to run a full node in these uh low power and resource constraint devices and so this way we can facilitate this uh for for people to run to run all set home in values uh my name is Nick zorkish I am um I work at East acre at Staker is a completely volunteer run organization and we facilitate Community Education and uh uh uh staking encourage people to stake we provide guidance and do technical support for stakers so we're just connecting stakers to the appropriate uh protocols that they can use to stake oh my favorite animal can I be boring into say a cat uh yeah I think it's kind of right ethereum culture Vibe but there are so many cat people here and um I mean kids are great source of entropy right I mean it's also something which is which is important like cryptography okay um anyway thank you so much for the introductions uh and I'm Mario I'm from ethereum Foundation protocol support and um yeah so um the merch um something which I've been also contributing to which is the congrats no oh yeah congrats whoa it happened before we celebrated a lot to merge so like uh we are over all the celebrations and now I think it's time to discuss what's next because the merge the merge was just the beginning there is no time to rest and we need to build right um and uh yeah so maybe let's start with some Reflections so um how how is the merge for you for your project but especially what it changed was there some challenges that you had to overcome and now after the merch what are you facing some uh some new issues new challenges new changes in your projects okay uh yeah for in from our side uh we were running uh nulls on Raspberry Pi's we have several several boards and after the match we were facing two too many issues uh first of all we we've seen uh from some clients some issues with uh performance and later we realized that the Raspberry Pi was a little tight because uh the hardware requirements after the merge rise a little so this is a device from 2019 and it's a little old and we are struggling to get uh notes in sync with a Raspberry Pi we really uh achieved this with uh Geth and and nimbusbat it's uh it's becoming a little uh harder for us to to keep up with the chain on a Raspberry Pi but uh hopefully uh there is a new device coming up this this month called uh rock 5 that is much more powerful and actually we are validating uh with this with this device and we so it is a really powerful device and it's it's running great so for one side uh Hardware problems uh in terms of of performance but uh from the other side we still can run uh full nodes in ethereum with these kinds of devices so it's running okay I guess that's my turn okay so uh as I said probably I'm the the least experienced person here in in terms of if POS but uh like just barely like one month ago I always pinned up about you know uh fpos nodes to to prepare for the merge everything went smooth so my experience with the merge itself was was really great but in my previous life I was a defy engineer I worked at maker Dao and now I've started running a survey uh within my colleagues you know blockchain Engineers Define Engineers like do you know how if Works after POS and the answers are no they don't have a clue and I I know like 50 about I think the word after POS I think the biggest challenge now is like education there's a lot of like you know talks about if 2.0 whatever but they are outdated because you know this respect changed so much and yeah that's that's a huge challenge I think right now I'm gonna Echo that um just because uh progress is happening so crazy quickly everything becomes so out of date so quickly that people come to us and they say I'm using this uh guide but this thing didn't work and all of a sudden this command is different and so education is one of the things that East acre focuses on and it hasn't really changed before or after the merge we did have a month or two around the merge where we were helping people prepare for the merge but now it's it's just focusing on education documentation and matching people to the appropriate protocol cool uh yeah so uh working on a consensus client implementation the merge was kind of like a four-year project for us so it's been going for a while um it's been stressful it's been fun uh it's been really uh enlightening the merge itself was um it was a little bit of a known event actually which is kind of what we wanted it went better than any of our tests um so yeah it's been um it's been interesting just trying to figure out um what's doing a post merge world there's there's plenty to do but yeah it's picking um picking up the pieces doing a lot of monitoring analysis um keeping the lights on yeah um thank you so much for all the answers uh we uh well we touched many important topics there um and I would start with uh with the node requirement so or also complexity of running nodes right so uh now after the merge uh we need a bit more complex setup before I was doing workshops for running nodes which is like download and run this right maybe a specific comment uh now you need to download two things you need to figure out the JWT path and um and uh also the hardware requirements went up like a huge good restroom an arm on uh for for uh being able to validate on on armed devices but um so uh what what is your take on this what do you uh what do you think is uh uh now how can we tackle this challenge of not just the node requirements it's more like the protocol topic but uh uh especially especially the complexity of running basically different piece of software at the same time uh in terms of validators is very important because we want easy homesteaking experience so yeah how can we achieve this what uh what is happening there and with l2s by the way it's also uh uh it also touches this because especially with optimistic relapse we want to be able to run altu node we want to be able to create fraud proofs uh and should be also easy experience so um yeah how do you how do you see that how do mm-hmm yeah I I think that uh complexity is is not a problem here because uh the difference between uh running a naught permit and Postmates wasn't that different um but what happens what we've seen in in particularly in our uh Discord channel list people that as they cannot get in sync the notes and they are struggling with uh with constitutional execution layer um they they were stressed because they validator were on offline and and that's something that concerned us because uh yeah if you uh cannot get your validator online it's it's a problem because people get the stress I mean uh for us it's not a problem because we are in I.T maybe 15 20 years we are used to it but regular users are not and my main concern is well if there are no much difference between running at home a validator at home uh why I don't go to a pool I take my 32 weeks send it to two upon and then and that's it I don't take this responsibility I don't get stressed so that's a concern in terms of of centralization of a staking yeah and I think that it could be really great if there's some way of incentivizing the the steak at home I don't know a protocol level or how but it could be great because if not maybe people are going to go this way of go to to pools or or this greatest changes that give you this service so yeah it's after the match a little bit of history on me I I got into East acre because I started staking and when I ran a validator I saw a lot of comments in the subreddit that I was using to set it up that like oh you have to be really good with command line to do this you have to have Linux experience you have to be a developer and I was like I am none of those things I I successfully set up a validator and so I started commenting in there and being like no this can be done by anyone there are lots of tools to reduce the complexity of it and we do need to like support those tools uh support those tools more so that they are well fleshed out and have good documentation and e-staker aims to maximize the decentralization of the network which means that we want people not depositing their 32 Ethan all in a centralized exchange for example and so we really just need people to we need a centralized a centralized place where people can go to to get the uh information that they need because right now there are a bunch of different protocols putting out a bunch of different information and it's really difficult to find all of your options all at once and find what the best option for you is like you don't have to be you don't even have to use command line really to set up a node if you use dap node or any of the other avata or uh wagyu or any of the tool staking tools that are built out um and so that that sort of needs to be illuminated a little bit better yeah I'd Echo that I think the um I think the clients generally are not particularly good at making things um usable for non-developers I think the the way that we hire is usually for people who are more systems programmers so people who are like kind of thinking about you know like what it what are the bits look like coming out of your networking interface um and we're not it's not something we're interested in and in generally and not something that we're particularly good at is is like thinking from the perspective of someone who doesn't isn't familiar with using the command line and stuff like this so we really like we do our best to try and make it so that it's um like it's straightforward and it's clean and that usually leads to it being simple but I think it's really a place where um the community needs to come in and help is when it comes to usability so like at Staker with like supporting with people how to use things and I know there's like nice nose and some other um initiatives out there to make it so that you can kind of wrap the creating a node into a like a user interface or something that's easy to use I think it's a really good place for people to come along and and um and innovate if you think about it the amount of people who are like comfortable with um maintaining a node on a command line interface who are also interested in ethereum is such a teeny tiny amount of the world and who has 32e that's like already a teeny tiny amount of the world um yeah so if we if we try to uh keep it at the people who can like run a validator a 32 week validator with 100 Effectiveness you're going to get a very centralized set of operators yeah thanks so much um right so I noticed this totally on the workshops when I'm setting up uh help somebody to step node and the command line comes in and wall just get scared like that's that's crazy thing why would I use this Devils whatever come online yeah that's that's too much for most of the people and uh I really very much appreciate tools which make it easy also to remote arm I mean just flushing it and and running uh it's it's it's uh there are a lot of options like tab download or that make is half of these great interface uh web uh web menu or this two clicks and you have a node so yeah I mean uh there are other solutions that command line uh our our project is uh yeah it's focused on on this kind of of devices or low devices but overall there are other places that make it easy for users uh to to run another or validate so yeah I mean even with ethereum on arm or this eat Dockers kind of stuff it's still command line but it's minimum amount of commands right so it gets like doable even for people who can get scared first um yeah I think that people with without the technical skills would be able to run anything on a nap no yes it's documented it's you you have to follow the guys and and yeah I mean it's better that you have if you have Linux experience for example he said but uh yeah but yeah you could even even run on those with no with no technical skills yeah sure so I have pretty fresh experience of running uh if POS node and actually that the biggest problem for me was that you know you need to realize that now you're running two pieces of software together right like consensus client and the execution client and um I I think like the documentation that I read didn't make it like crystal clear for me and that was like one thing but then you know the merge I I was totally not sure if it's gonna work if it's like correctly configured because it's like it's it's it's it's been working but I wasn't sure if like that you know engine RPC works correctly like I I think after merch it's like gonna be easy to realize that it doesn't work because it's not going to be syncing but you know during the merch I was like okay I have totally it's gonna work like we're not staking so it's not it doesn't take a huge deal but there is a kind of stressful and you know like now we have like you know this modular to to part of the clients and in the future we're going to have new module new layers like I don't know data availability layer on top of eip4844 and it's going to be even harder for people to to run all that stuff so yeah again we're getting there and uh yeah first I want to ask what kind of documentation you read whether you think it was prism okay okay I wrote ethereum org documentation or running node so I I whether it wasn't clear yeah um and but I I was I had trouble explaining it like you realize how much how much complexity I'm adding to the documentation there but like so so just to sum it up what I believe we agree on it's very useful to have uh the automatized setup tools uh which minimizes the the technical requirements also because like I'm not sure whether I set up correctly or I leave some secular security hole there and so on and also the education it's really important huge kudos to each taker for that and now so regarding upcoming stuff now like uh the merge opened many doors opened many interesting possibilities and um like uh generally before for 844 we already mentioned one thing but uh uh but uh are there any maybe okay other than that or uh uh even even including this what are you now excited for or maybe afraid of what are the biggest uh uh biggest upcoming uh things for uh for ethereum which will influence you uh that's coming here okay I I can kick off this discussion so for me uh like I was surprised to learn that um you cannot take more than 32 if at the same time right right now I mean you need to have separate like uh validators evaluator clients yeah yeah and I I found it quite strange especially for like you know this big staking pose like uh and I'm excited that this is gonna be fixed I think in the in the future even in your future I'm not sure here but that was something uh unexpected to me and also like you know the the reasoning behind like you cannot do this right now was also pretty interesting to to learn thank you uh so coming up soon in the protocol will be withdrawals which I think are pretty exciting so there's an EAP out there that is pretty well supported by the teams and it seems like that's going to happen so that'll be um automatic withdrawals to the execution change so it means that uh if you've already withdrawn or if you do withdrawn then at some point um your eighth is just going to appear as in your F1 address and you can claim it and then for people who are still validating once you get over 32f maybe once every couple of months or something like this it'll be just kind of dropped into the execution chain and you can get at it so that'll be pretty exciting I think um I think the panel is on inspiring people to build things is that is that right yeah okay so um the what's interesting with that is that it'll dump the F into a uh an address but it won't trigger any code it won't do code execution so typically like in ethereum if you do an F transfer you can you can when it lands in the account you can trigger some smart contracts to create a log and then people can know that they got the rest um that won't be the case um in this event because it's just just to clarify like because it's kind of country initialized by the protocol right Circle right so that's that's the current specification yeah that's right so there is a slight complexity to it that um when in the early days of um the beacon chain we weren't sure how withdrawals are going to work so um it was a BLS key that you put in as your withdrawal credentials um that's now what people now we decide that we're going to withdraw into an ethereum address so that means that users are going to need to users who have used the BLS withdrawal credentials which is probably most um we'll need to submit a message on chain um signed by that key that that that LX a um ethereum address to receive those funds um so I don't know how far I want to go into it here but yeah that'll so what once users have done is kind of one soft action then yeah the the withdrawals are going to be automatic and they just land in there which I think is really cool because it's less for people to do it's less stuff happening on the network it's cool one thing so you mentioned that the withdrawal can be yet another way to force if into an account without code execution it's like the self-destruct trick now that that's good to know it's actually how coinbase works at the moment so the mining reward is delivered in the same way so it's not it's not too too weird that Nick she had a question no no question but a comment go ahead um following on the withdrawals topic I'm actually really excited for withdrawals because a lot when people when the beacon chain launched there were very few options to stake and since people have staked they have not been able to withdraw and maybe change which protocol that they're using if they're not running a full node um and so there's going to be a lot of opportunity to shift things around and maybe address some of the centralization risks that have been brought up and so I'm excited for that and I'm excited for um any documentation that's going to be written up prior to that in about how to withdraw correctly and what protocols are the best protocols to deposit into if you're not running a full validator at that point or there could be people who have stacked 32 Ethan that time and want to run a full validator yeah from from our perspective uh after the merge uh realizing that you could create a blog in this and in ethereum uh with a device that consumes 10 Bots was like well I mean this is this is amazing this is great this is it feels feels right I mean uh yeah we didn't propose a block yet but we are still waiting so and on the other hand I I am very interested in in two use cases that I I hope that once the merge uh went through and and l2s are thriving that's uh they were part of of the initial uh ethim Design This web3 and sorry uh it starts decentralization it was swarm protocol we have ipfs as well and and decentralized messaging I think that these two cases are are uh I hope that in the coming months or years we get an input because I think that they're very important to foreign yeah I'd love to see non-financial use cases I did really never imagined myself in a fintech role that's not why I started here so and we came to see that and everybody asked you and what are you doing on the blockchain is here oh what can I buy I mean well no no this is not about that I mean yeah that's why why we're saying this um thank you so much for the answers um so yeah we tackled a few important things here and there is actually a big uh discussion uh in the community right now so what should be the priority uh whether it's for 844 whether it's with roles whether we should try to put them together in Shanghai or separate um and uh so your voice of communities folks so what do you think has the priority for you what is uh considering maybe the complexity and whatnot uh what would you uh what do you find um uh more important and doable right now whether it's withdrawals or 4844 which I mean there are many other things but I will mention these two because we get into them yeah so I guess for me obviously it's right for far which will help us with scaling ethereum and I guess uh you know the the teams working at uh working on uh different uh client implementations can have like you know better knowledge about what to prioritize but the funny thing is that prism Pride Labs got acquired by arbitians so I guess uh there will be more consensus there about prioritizing for eight for four uh but yeah like I believe that you know ethereum in the past failed uh scaling in scaling game and that's why uh not only things like you know Solana not only exists but also got traction but also things like polygon right side chains and uh like you know um user my users might not realize the all the different security properties uh of associated with using these different uh blockchains or Solutions so um I think that should be like the the very uh high priority on uh ethereum devs release to to to solve or improve the the scaling of the whole system and uh for eight four four is like the easiest way to to accomplish that yeah I don't have an a strong opinion or that but I would say that if if 4844 means uh scaling and lowering fees I would vote for that yes from the perspective of running the node uh what do you think about it for it for like whether I mean there is a big overhead on the consensus client side uh by storing more data um maybe maybe people can consider it is it that's significant poll I think I think it's the the numbers haven't been chosen for 4844 but it's looking on the order of maybe like 200 to 400 gig extra disk space um there'll be some more Network um usage as well so we're very sensitive to these things and we really don't want to set them high so I'm I'm one of the guys that wants load values so um yeah less less usage but yeah we're still figuring it out but there'll be an impact on disk uh and it'll be impact on network I mean uh I think that the butternut wouldn't be a problem with with network disk so I think it could be we will be good with this I'm not going to pretend I have enough expertise to talk about um what should be a priority at the protocol level um but I would say not at the protocol level I think that just abstracting the technical um barriers from the user or from the Staker is the things that I see like um and interoperability um I don't know if you guys have had any issues but like when you make a deposit to the deposit contract going through like Ledger metamask the deposit contract can have all sorts of issues and it would be great if those teams focused on the interoperability of those things yeah sure I thought I should let uh everyone else speak before a client have talked to to hear what people want because that's what we should be driven by uh and so the question of um withdrawals or 4844 uh for me personally I am prioritizing withdrawals I feel like the merge is not really done until withdrawals are complete because I think that we're kind of um we did the beacon chain on this promise that will let you withdraw one day and the longer we do that it just feels a little dishonest to me so in my opinion the next Vlog withdrawals have to happen uh and then it's a question about whether or not 4844 comes with it um yeah I think myself and the other developers kind of flip back and forth between like yeah yeah we can do 4844 no no we can't do four eight four four so yeah I'm not sure what like we've we've had a bunch of meetings on it this week making progress checking it out be cool if it came in um not explicitly for it or against it um yeah thank you um yeah so like considering the topic should be like what to build after the merge right um uh I think we mentioned many great things uh I mean contributions on the protocol level to withdrawals uh to First for for the testing and uh the stake infrastructure the education there is so many things to do right now uh one one thing I wanted to add is light clients uh which after the merge uh can actually become usable thing and um and uh this can I believe also are largely um uh a better experience of running the node or uh just being able to participate without a third party right um do we have time five minutes um um so um I have one more question but maybe are there any questions from audience as well feel free to raise a hand uh I just wanted to give a space yeah one hi um thanks for all the work all you and your teams do by the way um curious what you think about homesteakers and other folks using Mev boost and if you have any concerns around that yeah I mean I definitely have concerns about Mev boost and Mev um I have been a little disappointed with how Mev um has gone since the merge I think uh it's it's probably not anyone's fault perhaps of the way that we we coordinate the merge but I think that Mev has kind of showed up not in the same Readiness State as the clients have so there's been some problems with Mev booths that have been um a little disappointing from my perspective um things that should have been caught already things that should have been caught with testing so I like I have a lot of respect for the flashbots team I think they've done a lot of good stuff but I'm still very very cautious about any Mev boost um I also worry we've also seen some problems with other um Builders where they've um uh that they've been unable to produce blocks and they can get themselves in a position where the validator is already committed to a block and then they fail to actually follow up on their part and produce it which means the validator can't actually produce its own block so we've seen missed proposals due to um Mev we're seeing a lot of late blocks due to Mev um so yeah and I've also seen some of these scenarios where Mev um like MVP boost or Mev Builders have problems and they don't alert the community quick enough they don't come out with an announcement saying hey like we have a problem and we're not critical so please just turn us off until we fix it I haven't seen that from them so yeah I'm I'm actually a little bit unhappy with Mev in the state of it I think it'll get there my advice to people would be like if you if you cannot run Mev just don't because um like it's better for the protocol and you're also not like you know sandwiching people just trying to like you know sell their money so but I can also totally appreciate why people would do it and I was one of the people that pushed for including the Builder API into like including support like I push for consensus clients to support Mev boost because I knew it was going to happen um by some people and I wanted it to be safe um so yeah that's that's my thoughts on it I'm particularly concerned with uh censorship because uh some uh some relays that are uh running on on countries like USA that you know some regulator says now that uh tornado cash is legal so this these Builders will not include this uh these blocks with sorry these transactions with uh through tornado cash and this bothers me because uh if we are using Med with for example a flash boot uh are we are we as a sensory in transactions uh maybe so if we are using their their layers so it's it's a little concerning yes um thank you so much for the question by the way it was it was my last question the censorship I wanted to drop the negative bump on the end right um uh but yeah so like the the like you also mentioned it like there are because of the Mev booths because of the this relays uh the values are currently censoring blocks and it's over 40 right now of blocks which are over compliant so no good um and there is a the PBS which is I would say just the meme right now it's just like uh yeah there will be ppas will solve it in the future but like the problem is that the the longer we wait the harder it might be to solve right so uh Paul thank you so much for mentioning or like encouraging people actually not to run the Mev boost or I would say uh you can use non-sensoring relays uh there are these options uh do you have do you have any other other sorry the idea is to include I may be in in at protocol level I'm right or uh so I as in what I said about including in the consensus client yeah yeah so the so what so before the merge we had Mev Geth which was like if you wanted to do Mev stuff you needed to run this Fork of Geth in order to like to do this stuff and a lot of people were running it which meant that a lot of people are running um a fork of a consensus client by a team that is not considered to be client developers um so I thought that was not great so what we did with the consensus clients was was to add was to get flashbots to to instead create a separate service and then we support an API to talk to it um so what this means is that um like flashbots no longer needs to to Fork all these against census clients and then like either do a lot of work or jeopardize client diversity and it also meant that now the consensus clients choose when to use these Builders these Mev Builders so we have um like fail-safes in there to say like if the chain is unhealthy in a particular way then let's just like cut Mev uh because it could be the problem for it um so at the moment like the protocol doesn't support it the implementations kind of acknowledge that it exists and try to and try to do it safely I like my um the way that I think about it is like drug use like it'd be great if it didn't happen but it's going to so let's make let's make it safe um yeah and then yeah and then in the future the protocol wants to support the ability for the validators not to be the ones that that build the blocks so whether those blocks contain Mev they probably do but that would be a bit of a better solution to it anything to add folks uh if not I think we are yeah we are uh passed over time and uh uh I would like to say a huge thank you uh not just for attending this panel but especially for all the work that you do uh in your project it's it's uh what pushes ethereum well to succeed right so uh thank you so much again uh for your work for being here and uh yeah thank you so much for attending the panel thank you [Applause]