AMA with Alex Stokes | EPF5 Office Hours

recording going yep recording is going yeah awesome yeah very well uh so we can kick it off um yeah we have another EPF five office hours here uh with another very special guest today uh Alex thank you so much for joining um yeah it's honor to have you here uh so yeah I guess many of you I hope maybe all of you know Alex uh or at least heard of him or at least saw some of the ACD that he's been running maybe some of his work um uh on uh uh the MV on uh uh the uh con specs in rest uh or maybe just heard of him as a CO researcher uh maybe saw uh the EPS uh the study group told he did so yeah uh you're right then so I think you are familiar with Alex uh it's uh and today I hope to dig deeper to give you guys opportunity to have some um to to uh some more questions for Alex to get him know better and uh explore some more parts of the protocol consensus and so on uh but first like yeah I want uh want the fellows to to get to know you a bit better uh give some context and uh and learn a bit about you uh yourself like if you can maybe um before we get to the questions uh if you can share a bit about your background uh but how you get into etherum what led you to this journey yeah yeah definitely yeah thanks for having me excited to be here good to see you all um yeah my crypto Journey it's uh I feel like everyone has a very like unique and interesting story um which is one of the cool things about being in the space so let's see let me think where to begin yeah I mean so I I guess I first heard about ethereum just like on the internet probably like Reddit or like Hacker News or something a very long time ago and I was like oh this sounds pretty cool you know like I had been following Bitcoin for a bit and um yeah this sounded like an even cooler version of Bitcoin with all these smart contracts uh but I didn't really like I guess at that time I wasn't like as into crypto as I am now so just I found it like technically interesting and it was like uh yeah like you know definitely something that picked my curiosity but um yeah it was more kind of like a remote thing and yeah at the time I was working in like some tech startups um this was many years ago at this point um but it just kind of got to the point where yeah I kept hearing about this ethereum thing and um it really just got to the point where like I just wasn't even like thinking about my day job because I was just like thinking about ethereum and like all the cool things that could happen I think I first got nerd sniped like with yellow paper um cuz I think you know like you through it and then I think I got to the point I think there's like an appendix or at least there used to be of like the pre- compiles and I was like oh okay like you know you could imagine like enumerating every possible computation you know in some sort of shared State like this and that just like really fascinated me like intellectually um so long story short I guess I got nerd sniped and then I was like figuring out ways to like uh yeah just get more involved and and sort of navigate that and really how I started was just like open source work um you know I've done a lot of like engineering in the past and um yeah it seemed like a good way to get started like I didn't really like know anyone or anything really about ethereum so much um but definitely there was like many opportunities for like open source engagement and so at the time that I like you know started getting more involved with ethereum there is this notion of like e2.0 and you know fast forward maybe even like four or five years and that resulted in the merge and how things look today um but back then like there were many different ideas around How We Do prove a stake like we were even at one point going to have a separate blockchain that would essentially be the proof of stake layer that would then just finalize the proof of work chain so um you know point being there is it's to see like how the research evolves and like how these things sort of you know develop in time but yeah I think the like first concrete thing that I did was actually part of this prois system I was just describing the thing that we wanted was I mean so you've probably heard another context of account abstraction maybe with like ERC 4337 or 7702 and things like this where you know we have like a counsil on ethereum today and they do some things but you could imagine them doing a lot more um a very simple example is even just this notion of like a multi call where you know when I make a transaction today it basically just launches like sort of one set of actions you can imagine instead that you know I want to do like multiple of these actions within like one transaction simply because like um you know it might be cheaper or more convenient or yeah just easier for me as a user to like do this a good example is like if I make a token swap on Unis swap right now you have to do this like year20 flow where you have to approve the thing and then you have to make the trade and like this is just annoying um and so account abstraction broadly is like thinking about ways to change how accounts work on ethereum so that you can um you know get better ux basically around these things so you can imagine doing this approv and transfer in like sort of one transaction rather than two so point being is there were similar ideas at the time around having account abstraction in this way but for validators on in the proof stake system and what that would look like uh in this case is like right now we have uh the proof system with the beacon chain and everything in ethereum today and we use this uh signature scheme called BLS and the the sort of Desire at the time was to like also have like a sort of abstraction around the signature scheme so you could imagine a validator signs up with whatever they want maybe not just like BLS but like it could be anything and like concretely uh anything that you could express in a smart contract and so that was like my first open source contribution was um we call it like the Purity Checker Purity in this sense refers to like uh like notion and programming where it's the fact that you know you want a piece of code and you want to make sure that it can't change the state and the way this comes into the account obstruction bit for validators is that it would um you basically need a way to have the validator register on chain the validator can then point to some smart contract as their like signature verifier and then what you just want to make sure as part of this registration is that uh the smart contract they point to was like pure in this sense the reason why is because signature verification needs to be terministic you can't have like you know you could even imagine something crazy where it's like as a validator uh want to sign my proof of stake message and I want it to be conditional on like the token price on Unis swop or something right like this is very possible um but token pric is all over the place and so you know a signature might be valid at you know today but then maybe and say the next let's say it's valid now but then the next hour it's not anymore right so this doesn't make any sense and anyway long story short I WR some code to do this you couldn't actually do all this in solidity at the time and I still think you can't so I actually had to like drop down to evm like bu code um which is just like a fun way to like learn more about the evbm and I ended up writing it in this like weird thing that I think actually vitalic or someone wrote many many years ago is like a sort of like high level language rapper around Ean bik code uh called I think it was serpent um this is like all very deprecated at this point but anyway it was very fun um it was an exciting way to like get more involved with ethereum and the research and development there and and get engaged so I was doing that and this is probably around the time I don't know this is roughly around the time with like Devcon Prague and I went to like Devcon Prague and I saw like oh like there's such a big Community here and like there really was something special so that just kind of pulled me in deeper and deeper um then that I think would have been 2018 so then I started the EF in 2019 I was brought on to the EF to work on a python client actually the time of this e 2.0 stuff um that client has also since been deprecated but that was like a funable Arc and yeah then in the meantime I've like drifted kind of between more research things more development things and yeah from there you can kind of fast forward to ethereum today we have the merge um that's happened we have Pro stake now and yeah now we think about you know the next phase of ethereum's evolution whether that's like scaling data or like C resistance or things like that and yeah maybe I'll just pause there that was a lot yeah yeah um yeah you went pretty uh into weeds and details there but I really appreciate it actually it's great uh thank you so much Alex um I mean you also getting to the the current state of things like I'm wondering um yeah so what are you working on these days what excites you now uh What uh what do you personally work on about maybe just you know find exciting in the ecosystem yeah yeah um I would say like around the time of the merge like very much like meev had become a thing and it was like something that everyone sort of we all sort of realized was like a thing an issue on ethereum and that we need to like you know focus on it um so I would say I like kind of got nerd sniped by m as well um and the angle I approached it from was more like cons security like I guess that's kind of always been my thing is I like okay we're going to have this new consistency protocol for ethereum like the thing that must ensure is that it's secure right like when we think about blockchains like what is a blockchain good for it's good for getting these like commitments basically in this like you know decentralized or like credibly neutral way which maybe we can jump into what that means but ultimately like you just have the system that gives you a set of facts and sort of the guarantee that you have is that no sort of single party can like change the facts on you and as long as you have that primitive you can build all of the stuff on top whether that's like tokens or like stable coins or you know Dows and like different like governance Innovations and experiments like you know and this sort of then points to all the crazy stuff ethereum can do and like everything you can build on top so point being is you need to make sure consum is secure and um yeah so that's kind of I think the frame generally with like my work is just like making sure that you know ethereum is secure um separately now we need to think about is ethereum scalable and all these other questions um but yeah so because of that yeah done a bit of stuff with like me especially like with respect to like infrastructure and like the protocol level um and then what that looks like today is like I talked to like a bunch of these different relays a bunch of different Builders and like teams in the ecosystem there and yeah I mean I think over time I've kind of shifted my work to more of like a coordination role um simply because it's really important in a space like ours and yeah so like lately um I've been pretty involved with like yeah alra dos and yeah just I guess engaging with like the AP process and the governance around that yeah awesome um yeah it's it's a lot of things that uh Alex is touching so if you guys have questions about any of those uh feel free to raise your hand feel free to ask I see we have a question in the um in the comment but if you just want to read it Kya feel free to speak up uh yeah hello I can just ask this question uh yeah this um this question uh I can hear it a lot like discussing it and many people concern of it uh how to avoid stake in Monopoly um yeah thank you right yeah so It's Tricky so the the question here is uh I don't know Dan did you need to did you want to hop in right now no it's a queing for like okay okay cool um right so what this is getting at is that so we have this Pro estake system today right we have this like notion of a validator and these again are these protocol actors that are providing this like you know I said a few you know minutes ago let's say uh this process of finality and like really this is where we get the security of ethereum when we say finality what we mean is that um once this like this process has happened and it's this thing that's continually happening but once you know once it has done its thing you get this guarantee that you know some amount of the stake uh like the eth that is sort of bonded to secure their protocol is basically saying yes this is the canonical chain and there can be no other chain um and moreover you have these like very strong sort of we call them cryptoeconomic guarantees around the fact that if there is some you know um malicious behavior with respect to this guarantee there's like steep penalties you can apply and so that's all well and good the interesting bit of course is when sort of theory hits practice so to speak um because now you have the system uh the way that you incentivize validators in the system is you like give them rewards for their behavior and you know that's all well and good but then you get to this situation where now there's almost this like Financial element where um you get paid to stake basically um and this I guess is a kind of double edged sword because now you know people are directly advis to do this like people even then start to see like staked eth as almost like a financial product or like an investment um they start to go and and do all this stuff and what happens then is you know it almost becomes competitive where if I'm sort of better at operating my validators I can maybe get a higher yield and this then kind of circles back to like me and like now with reaking and like all these sort of things where you know and again this is like a very cool thing about ethereum is it is so like flexible in general is you can build stuff on top now the thing then is that these these like other applications kind of funnel back into the core protocol and have direct implications for how it looks like and like in particular what can happen is again as a validator like as an operator of like some amount of the stake if I'm better than others at you know doing my job then basically I can you know you can imagine I get higher returns I can then invest that money back into getting even better and then this is where we get to these like centralization vectors and this is what you'll hear people talk about all the time um because yeah basically more sophisticated operators can do better they do better so they have you know more resources to get even better and what that means is that if I come to the protocol you know I can imagine maybe rather than running my own sort of you know note at home I want to go with one of these big operators because it's simpler it's easier and they'll just give me you know more money if that's the thing I care about so if you play this process out you get to like a very sort of concerning instate where you know you can imagine there's only one or like a few like very big operators who like just do this professionally and then all the stake sits with them this is a problem because it directly goes against like this core value of of blockchains and which I said a second ago right this is what we want is we want this ability to have these like very hard guarantees around like the commitments we make say the ethereum protocol State at point you know yesterday was this today is this and the issue is you start to erode these guarantees the more centralized you know this sort of the set of operators of the protocol is and this is like the Monopoly here so when we when we talk about this stuff um this is yeah this is why it's a problem It's tricky because there are just like natural you know there's this this notion of like economies of scale where basically this is just you know you see this even outside crypto outside of E I'm practic it's just this is everywhere in the world that usually you know many things are just cheaper or easier to do if you do them bigger and so now there's this tension between you know this like phenomenon in the protocol while still preserving this like core value or like property that we want and yeah so how can you avoid it well it's very tricky there's a lot of research around this topic many people think about it um I think the most most like probably the most like compelling answer is just changing the actual staking economics where uh this is no longer an issue um and maybe even before I get there like there are other ways we thought about like sort of isolating these different concerns so maybe this is a better point to start actually like so these these these you know economies of scale exist and so you could say okay what can we do about them we can either say okay they're here and we just accept as a fact of life and move on or we can say Okay are there ways to like isolate it in various senses so that it doesn't have impacts on like key parts of the protocol and if you follow any any of the MV stuff or like you probably heard PPS or like proposer Builder separation this is like one way to like have this separation so you know I said MV is like one of these domains where more sophisticated actors can you know more readily take advantage of these economies of scale and yeah you could very much just it's actually like kind of a crude idea you just imagine splitting off in some way so like if muv is a problem just try to like silow it away from the vator set and this is exactly what PBS tries to do um so that's like one answer there's more speculative research around like you might have heard of like APS or like a to a Testo proposer separation and this is kind of even going a step further so like with PBS we separate out this like execution Builder role and that's kind of where this whole M domain sort of lives and is generated from separating that from the validator you can then imagine there's even now a notion of like testing and again this is kind of getting down to the weeds but as part of this process of the perate consensus you perform this this Duty called at making an attestation or testing and you could even imagine them separating out the proposing role so like the the role of building blocks today and actually like getting them on chain you can separate that out from a testing and so you can imagine then in like as we go through time there might be ways to like carve up the protocol in a certain sense that uh you know isolates these more problematic issues from um the very like core core parts of the protocol that give us again these guarantees of decentralization of resistance and all this um via testing so that's one way the other way that I was kind of hinting at a second ago is like more economic you can just try to change like either the reward schedule of the protocol or like the ways that the incentives around being a validator in the first place that uh might naturally also like sort of uh put back pressure on this like tendency to monopolize and there's a couple different ideas around this I mean one thing is you could just imagine making it less attractive to stake so you could literally just like lower yields um and that's like a pretty simple solution but then again at the same time we want like the reason there is this like notion of yield in the protocol is to attract people to stake and to actually secure ethereum and so if it's like too low then maybe we don't have enough people securing it and then you start to harm the crypto security we have so that's an issue as well um so yeah it's it's not quite clear what to do the other thing here that's worth mentioning is just that you could imagine other ways to think about how the staking actually works like in an economic sense and this is where something like s targeting comes to play where uh what we mean by this is basically say Okay regardless of what the supply and demand is to actually like stake and secure the protocol you can just say um there's essentially an import call Target let's say of uh you know we want this much this much steak uh say 20% we want 20% of all each steak just for like a random number for the example um there's a way to like have the protocol work where it looks like this and then what how that would like look like in practice for like a Staker is that um you know let's say that uh there's less than this 20% number Stak then like yields are higher so then people are incentivized to come join and participate if it goes above the 20% Then basically yields are actually negative um in the sense that like you're paying to stake and in this way you can like tune the protocol to like have an equilibrium around whatever this this target is the sck Target that we have this I mean yeah so how does this help with monopolies well the first thing it does is it it has this like pretty hard constraint around only so much eth Stak um again the like sort of pathological scenario is you imagine all the eth is Stak it's Stak with like one saking operator say Lio not to pick on them but just for an example um we want to avoid that probably and there's very much active debate around all the nuances here um but yeah so you have it in a way where even if there is sort of a monopoly of staking operators or these staking pools you could imagine constraining them to like some subset of the E and so then you have like say 80% if you know if our Target is 20% and there's 80% that just is not rational to stake and so then it's available for other purposes whether that's for like more money purposes or for using defi or you know buying NTS or like whatever have you so yeah long story short is it's definitely a tricky problem there's like a lot of Rich research around how to handle this and yeah it's not not super clear honestly like I think there's many directions to go in and there's very much active debate around what's the right way to do this like as ethereum grows like more more players come to the table it then starts to become a very political sort of you know question because at the same time we could say oh you know maybe economically or like if we zoom out like some stake targeting regime makes a lot of sense uh but then that's directly basically telling the staking pools like hey your revenue is capped which like they don't like right so it then becomes uh not so much technical conversation but a different one thank you thank you so much Alex it's been this was tough question to start on went from like Zer to 100 pretty fast it's a good yeah thanks for elaborating on all of that yeah uh yeah it's been it's been a tough discussion and I'm sure that uh Dan has something uh related yes uh not uh not to ease the whole uh topic which was very well uh brought up so in the in the context of of not of the economic part but of of separating the validator role with uh epbs with APS my EPF project is eods operator in shrined operator delegator separation so it's basically a a conceptual design and hopefully a specification of introducing delegations in protocol and separating you know Capital providers from node operators and the specific question would be if um you know with your expertise and and knowledge about ethereum sorry uh what would be the validator attributes from the Val from the current uh uh action set that could be uh considered not strictly related to finality so not strictly strictly related to Bringing economical uh security because at least in in what I'm uh prototyping with eods that could be abstracted from the validator set and added to a delegator set that could back a network of light validators you know bringing Security on chain with censorship resistance but there are other roles like I don't know whistleblowing for example which so can you think can you I'm trying to pick your brains on this if you can help me because this is uh the point of my project where I identify this attributes that we can we could you know consider removing from the validator set and moving to the to the light light consensus so to speak right yeah and so yeah it's it's nice that you brought this up uh I think that's another you know actually good point to my previous uh answer around like what are ways to think about this um there's this notion of like rainbow staking from barnab mono and like some of the rig team and yeah barab my mentor yeah okay great um and so yeah maybe just to like briefly explain what this what this means is like yeah it's kind of taking this this notion of like okay can we sort of analyze the different parts of the protocol and imagine ways to like maybe unbundle them and like rebundle them in another way to uh address some of these centralization vectors and so um right you could imagine that you have like um a even like a very lightweight sort of validator set that's maybe even just doing a testing or something like this um you could even maybe go a step further and say there's even like validators who are just doing like uh sensorship resistance functionalities in terms of like just doing inclusion list or something like this so you know we'll see like it's a very interesting research and I think very exciting um at a high level there's a tension between things being like too complicated basically because like the more complicated the protocol is just the more hard it is to secure and so um yeah I guess there's no clear path there's no clear path yeah yeah but then to answer your question uh to answer your question so there's a couple things I mean we can go down like the PBS route of taking out like execution building um we can go to like the aps world where now there's like an execution proposer and that kind of subsumes PBS then you have a consensus proposer so like that's kind of a distinct thing is proposing like the consensus block um another thing then is like the role of a testing where this is also kind of the aps world where it's like you then take out like the the consensus functions from even proposing blocks in the first place there's like a few other minor duties that you could at least think about um one of them is like sync committees so like there's today we have yeah we have validator who like basically provide the infrastructure for like clients and that's like a different thing that you could imagine pulling out um as some separate thing um yeah and then yeah like I think it also would be worth thinking about like uh any sort of like inclusion list design or like some CR design where like that could even be like a separate task that like uh validators sort of opt into and yeah I mean even when we put it like that it kind of points towards restak and then you could even start to imagine like with some ODS thing it's like somehow starts to even enshrine like um you know these different like reaking ideas people have that's like all kind of worms but it's it's an interesting uh L like like an in protocol AVS somehow yeah so what I what like a delegator uh stake uh delegating value to a node operator that that participates in finality and gets some sort of a liability proof that he delegated and then uses that to back a light operator that's nonf finality so non non non slashy so to speak that's some sort of RIS taking yeah and like something else this kind of makes me think of is like um like another use case that's would be interesting is some sort of like native LST where like you could imagine like you pick a part yeah the right parts of all these responsibilities in a way that um you could now have people like have their own LST and so this is the whole thing like this is I guess even another point of like how do you avoid staking monopolies is like if the issue is these like big staking pools with like their own sort of bespoke lsts or like they're staking derivatives you could imagine having this in the protocol somehow um this is tricky because um the whole like the reason LST is like coherent or makes sense is because there's like a similar risk profile with respect to all the validators that the LST represents and so then if you imagine a world where like any validator can make their own LST then like there's no way to know right off the bat like if they're fungible at all and then it is not a useful sort of construction if there's like a million validators and a million like everyone has their own LST clearly yeah so I have I don't know I've thought about some things around here and like D I think had some research Mike norer had some research around this like two-tiered sticking idea that I think gets in the this direction um so yeah that's worth looking at another thing would be yeah like I I mean I even had the idea of like you could imagine like a Unis swap hook that takes a proof of like Val performance and then that is then like somehow rating like the risk profile of a certain batch of stake and then you can imagine that's like underwriting um an LST and then from there uh yeah maybe you would start to get like buckets of things and you start to have some like uh even sort of more decentralized LST than what we have today pus proposed something but it was in the epbs context of uh of clients uh building their local rating system for uh for Builders we could also I I also consider some idea of having a validators validator rating system that the delegator can you know opt so I want to delegate to to who what validator what node operator so uh you know we could have like a rating system and then just go with the validators that do not work with censoring uh uh Builders or with timing games playing uh Builders or or stuff like that yeah it's it's going into the Weeds now so just just for my own uh reiteration so it's uh uh you said s committee uh something that it's absolutely not related to finality uh uh whistle whistle blow whistleblowing yeah yeah that's an interesting one um like think about the incentives right well it's more just like anyone can do it and like yeah there is this sort of like dilemma of sorts where it's like yeah it's not really incentivized um well okay so you do have like this whistleblowers reward but like you know you can imagine that like no running into SLR mayet and like that like I don't think it's the case it's not the case but you could imagine it is and then that's like kind of a huge hole in the security model um so yeah I don't know maybe you could imagine some like part of this ODS service that's like yeah somehow incentivizing more directly or more strongly like uh this whistleblowing um and and and inclusion lease that that's the first that's the first major stuff obviously yeah okay Alex thank you I don't want to to take too much uh of the fellow's times uh time thank you very much uh maybe we can chat on Discord more thank you yeah thank you both um yeah we dived uh pretty deep into into the consensus there I wonder if anyone else has a question related or maybe this switch topic um yeah p uh hi Alex my question uh was more on your role as a coordinator for these all core developer calls what your like biggest takeaway or most surprising thing about development in ethereum compared to your more traditional role uh starting in startups yeah uh okay this is there's a lot um I think the biggest thing is just like understanding and like having experience with how like open source development Works um and it's just like such a different frame of mind from like everything else uh you'd see in like more traditional context like you know a for-profit company or like a VCB company or something like this um and some of these things maybe are specific to like crypto or like blockchains and how like ethereum works but you know at least how we do have it on ethereum today is like you have a bunch of you know different people they have different interest and different reasons for being here um and so yeah like that's great because you have like more perspectives and like just more engagement and like it's the system overall is more resilient like the more actors there are but at the same time you know there's not perfect alignment and then there can be frictions there and like those can be interesting to navigate so yeah I think just like like seeing the system having some view of it at like a high level is like very interesting just to like see over time like how different decisions are made or like why they're made and how that plays into everything um yeah yeah and especially in like a context like this where it's like very decentralized like it a lot of it just feels like cat hurting sometimes where it's like people usually like you know generally there's like reasonable suggestions made saying like all cevs and so um it's not even so much that like it's not even so much about like needing to make a decision it's more just getting people to even pay attention in the first place which I guess is like yeah I guess the point here is just like attention is finite and yeah it can be hard even just to like know that someone's even like seen a proposal let's say um so yeah that that's pretty interesting and yeah I don't know I mean I think I think that's like the biggest thing is just like the scale of the process the scope of the process and like seeing how it plays out and um yeah again both like in a very nice way because you see there's like so much like interest in engagement with ethereum um and then yeah at the same time it can be uh yeah there have been some spicy ACD calls like Petra is a good example where it's like you know the fork is probably too big and like we need to do something about that um but then you ask different people and they're like yeah well you know we think EF should be in be for these reasons we think you know Max CB should be in for these reasons and then like no one's wrong it's just more question of like scheduling almost where it's like what's the sort of highest priority thing and like when and like when when do you do these different things um yeah so yeah it's it's fascinating thank you thank you it's interesting I'm glad we could uh ISO the topic but um guys U please raise your hands otherwise I will take the opportunity of questions for myself uh yeah K again yeah uh my next question won't be so tricky just about my project I'm working on puras metrics it would be nice to hear your opinion is it possible and does it make sense to standardize metrics across clients uh yeah no it definitely makes sense like it' be super valuable um this comes up in many different other contexts like given just like the Json rpcs for example where it's like you know we want anyone to be able to use any client again for like security reasons because we have this like very nice sort of multi client security posture um but then if you like go to use one client and they have some nice RPC feature uh if the other doesn't support it then like that's not nice um so generally there's like a lot of value and standards um and similarly with like metrics where you know there's things that we all want to know about with respect to like I don't know if I'm a validator I want to know like what my adastation performance is and so like I think at this point all this y clients expose this in some way um some even have like very nice dashboards with like raana and these different like metric systems that are really nice um so yeah then you can imagine the same thing for for pios and like different things like that and so then you want to see like like you know even just knowing like bandwidth consumption and like you know when I'm going to sample sort of this column or that column of like the block that you know uh is it happening like how much bandwidth am I using and things like this um something we'll probably see is like some ability to like customize how much data you're downloading like under pios and so you could imagine like maybe you have a default setting for like a solo stier where they're only downloading like the minimum amount of data required for security uh but you might actually say you know for example say you're like an L2 and you actually want the blob data in its entirety you'll probably need to subscribe to like a majority if not all of the subnets and so again like having metrics around um what your node's actually doing with respect to like your preference here is like super valuable so yeah it'd be really nice thank you nice anyone else maybe feedback for the project or or any other question there's one in the chat here what implications would standardized metrics have on attacking the core protocol hopefully none um but it's it's like it's good it's good to think about like nothing with metrics like nothing really sticks out with metrics but there are things where like you could imagine like even like again because of like the process being so open and like you know public uh you could imagine that like some you know someone here is like looking at something with like some client and they realize that uh you know there's some discrepancy between like how client a processes blocks compared to client B and in the limit this could actually be like a Dos Vector um and for example if it is a Dos Vector that's like on Main net you don't want to just like go to Twitter and like broadcast this because then you know we would have chaos someone would then go and Trigger it on Main net it would be like a whole thing um so I guess that's like important to say is just like try to have this like adversarial mindset kind of throughout everything you do um because you know you you could find an issue and very much it needs to be fixed Say by like the the developers of this client um and it could be multiple clients but either way there there could be something here and so it's like good to keep that in mind that like um yeah like we're on the bleeding edge the software is like you know very Dynamic and influx and like there are there are things um and so if you find something maybe yeah at least be mindful of the fact that it could have like very big security implications and so that just needs to be like managed appropriately um yeah otherwise yeah with respect to like the question like standardized metrics I don't know I feel like the things you standardize like I guess one is that they just have enough scrutiny like otherwise we wouldn't have agreement on the being standards so like it's things that people feel like you know they've been looked at a lot so they probably don't have these sort of security considerations and yeah like metrics are very sort of like you know it's more just a description of what happens and so you know I I could imagine a scenario where like metrics show you like maybe they call your attention to like some issue uh but yeah I think otherwise it's it's pretty okay U maybe more of a Jason RPC like some debug namespace which is also used for some or things like metrix might trigger some heavy computation that that's that's a DDOS VOR there like in right these sort of things I can really see the metrics there but I like what you said there like think about security implications of everything you do like you know any changing the client basically yeah um yeah CH you wanted to elaborate on that you um not elaborate for say but like there are there are these attack vectors that uh uh nobody executes like nobody attacks using that like specifically targeting validators in the current ethereum Network right uh anybody can do it but nobody does it that's another scenario but I just think that with metrics uh linkability of identities on the blockchain would become much more easier uh and I think it would it would open up you know I I'm not sure uh if it would but it feels like more information taken out would allow me to l link identities to uh ethereum addresses and like directly Target them in a sense right yeah and you could imagine like some periodos Metric for example that um you know maybe I get some data on like uh which columns I get from which peers at a particular point in time and then you can actually use this to like correlate validator identities to um particular nodes just given the way that pios works uh and yeah this would start to like give you this information to like have this Sy ability and then yeah like if you have the IP of a validator you can just wait until it's their turn to propose you could even like on demand Doss them and then suddenly you know there's no block for that slot and like ethereum has gone down in some sense so we definitely want to avoid this and like V privacy is a big thing um people definitely think about it all the time and um yeah it's just kind of one of these like cat and mouse games where like you just need to kind of be ahead of the attacker and um yeah just just keep it in mind uh yeah like it's much broader issue of the Val privacy like you know metrics are generally used locally like it's some end point on the client that you listen to with prome influence DB and you just follow what your client is doing it's just uh some it's like logs of your client but more variables basically so but the general privacy of validators like the what Alex said like the IP addresses and the the valid keys are all public it's all all linked to your deposits so uh it's it's not ideal right now like uh yeah but yeah I haven't seen we haven't seen any attacks so far like you know after after so many years without any attacks I can think of like maybe it's just because secure somebody had to try at this by by by now right I don't know um um any any other question folks we have still 10 more minutes to go um um I had a question earlier uh going back to the um to the monopolization to the uh pools and stuff because a few months back there was The Proposal on the um attester penality they correlated correlated penalities uh which uh was pretty I don't know like it was it was pretty cool at the time I remember uh Tony had a feedback for it and there was a conversation going couple of months ago but I'm not sure if it's still happening I'm not sure how it uh was is the current state whether it's still still active research or yeah what what do you think of that yeah no I mean it's it's really interesting work um the idea would basically be to like have um yeah let's say like additional penalties if you could like detect correlation within the protocol um and the way you'd start to do this is you could actually look at attestations on chain and then start to understand like you know essentially if some some group of validators like moved in unison with respect to like their adastation behavior and then in doing this you could then pretty confidently say oh well they're probably like a single unit um and then if that unit has like some you know uh say they like uh you know Miss the head of the chain in their attestation or like attest to something that is non-conical then you could actually like penalize them even more and that's like the correlation penalty and the idea here is to have this it's like it's not really an incentive as much as it's a disincentive to like have like a shared setup right so to not correlate and um yeah no it's it's uh definitely interesting work and you would imagine that you know the success case is one where now as an operator you have to think about okay even if I am operating many validators I still want to have them in some decorrelated way to avoid this penalty and you know that could look like say I'm coinbase and I have a lot of like validators that I run I could imagine putting like some even across different like AWS domains or like different data centers um even across different continents and there's like all sorts of things you can imagine around like resilience here um and then you know on the margin this like goes all the way out to even like home stickers where it's like there's like a home sticker you're probably fine uh imagine though that you're like sticking in like New York City and there's happen because you're in a high population Center there happen to be many other solo stakers with you so like that's like an interesting thing to think about and um then if you're like a solo Staker in like you know some random City and like Brazil like yeah I don't know Brazil or somewhere like say South America like uh you know you're you're probably fine here and then part of the the work there with with like that EIP or like the proposal is to figure out like the right parameters to tune it such that like you don't overly penalize these like very uncorrelated people um while still having this this like sufficient disincentive for uh these like sort of bigger scale operations yeah awesome thanks thanks Alex U yeah it's great to hear that uh it's it's in progress uh it's been of months since I last about it thanks uh Dan has some more questions only if there are not other questions from fellows I just had a question in popped up uh so max CB and what follows after maxb so we get some reduction of the validator set uh we get consolidation yeah I reckon that's that's something that uh we all should be looking forward to yeah I think so I mean one thing even already is like there isn't like a strict incentive like there's not a clear benefit necessarily to consolidate like there's no in protocol rewards for consolidation so like that's interesting um you know the people working on this uh EIP have definitely talked out like reached out to the big sticking pools and like at least given everyone like a sort of soft you know thumbs up that they will consolidate um but yeah like that's an interesting bit and yeah I mean what's next um so I mean I guess there's a couple things one is it kind of lines up ssf work which is like really exciting because with the smaller batter set size uh it's much easier to like do like to get actually two finality within a single slot ssf that's that's what we mean um so that's pretty cool and yeah there was something else but I just lost it no no I think ssf is I think this was where I was also thinking about and fant's uh orbit ssf and that tackles the consolidation you know uh reward and but and again just high level discussion does Max EB gets us one step or two steps closer to ssf because we see ssf as a future event in the in the in the far future but it it just might be closer than than we think or not I think it is closer I don't know like ssf is this thing where like at least in like the research Community we've been talking about it for like even a couple years now um like even this year myself I've like been like oh I need to like change my perception of this because like I think it's actually much sooner um um than than like sort of what I at least I was thinking previously especially like orbit and things like I think there's parts of ssf as a vision that we can start to do like in the next couple years definitely um that give us a lot of like nice benefits and like yeah it's I think very much more accelerated uh timeline wise than uh yeah at least that I was thinking even over this year let's say um so yeah super exciting there and uh yeah we'll we'll see we'll see what maxb does how much people consolidate and uh yeah that'll just kind of te us up to make ssf easier to ship yeah awesome um Bon uh hey yeah thanks so I wanted to get your thoughts on uh like the current state of light client and what the future could look like and what's uh like missing in your opinion right now yeah this is a good question um so like clients are good because you know we like the the instate of ethereum is like at least what I would like to see is like this ability for like a very heterogenous set of nodes to like engage with the protocol right so you could have like maybe these more like more professional like very big operators who like are offering the service to like uh either produce blocks or like you know perform the staking function and like come to consensus and things like this um separately then could imagine like sort of smaller actors who still want to like both sort of read and write the chain so that's like both validating the chain and knowing like which you know if there's like some state route or like block root of the state of the chain right now that it's valid um but then also like interacting with the chain in the sense that like I might not necessarily want to like run a full node to like send my transaction and there's a lot to say here um statelessness is like a whole thing that's like very relevant on like the spectrum of like node capabilities uh that's very exciting with like veral and things like that and then you get into like I think what people more traditionally think of with like clients where you know you could even imagine that like you know metamask or like uh Robbie or like whatever like your wallet actually has like client in it and this is you know a huge Improvement on the trust model today because like if you go to use metamask you're basically just trusting consensus that like the chain is what they say and like when you go to make a transaction like everything checks out you know very obviously then this is like a single ponent failure uh and kind of you know it it's not it's not the best way to interact with ethereum at Le like we can do much better um and the way we do much better is by having yeah we like take more and more of this like uh sort of sovereignty almost like into the end user hands um and the way you get this like is with uh is with like clients so they're super important and yeah I mean where are we at I mean I think a lot of it is just kind of blocked on even getting wallets to start to adopt these things in the first place there are there are some like protocol things like again I mentioned like statelessness or like ccle and that that whole line of work um there are things we can do to make the chain easier to validate and like in the limit you even get to like ZK EVMS or things things like this where you could imagine that like what does the execution layer look like in five years well maybe the whole thing is just essentially a snark or at least what you're like what most users see is just a snark so like rather than have to like download every payload and execute it and make sure that it's you know it's valid with respect to the protocol um you can skip all that you can just like be on some gossip Network you can get a snark that like hey here's the next state route of the chain and here's a proof that it's valid um so that's kind of like the end game of a lot of this and then how that ties into like say this wallet use case is then yeah your your wallet just has a like client bundled in it where then it's consuming snarks like this and um it be now becomes a very lightweight to like validate the chain all you have to do is um yeah download say a small proof and like the state route maybe and then uh you're kind of Off to the Races um so yeah super important and I think there's like a number so yeah there's things kind of across the ecosystem that that kind of need to happen to like get to this nice Vision um but people very much are working on it and we'll get there so there's that maybe like yeah even zooming into like today uh I call it Helios Helios is like a super awesome like client implementation um and I think even I saw on Twitter today or something like uh they're trying to reach out to more wallets and even think about integrating um Helios into say metamask or something like this and um yeah then you just you know sort of keep iterating and get to hopefully this like nice Vision that we all have around like a super secure and decentralized ethereum yeah awesome uh it's great yeah I was I was wondering if you bring up helus uh it's a rest client and I think youve also contributed there right um little bit Yeah so it's it's great we had a we had it part of Fellowship as where before we we have we have two minutes I wonder if there is some quick question anyone um otherwise I have one last quick question and then then we can repeat up oh there is there's a question from Caleb foral Ms I'm not sure if that's that's quick but if you can do it in two minutes yeah yeah no this is a big question that I think about probably too much um yeah I don't know the I guess the short answer is kind of what I was pointing at a second ago where it's like you can imagine that like um we move away from even where we are today where it's like things are somewhat centralized where it's like there are like to a large extent you do need to like trust some other sort of big player whether that's a company or like something else uh to actually know the state of the chain whereas like what I would like to see is like again this very sort of like cypherpunk Vision where it's like via VIA the application of cryptography or different things like this you can imagine that like I have the same guarantees as these