Protocol Berg v2: Geth AMA

the team lead of GU. Uh, do I need to say anything else or Okay. Um, hello. Is my mic on? No.

Hello. Is this mic on? Yeah. Okay. Hey, I'm I'm Matt.

I go by Light Client. I am a member of the GU team for the past few years. And I think that's all I'll say. Um, I'm Guom, also a guest member of the guest team for the past few years. Oh yeah.

Oh yeah, maybe. Who knows? Um, I'm I'm Marius. I'm very surprising also a member of the guest team, otherwise I wouldn't be sitting up here. Um, and I've been also with the team for I think five years now.

Hey, I'm Jared. Uh, I've been with the team since 2021. Let's get some, uh, let's get some questions going. Uh, okay. So, one of the questions that has been voted the most has been, if you couldn't work on Ethereum, what would you work on?

Yeah. I don't know. We we don't really know honestly. I think is it possible to put the questions on the screen here? Yeah, it would be good.

But I mean we also have to just answer the current question. We cannot it doesn't help us to check the other questions because we have to answer. Oh, you can start. I I I I actually really like fuzzing uh and and testing. So um I always had this idea of doing like a like a company that does fuzzing for for other projects.

Uh so that is something that I would probably uh I would be building blimps. [Laughter] Not even a joke. I would probably work on Bitcoin. Yeah. For myself I I don't really know.

I I think I would just have a job at a startup or something if it wasn't this. Yeah. Likewise, I I don't really know, but I definitely want to stay in space. Yeah, I think all of us would probably stay. We would prefer to we would prefer not to.

Okay. Okay. Um, some of the next questions here in this list, um, what are your thoughts on Wrath in Oh, come on. Like, we we're here to answer your questions, so we should just like answer your questions. No, no, no.

It's okay. I mean, this Yeah, I mean, that's this one's not the top world. Oh, so uh it is his job to how do you like working with Tomage? Oh, that is a good one. Yeah.

No, I mean so seriously it is we are I mean we have we have uh you know we've we are we've been getting accustomed to the Tomas experience for the last uh couple weeks and um it's an interesting thing like he is bringing a very new dynamic to the Ethereum Foundation I have to say and is good and bad. So I mean it's not I mean I don't want to say it's bad I mean he's my boss. No I mean ah damn it I'm too honest. So it's no I mean it is it is it is a it is a big shift I have to say like he is bringing this energy where like you know we are laser focused on you know like the immediate goals and also like the very big goals and I think this this focus was maybe not there before as much you know and so he's he's honestly bringing this and um at the same time it is it's that is I guess the good part And like also the bad part is maybe that like with all the recent reorganization, it's hard to say like how much of this is like can be attributed to him but specifically but it's like also just all of these changes we are currently still kind of processing all of this and it's it's very um yeah it's not all like good feelings. Yeah.

Yeah, I mean maybe just to add on, I think it's a little too early to know what really the outcome of this transition is going to be. Like a lot of things have happened in the past few months and I personally am excited about change. I think that like for a while we, you know, for the past couple years it felt like we were sort of stuck maybe in some sort of rut and now I appreciate that there is some vision and some ambition in what we're going to be doing with the Ethereum protocol like vision ambition that wasn't it wasn't as you know quite there or communicated as clearly you know to the community and I think like he is bringing a lot of that and he's trying to like mechanize that across the Ethereum Foundation again like this you know he's been officially with the Ethereum Foundation for two and a half months It's very early in his tenure. We'll have to see what the next like six months hold in the year after that to really be able to answer how we like working for Tomas. I think we also like we worked with him before in a different context, right?

Like uh he was a he was a core dev, he was a client dev before uh he like really built up Nevermind and um I actually really liked working with him uh back then. So, um I haven't had much contact with him uh since he joined uh DEF, but um I really liked um uh back in the day working with him on the on the engineering problems and I hope that um we will uh we will have a maybe more re more engineering focused organization going forward. Yeah. Um he's um he's questioning a lot of things which is a good thing. It's also a fairly uncomfortable one.

But uh yeah, we like like other said, we are going to have to wait a little bit to figure out if uh if this is what we need or what whatever happens. We'll see. I guess not really a comment on specifically, but I think I'm excited to see what this new uh laseredeyed focus on accelerating scaling brings. And I think that if we can push that along more quickly, then his his entry into the EF will be a good thing. Next question.

Okay. And remember when you use a mic that is Yes. horizontally. Uh, next most upvoted question as we're continuing this. Um, do you think there would be benefit to spinning out of the EF and become a grant-seeking organization?

I guess so I can answer that one. Um, um, we have discussed the spin out question many times uh over the years, but we always end in the conclusion that specifically for GU it makes no sense. uh GU itself will never be profitable in any way like it is not it is more like this type of project where we in order to be fully committed to just the Ethereum protocol we have to be supported by an organization and um at the same time we are super deeply integrated into the Ethereum foundation its structures and also the people there so I think I would say a lot of the value of GU within the EF is just this integration, us like knowing everyone, having all these ties, having all these projects within the EF with other teams and so on. And uh if we would leave the EF behind and we would just focus on GU, it would be a big change. But I think it would also we would a lot of the value that we actually bring to the world would be lost by this transition.

So that's why we think this spin out makes no sense. Yeah, as as Felix said, we have debated this over and over over the years. Um there might be some benefits. We might I don't know be more independent, but um the drawbacks definitely outweigh for us outweigh the the benefits that we would have from from spinning out. So um and uh I think the the the other thing that is like we're we're not m we're not only maintainainaining guest right we're also we we feel very responsible for the protocol itself we feel very responsible for like the testing the standardization of of um of processes um uh like the the APIs and all of these things would be things that would need to go if we were if we were becoming a for-profit company and um I don't think that would be to the benefit of Ethereum and our biggest goal is to work on Ethereum and uh and to maintain cool software and uh I think this this structure that we have right now is the best way that we can achieve our goals.

So I don't see a point in changing that. Anyone else have more then? Yeah, we need to kind of go through the questions a bit quicker, I think. I'll just switch mics. Okay, so let's just go to the next next.

Yeah, next question. Yeah. Uh, what are your thoughts on ref? They added, "Why do you prefer it to GU?" But doesn't it make sense?

So, just what what are your thoughts on WH? I think it's a trick question. Yeah, that's why they're going to get an answer. Honest answer. A real question.

Um, yeah. No, I I I love it. It's uh it's a really nice uh heater for my uh like I run it on my on my knuck and it really hits my uh heats my my office very nice and also the the fan spins really really uh quick. So I really uh I really enjoy like the background music. Um and also it's a great test for my uh internet connection because every time I compile it, it downloads the entire internet and uh um yeah.

No. Do you ever do you actually run? Yeah. So in in my backup in my in my backup PC backup. Yeah.

I mean Reth is definitely like a a newcomer to the execution client space and I mean we always find it quite exciting to work with uh newer clients because they do bring some new ideas. Uh in the case of RET um yeah I mean they have a very different focus. Again it's a very performanceoriented focus as we are used to from the Rust programmers. they love to optimize and um it's um it's great. No, I mean we we very much enjoy talking to them and um yeah, we've been sharing knowledge and uh there there are no negative feelings towards GU uh towards ref from the guest side.

I don't really know how they see us. I guess they would answer this differently. I mean I I think now they they see us a bit more with more animosity, I would say. Oh yeah. Um, but I mean after what I said, but uh but it's uh No, it's it's a very strong team for sure.

Anybody else had something? We're going to go next question. I think it one aspect of Earth that's really neat is their close connection with Foundry and a lot of the dev tooling aspects. I think that's one thing that fundamentally a lot of the client teams have always been kind of separate from and the separation just grows with time. Like at one point, Go Ethereum was much closer to what was happening with developer tooling and now there's just so much going on with their protocol.

There's only so many hours in the day for us to see all those things. But because they have that very close proximity, they are able to bring a lot of the feedback that exists from developers to all core devs. And so just that aspect of rest has been awesome. Yeah. Like when we have the the like crossclient discussions, they always come with these suggestions saying, "Yeah, we know that there are these users uh doing this type of operation for example very often or so like and then this is something that we might not necessarily always know and they have this because they have a lot of users of their tools reporting certain things and so on.

So it benefits us actually. Um I'm going to combine two questions. There's something about top three engineering problems and then also about technical debt. So maybe could you guys go on those? Like what kind of top engineering problems are there still left with GU especially in relation to technical debt?

H it's a weird it's kind of weird to combine these I would say. So the but we can we can also start with one of we can just riff on this general area I guess. Just trying to go fast as we have. Yeah. Yeah.

Yeah. Okay. So I mean technical dep what are you even talking about? Um I think that's another word for the Ethereum protocol. Yeah.

Yeah. We just heard it from from from from Justin Drake. Basically all of GU is technical there. Like there's no single part of it that you know it is not um actually I think he meant all of Ethereum. Yeah.

Yeah. Okay. But like when it concerns us basically means get no I mean there um what I guess when it comes to technical depth and one thing to realize is go Ethereum has a kind of big API surface and nowadays because uh since the beginning of this this year we have had a bit of a pivot and we now care a lot also about uh the uh layer 2 forks of GU for example and this increases our API surface even more because effectively all of the core of G is now kind of like has to be maintained with the eyes of like someone else having to work around the changes that we make and so on. So um we have a lot of difficulty like trying to keep certain parts stable and not changing them so that other people don't other people's projects are not broken and I would say that um if we are planning to make certain changes to very core APIs and um it's a challenge because we have to maintain the backwards compatibility we have to really think through like what is the longevity of this interface and things like that and yeah so these there are some parts that we maintain where we think if we could do it again, we do it totally different, but everyone depends on it now, so we will not change it. Um, in general, we've been pretty good at just deleting functionality that isn't used anymore, I would say.

And I mean, some of this stuff is also like we're kind of on the fence about. So, for example, GU still has a library that can talk to hardware wallets, for example, but obviously like nobody's going to be using GU for their like hardware wallet transacting. I mean there are some projects using this as a go library but it's like one of these things where we've been thinking this whole like accounts account management direction is something we would probably just delete if we could but at the same time it doesn't cost us too much to have it so we're just like yeah you guys also want to chime in on this or I could speak on the top three engineering problems with gas uh like I also think like we don't have that much tech that in G. Um the only thing that we still not have is the small archive node but that's coming very soon. Um as we probably said like two years ago already but like it's really coming very soon probably.

Yeah the legacy archive node we kind of consider it to be a bit like this this whole like we have two big modes of storing your state. we have the pathbased uh state and the hashbased state and the hashbased stuff is like we want to remove that like but it's it is very hard to remove there are even like this week we are learn we learned about three more projects that like heavily depend on this mode of storage for example so yeah like some stuff would be great if we could remove it but it's dependent on um yeah one thing I think rest by the way does very well but we for us is perfectable it's the modularity Um like G is just a big chunk and extracting stuff is really hard. Um so that would be one thing and uh technical depth the Genesis code. I would totally get rid of that. Um the the other question was uh was top three things that we that we want to uh we want to improve or something.

Yeah. Yeah. um uh top three things that are problems in GU. I think the biggest problem is state like we don't have a like state is always going to be our main bottleneck. Um and we don't have um just the Ethereum protocol itself has no real answer to what are we doing with the state.

Um yes for most of the users it will be uh it will be solved by state statelessness but there will be there need to be these users that just have the complete state and have uh have the access to the state and can provide proofs over the state um in order for for light clients to work uh properly. Um so I think that is something that is um more in the like medium to long-term road map um where we see we need to invest much more resources to um to understand the problem to see if the if if we can find good solutions to um uh to to manage the the ever growing state of Ethereum. Anybody still more to add? No, I think we should go to the next. So, now I'll give you a choice.

Uh, we can do the Peter thing like it's okay. Come on. It's okay. Okay. Do you share Can you Can you also because this is an ambiguous, can you do you share Peter's frustrations?

And can you also explain what you think? Yeah. Uh, his frustrations might be. Yeah. Um, we I guess this is one of the main reasons people came here to hear about the whole Peter situation.

So, um, No. Yeah, it's fine. I mean, this is good. Um, Peter, um, uh, as many of you guys know, uh, Peter went on a sbatical, um, the end of last year and then up until basically now. It's like he hasn't been back.

Um, it's is is a complicated relationship, I have to say. So um Peter was always frustrated with EF and the specifically was frustrated with processes at the EF and management at the EF and yes we have at times felt equally like angry about the EF's processes like it's not been an easy ride. How he deals with that is a bit of a different thing. So he's famously very active on Twitter and can be very toxic even and this toxicity has led to the current situation where basically it's not really possible for the EF to work with Peter because of the way he has been behaving socially. um we value his technical contributions and we would love for him to be back at the EF contributing in a technical way but it is very difficult for us to deal with his behavior I would say and this um some of the latest frustrations that he has been sharing I personally would like to say that I have researched these claims that he has been making the last couple days over and over.

So, I do not share the sentiment of these claims at all. Um, the EF does not want to remove the guest team. The EF does not like somehow think we are worthless or whatever. And the EF has also not been trying to create a secret second GU team that could somehow internally replace us. like this these things have not transpired in the way that he has been portraying.

I hope this answers your questions. So like me personally, I do share some of uh some of Peter's frustrations about the processes as they have been in the past. um like there was like everyone knows there was the the u the idea of substraction and teams were getting uh were getting removed from the EF. Um and uh there was there were plans for for GU to be also spun out of the EF and we had a lot of discussions about it and in the end both we and the leadership team at the at the Ethereum Foundation um uh came to the conclusion that it doesn't make sense at this point uh for the guest team to leave the EF and um yeah I was like at at the time we were kind of I was I was also kind kind of like myiffed about it. But I don't think we should dwell on these issues that are like for me that feel like they're in the past now.

Uh we have a new um we have a new management team. Um we have a we have more direction. We get more uh we hope that in the future the EF will provide more direct feedback. Um and um I don't know I yeah my answer to the question I have shared his uh some of his frustrations in the past but um as as Felix already said I don't think um uh that Twitter is the right um the right place to uh to make room of those frustrations. And to be clear, um, Peter, you know, has, uh, has held the the team together for the longest time.

He's been the one dealing with, uh, with, let's say, leadership trying to find its way. And, um, yeah, clearly we do share uh, at least some of Peter's frustrations, like the exact accusation. Uh, well, okay, for example, I am not in a place to to verify all of them. Um, yeah, we we just appreciate Peter. I think that should be the the map.

We have a few minutes left. One one or two minutes. Yeah. Yeah, we can. Let's do uh would you still work on GU if your pay was cut 50%.

Yeah, we have been I have to say this question is not the first time we received this question. So I think one minute left. I think I think something that you that you need to realize when working with Tomas is he states a lot of very provocative questions like what if we fired this team? What if we like doubled the salaries of this team? And um so I think he's he's like questioning a bunch of the he's trying to question like the the common wisdom of the EF.

And um so I don't really see the this question of like what what uh what would happen to the guest team if if the salary was cut by 50%. As um as a real thing that he wants to do. It's just a hypothetical that he throws out there. Um and he expects to be pushed back on this point. So um yeah to that question I would say yes at optimism.

Yeah, that's I think that's a that's a thing for us like we we can probably go into a lot of companies in the ecosystem and probably get much more but uh we feel that Ethereum is our mission and Gath is um for for now the best uh the best uh the best place to um to to capitalize on that mission and to work on that uh mission. Um, but I think we all share if we feel not appreciated um by the by by the Ethereum Foundation um then maybe it would be time to move on. But I think right now that is not the case and so we're staying and um we love to work on GU and we also love that GU is the most client in the ecosystem. We especially love that part. Uh thank you guys.

Please, let's give a round of applause to them. Thank you.