EPF6 Office Hours | Giulio

Very well. Okay. Awesome. Yeah. Welcome everyone.

Um, thanks for being here again everybody and thanks for Julio for joining us today. Um, are you uh our next guest for the office hours AMA. So we have these regular office hours to um uh have a guest from some of the client teams to uh learn more about their work about their background to understand what is it like to be a core developer and um the audience the fellows uh have a chance to ask any questions. If you have any technical questions uh about consoles execution on go I believe that uh Julia is the right person to ask. Um but before we jump into that I would like to learn more about your background about where you're coming from what was your journey in the core development.

So uh why don't we start with uh with a bit of that if you can tell me uh Julio more about how you get started working on Ethereum. What was the what was the journey for you to get into the core development? I know you started early so I'm curious about that.

Yeah. So all right. So first of all um so I I actually so actually my journey is a bit weird. It's probably more unusual than other people because I started cur development at 17 years old. Right now I'm uh 22 almost 23 and it's and and and it's a bit weird because it's basically because when people ask me where you know where do I come from it's just Ethereum right?

Like I'm basically Ethereum native in every single sense of the word I would have word. Basically my career as a software engineer start in Ethereum and it's also continuing um in Ethereum right now. So it's a very boring answer but yeah I basically I'm I'm I'm just Ethereum that's it. And so the journey actually started with me when I was 17 years old starting to chat with Alexia Kunov which was the when Argon was called Turbo Gaff at the time there was a period of time where we didn't really choose a name and Argon was just kind of a better archive node for Gaff and so we just went with Turbog and at that at that time I was uh we um I I I decided to reach out to Alex Kunov and ask him actually to join the project just called text him whether I could just join the project and and uh he gave me some task and then I just I just joined eventually because apparently I was I was good at it and u and yeah so that's essentially that's essentially what happened he was actually Alexi was actually at the time uh looking for an apprentice of some sort and I just kind of spawn into the picture I just took it that I just took as an opportunity I actually was aware of Ethereum since I was 15 years old. I did some ser before then but nothing really serious.

like contribute to some project like Rideen Network which now is basically dad basically. Yeah. But nothing too crazy. No, but nothing too tra nothing too crazy before arrogant. And yeah, after just chatting with him for a little bit, we he gave me some tasks and I then I just got started and then slowly I became I got better at it and now I'm essentially the city of the company essentially.

That's essentially what happened. Um and yeah, that's that's my past, my background. Um, yeah, it's a bit boring, but it is what it is. I have a short life. Yeah, I don't think it's boring at all.

I think it's very interesting. Thank you so much, Julio. Uh, yeah, it's really cool. Yeah, so uh Ethereum native since 17. That's really awesome.

I've uh I've been looking into Ethereum since I was like 16, 17, but uh mostly focused on Bitcoin before. So, so it's um I think takes a lot of courage and talent to to start um so soon and

and uh yeah so if you can tell us more about about that like what was it like starting what were the first things you uh you worked on what issues have you been facing? So when I got started I think in Ethereum like when I entered Ethereum it was I was not 17 actually I was 15 right I I became a c developer at 17 but before but I actually entering Ethereum at 15 and actually I I I didn't like Bitcoin to be fair I never entered Bitcoin I I didn't I I still don't really like the premise of Bitcoin to be honest I thought Ethereum was cool because essentially everyone can build their own application and uh just get done immediately. You can basically handle real life money without any intermediary. It doesn't matter your age, doesn't matter your gender, doesn't make nothing, nobody cares. You can just do it.

And I think that's what was great about it. Bitcoin didn't really intrigue me. Uh probably kind of unpopular thing, but I never found anything really. I I don't like Bitcoin basically. I just don't like it.

and yeah and uh and yeah so so that's what appealed me to get into Ethereum it was there was smart contract development and I also yeah and then eventually so at the beginning I started working for Michelon's project in open source space like one of them I already mentioned it's red network for people who don't know what red network is which I think it's most of you because honestly it's dead so nobody probably knows about it anyways anymore. Um, it was a m it promised to be a a micro a micro a payment service for Ethereum. It basically use something called payment channels which was basically uh the early version of layer 2. Basically, you you do the transaction of chain and then you verify and then you batch verify them on chain later. It was this is basically the gist of it, right?

And it basically is the earliest basically I think it's it could be considered the earliest version of an L2 that has ever existed. And I remember that my first PR there was basically patching a security vulnerability which was kind of interesting. Um basically bad balance. It was still in test at the I'm choosing to mainet but they essentially add like they essentially didn't check some part of the balance when doing transactions and it could and people could double spend basically on the payment channel and so I so yeah that's the first issue I actually ever fixed in in uh in any in any blockchain project which I think is kind of interesting and and then I I did some contribution to Kyber network which is now a dead project and and uh and and other things which actually I don't remember on top of my head. Oh, I also yeah I also played a little bit with some of Barry White out project about zero knowledge but it was very minor stuff but I can yeah but these are I think my my early highlights basically.

No, that's awesome. And I think it's it's really cool because uh I I didn't know that you contribute to Raiden and I was a big fan of Raiden like again mentioning I coming a bit from Bitcoin. I was a big fan of Bitcoin lightning the payment channels on there. So the Raiden just immediately made sense to me. I like the idea of payment channels which are

they sent me between two parties.

Yeah. The the nice thing is they actually sent me merch when I fixed the issue because they realized I was 15 and they just sent me merch. Um I I have like three Raiden hats that are oversized now.

Yeah, that's awesome. Yeah, that's really cool. Yeah, I have a Ren t-shirt somewhere. That's uh that's interesting piece of history. But uh let's move more to the future or to the current day.

Uh what are you working on these days? I know that you've been uh contributing to um a lot of like um uh prototyping efforts like you've been working on ver trees before in argon and uh on the latest features. So what are you yeah what are your challenges today? What are you up to?

So right okay so I I think right now um I'm mostly up to to do like so I do so uh I think I I should separate what I do for work to what I do as a hobby. As a hobby, I just write EIPs and I make them go into hard forks. That's what I do as a hobby. It's not really part of my official work, but I just do it because it's fun to get uh to get to contribute to Ethereum protocol as a whole. If you look at the Fuzakar fork right now, I have a I think I have four EIPs in there that you can I I don't know.

I don't even remember the the numbers in my head, but I have four EIPs in it, for example. And that's one thing I work on. I just work on specking them and uh and into coming up with all with with the ideas behind them. They're mostly right now on L1 scaling. Uh well, there are security there is a security EIP's um direction at L1 scaling and uh per and when and and uh and in Argon specifically right now I'm uh I'm I'm working on well I'm the CTO so I I work on many things.

basically work out everything. But but the the but the coolest thing have worked in the past few months was our proving our was our proving data scheme for archive nodes which we didn't have and no other nodes add and we just essentially are about and we're about to roll it out right now. Essentially, it allows uh big info providers to provide Ethereum proofs to be for basically any every single block range, which is great for uh for witness generation in ZKVMs. It's great for trustless bridging. It's great for a lot of things that that requires um Merkel Patricia proving.

So everything everything that claims to to prove Ethereum um it's is made easier by this um by by by by this by this project I've been running and and I actually finished it very uh very recently uh two days ago actually I think we are it went into testing by from from some big infra provider and they're testing it for me and we're going to see if it's going to work. So that's what I've been working technically. Then I'm also manager of projects. So that's a bit more boring stuff. Uh it's just keep track of people.

Uh and yeah, that that's pretty much it what I've been doing recently.

Okay. Yeah, that's really cool. That's very interesting also from uh coming from like the early contributor to CTO to managing people. That's interesting experience. Awesome.

Yeah. Thanks a lot, Julio. Uh I think it's very inspirational, but I hope that folks have some more specific questions, something that they can ask. We get some people who also contribute to Argon as a part of EPF. So folks, if you have any questions, feel free to raise your hand.

Um don't let me use all the space.

Someone already put a question on GitHub by the way. Maybe we should ask just by answering.

Uh yeah. Yeah.

Okay. So

yeah, yeah, we can start with that. It's great. Yeah.

So just so I'm just going to read it out loud or if you guys want. Uh so the the question is I recently saw nethermind gaff and class lamba teams published articles on benchmark testing but the results display are different. Should there be unified standards? So um so probably so I so it's a bit vague. I the question is a bit vague here.

Um I don't know what should have a NF standard if what it's meant is the performance optimizations probably the answer is probably no because each client is built uh uh differently uh it they're designed differently and it's very hard to do the exact opin the exact optimization to the exact optimizations to the minuscule detail um in every single client at the same time probably that's not possible uh there are unified standard in the sense that the protocol specific ification is the same for everyone. But aside from there, probably the answer is uh no. And there probably never be because it's a huge coordination issue. And it's and also and also it's and also it's very hard to enforce, right? It's protocol rules are easy to enforce.

This stuff is out of the protocol. You you can't enforce it. And then the second question is I originally considered implementing uh EIP 7828. So it's block level access list I believe just by looking at the number in argon but when I look at source code I found that argon state management implementation is quite complex in addition to the basic database cach and also c and also uses right file optimization what is the trade-off here so actually block level access list should not touch any of that to be fair that's the first thing I'm going to say I don't think it touches that you should just listen to the to the touch state and that's it you shouldn't really you shouldn't really touch the the internals of the database. Um, Gaff also have op and other mind also have cing.

Uh, and the gaff well yeah the train so okay the the first comment I would say is that you probably shouldn't touch those things as part of bal in my bed. That's the first thing I'm going to say. The second thing which is the answer to the question what is the trade-off here? Basically um the tradeoff so the so the trade-off is uh well there is no trade-off there optimization so they make the thing the the the the node better they make some low-level API more complex as you mentioned I don't think you should touch them as part of 7 7 928 I don't I'm not sure why you look into that but but uh but yeah so the k issues are also present in G and other mind the database separation from static files and and um and the state is actually also present in G right now uh but it's a lesser but I think it's probably lesser known we probably abuse it more and I mean the compression it's just an optimization so the trade-off is yeah it's some low-level API is more complex but overall you shouldn't touch them as part of that's the question and it makes everything better overall apart from that.

Very well. Thank you so much for answering both of those and uh yeah, thanks for the question, Gaba. Um it's great that we had the the one in the disc in the in the GitHub. Um yeah, uh thanks for technical answer. I hope you guys have some more questions.

Please raise your hand. Let's ask something live as well. um you have a chance to chat with Julio in uh in in person live. So please use the opportunity. Yeah, Lee.

Yeah. Can you hear me?

Yes.

Yeah. Cool. Thanks a lot. Uh your your experience is amazing. So I was pretty surprised.

And also I'm curious about uh when you started to dive into Ethereum it was already a huge project. So uh where did you started like to learn like to accumulate all this knowledge and also like uh code B uh the understanding of the code base of the Ethereum. I mean now when we look back to the Ethereum codebase now it looks like it's pretty huge and honestly speaking I can only understand the part I'm working on. So if you are talking about other parts it's really hard to say like I have any understanding for them. So I actually started so I I I don't remember if it was immediately or one or a bit in my in my blockchain journey when I started at 15 years old but uh the first time I looked Ethereum specification I referred to something called the yellow paper.

I don't know if people still talk about it because it's a bad documentation as far as I know because now we we have also the consensus layer and it will be huge to describe it all in one and in one single place at least. So I started from the yellow paper. I read the yellow paper. Um I it's actually I think people I think the yellow paper is is a very nice way to start learning about Ethereum. uh I think people kind of demonize it a little bit because they think it's complex because it's full of mathematical notation but if you get used to it it's actually quite nice.

So that's actually how I started um to look into the internals. Then the second thing I did while I was looking at yellow paper is also I was going over gaff a little bit. Uh I was really at the beginning just curious to know how everything worked because you know yeah I mean I can write smart contracts but smart contracts are very easy right they're just like JavaScript like language entry points easy stuff but you know you go into details and then you look at how the EVM works how the encoding works how transactions are are broadcast are broadcasted and yeah a funny part is actually I never looked into Bitcoin code I actually don't know anything about Bitcoin technically but I know quite a bit about Ethereum. So that's how actually started learning. I just looked at the yellow paper and while I was looking at the codebase and you know some things are just onetoone with the yellow paper.

So you kind of get to understand at a surface level what does what very quickly because a lot of is referenced back to the yellow paper at the time at least now it's it's a that documentation I think

yeah thanks I mean I also started with yellow paper and uh also

it's not yet well

yeah we also like uh you know like how to see how the compeller works how the uh stake virtual machine works that's actually the first time I understand how the ST based virtual machine works.

Yeah, I know. I know. It's great.

Yeah, that's also amazing. But I didn't really get into the guests. That's maybe the difference. Uh and uh yeah, we actually discussed the last week or the week before we wondering like there is actually like a yellow paper for the current version of the Ethereum uh technique. But I mean, I didn't find that.

I I know it's it's I think it's that I think I think they give give some the people that maintain it gave up on it because after London because after DP159 art for because they had the consensus layer then they added other things and if you were to actually formally specify it it would be bad probably it would be probably very long and and also I yellow paper is people are scared of it for some reason so I think it's the best piece of documentation for criterium that has ever that has ever existed. But most people don't think so actually. So

yeah,

it is what it is.

I enjoy I enjoy it, but also I know like people some people really

kind of enjoy it. Yeah.

Yeah, I think it's just taste most people are don't like it. I don't know why. Anyway,

yeah, thanks a lot. Yeah, thank you so much. Um, I'm honestly surprised as well like I I think it's hard to parse the yellow paper if you don't have the mathematical training if you're not used to it and uh respect for like diving into Ethereum like that. Like I think the Python specs are for programmers at least are much more readable. So it's yeah it's interesting approach.

I respect that. Uh Josh has a question. I I guess I'm curious more about your sort of early life education and like how you got into programming generally as as uh such a youngster and and like where that motivation came from as a as a kid.

Yeah. So actually yeah this is interesting. So, I think it actually I actually wanted I I I knew I wanted to to do computer stuff when I was like u I think four years old because I was playing I remember I was playing with I remember that my my my family uh gifted me a Nintendo DS. I I don't know if Yeah, you probably know what a Nintendo DS is, but it's And I remember that when I was four, I couldn't really get my hand around it because it was basically magic box doing things like it was like how how do you display things on a screen like that? Uh it's kind of crazy if you think about it.

Like from at least from a four years old mind, it's it should to me it was kind of crazy, right? uh everything seemed to just work like bowl like ball is here and and it falls and then all of a sudden you have a magic box that has colors and stuff and so I always thought like wow this is sorcery I want to learn about it right and I and and then when I then I remember when I was uh when I was around um I think 13 years old 14 years old I actually can't quite remember I think 13 years old I started looking up tutorials how to program in C that's my first programming language which is probably it was probably too hardcore. I probably should have started with something easier, but there was like this very good. So, first of all, I won't say I'm Italian and so I was looking up at this like Italian YouTuber. I was lesser known one and I and I just was looking through all of his C tutorials.

Then after a while I decided to I decided to test myself with um I with some uh with some exams from the University of Rome um which we're first year exams. So imagine like implementing link at least in C or stuff like that, right? It's very not very complex stuff. And I real I think I realized I had a knack for it after I was able to complete basically all of the question on some of those exams and I just started going down from there. I then I I did some some then I I I I'm basically 100% selftaught.

I don't think I had my family gave me some tutors at some point because they kind of realized that I was particularly good at it but I think I'm mostly self-taught. I used the tutors mostly to answer question and not to answer I I you know those questions that they come up but you can at the time there was no chip right so you you needed to have some ref and the internet was not actually that developed either. So if you had a question, it was really hard to get an answer for it on the internet. And so that's essentially uh what the tutors did and but most of the things I I think I'm I'm pretty self-taught even going forward. I think after my my in learning C phase uh I just uh I just learned all the other languages by myself without any help.

I think C is the best language to learn to start probably with action and second thought because it teaches because if you can do C you can do everything else. Uh and then I just went over from there.

Awesome. Thanks.

Yeah. Thank you so much. Um awesome. Any more questions? Anyone free to raise your hand?

Um yeah.

Hi Julio. amongst Italian. So, Julio,

I think I met you.

Yes, exactly. We met at the interrupt. Yes, nice to see you again.

Nice to see you.

So, my question would be more not on the code base. So, like stepping back from the code and more the technical side. What do you think on the about like the actual products that are built currently? So, for example, last summer we had the NFT saga. This summer we had the meme coin.

So what do you think about the products that are built on the top of Ethereum?

H okay that's an interesting question. Um so I so to be completely fair um so I I want just to start saying that this is really not part of my expertise but I I I might as well just answer but I just give an answer. I think that uh I I think that first of all I I'm of the opinion that uh Ethereum should build the useful things. That's I think my my first my mantra when I came into the chain. I think that the best things built on Ethereum so far is DeFi and it's not even close.

I think NFT is NFTs memecoins are generally not bad use cases and I think this is kind of an unpopular opinion. I think I I I would even go as far as saying that that stuff like um the stuff like um you know this you know that accessible or permissionless gambling is also a nice use case. I know that sounds bad saying it like that but you know there are people who wants to do it so why not let them right that's my opinion on that and uh but I think that the epidm of Ethereum so far has been defy I I'm it's I think the number one use case and I think you know if you look at hav right now it will be the top 30th bank I think that's great um it's it's it's just great I think I think I'm I'm I'm a really big fan of the fi u I'm also I'm somewhat of a fan of everything else. Everything that people wants I'm a fan of. Uh I think that right now I think that I think that right now these products are maturing and I think and and and I think and my honest opinion is that the fight is going to be is going to is going to make a great return soon probably because that's I think uh now there is like this this narrative about treasury companies, Wall Street and I think that defy is uh is going to be pretty much the the hot spot of this of this of this new narrative, I think.

Maybe not. Who knows?

Amazing. Thank you.

Thank you. Yeah, it's amazing to see you bullish on DeFi. That's uh that's really cool coming from a cord.

Nice. Um what's up has a question.

Uh hey, can you hear me?

Uh lightly. Maybe you should I can hear you. I can hear you. I can hear you. It's just a bit far.

I hear you a bit far, but you can.

Yeah, it's fine.

So, we were talking about that D5 was the one of the best inventions in Ethereum. Uh, so let's talk from an economic perspective. Uh, we can easily say DeFi uh firms make money and security firms make money and even builders, everyone makes money. So, uh what are your thoughts about ZK? How do ZKVM teams or uh yeah these teams make money?

So my okay my thoughts about ZK right so so I this is actually kind of a loaded question. So uh so this is actually yeah this is a very loaded question. There is a lot of things to unpack here but just to let me just think about it for how can I So, I'm bullish on ZK. Okay. I'm I'm completely bullish on ZK.

I think uh I think it's the right direction if it works. Um I think Yeah, I I also do think that uh I also I don't I don't know if I Yeah, overall I'm just bullish on ZK. Yeah, it's just great, you know. It's just great. I don't um I uh I I I I you know if it works I think it's it's great if it works right.

I think that uh I think the only thing I might say that might be probably unpopular is that the timeline that the timelines right now are not what I'd like them to be. Um but that's pretty much it. I think I think if we can find if our compromise is with ZK we can uh we can do some serious scaling. I think Zik is great. As simple as that.

I I just hope it could come. Uh I just am not really happy that it's coming in potentially like 10 years because I'm probably not going to because I mean in 10 years I'm going to be 33. So that's the only thing I'm not really a fan of. But that's that's great. I'm bullish.

We both have same views on Z.

Sorry, what?

Uh, it's nice to say we both have same views about ZK.

Yeah. Awesome, ZK. Um, any more questions, folks? We have uh plenty of time. Please do feel free to ask.

one

curious uh what you feel like the hardest part about your job is especially as you know kind of stepping into this CTO role and being a people manager.

Yeah, I kind of thought that might have been the answer.

Yeah, I mean to no I mean know there are different definition of difficult, right? I think for me it's talking to people uh because I'm also young so I think that so you know it's I know it's not really uh an optimal dynamic to have uh I'm aware of it uh because people I work with sometimes are twice my age and I am aware that you know being young I don't really interface myself with people that are that are older than me so I have to be careful careful about what I say. So, it's a bit more, you know, it's not hard in the sense that I don't really need to put like time in it, but I need to be careful. That's what I'm going to say, right? It's uh it's something I I probably don't want to mess up.

Then uh the hardest thing I I work on well if you if you want to go technically probably uh the proving scheme that I recently finished doing is probably the hardest project the most complex project I've ever worked on so far in my life. So that's it. Yeah, that's like yeah that's I think generally my question the the gist of it.

Cool. Yeah. And I guess to sort of follow up, uh, I don't really see you participating in the live ACD process very much. Is that

what would you say?

Sometimes, but I do sometimes. I don't usually my talking is internally to my team. Um, I do sometimes uh because I have because I'm championing four APs. Uh the thing about the ECD is that um is that I guess I will attend when uh when I when uh when because usually ECDs are either like testing update updates which it's not really my responsibility um or it is not my direct at least not directly my I'm indirectly responsible but it's not my direct responsibility or it's uh or or in some cases it's EAP decisions and I usually only attend the EAP decisions because that's where I usually have where I usually have stake and where I actually have a direct responsibility to. So that's usually my talking is internal.

It's not and it's not really external with other teams sometimes but makes sense.

Yeah, thanks. Yeah, please uh raise your hand, guys. And Shane has a question. Yeah.

Yeah. Just curious um since you guys are both uh goling based like GE curious what are some of the main like optimizations you guys have built uh comparing to them since you guys are you know newer feel like you probably uh made some optimizations based off of uh the original codebase. So just kind of curious how would you kind of compare the two? Well uh is uh an infra provider client. We we work with our PCs uh if I mostly so we actually have optimization for different use case than gaff.

At the beginning, I guess we built Aragon because GAF didn't want our changes essentially. uh they didn't want any com additional complexity and so they just basically told us no and so we just built our own client our own fork and the main difference is that we are RPCentric we are infra provider centric we work with RPCs and we are and we optimize on that axis which is something that I don't think any other client is actively doing um so that's I think it's a pretty big plus for Ethereum as a whole. I think it would be I think it would be less of a plus if we just did what everyone else is doing which is uh like which is just be focused on staking.

Okay, cool. So you guys found like a niche use case within like the world. Got it. And that's RPCs.

Good to hear. Right.

Oh yeah, I mean it's a great use case to it was argon was the only viable way to run an archive node uh

before maybe red is another another example but uh I had a server with like 12 terabytes of SSDs just to run get archive note before I switched to Aragon. So yeah that's the that's the main factor there. Um, yeah. Any more thoughts, questions, folks?

Um, I have a question about uh Okay, I go ahead.

So, you're going to go next, Mario. So, followup question about the ZK. What specifically do you think is going to be like uh the pivotal point of the ZK in the next two years, let's say? So, so why are you super interesting in ZK and bullish in on ZK? Well, I think the pivotal point, the most important thing to figure out is how to incentivize people to generate proofs because uh right now as far as I as far as I understand uh the proof generation is very expensive.

Um verifying the proof is not but making the proof is the um the the provers if you go you might have heard there that now we have sub we have subtop proving we can prove ethereum blocks in 10 seconds and if you go look at the specifics this the pro the setup is for generating these proofs is around uh basically it's it's it's I what I what I read on saxing website was $124,000 worth of comput worth of computing which is a lot. Um, so we need an incentive for that and we need to figure out uh how to incentivize people to prove. So I think that's the most important thing to do and I think it's and I think although it sounds a bit weird maybe a bit easy because I mean just give them if just make a mechanisms to give them if from the consensus right but it's not actually that easy because if you think about it um there are a lot of stakeholders here to consider and I think that you know it's very hard to push things into Ethereum as is if you if you also have differing stakeholders that have that have you know kind of an differing interest is a bit hard to po to to pinpoint a point of agreement between all of these stakeholders and I think that's what we need to figure out. I think comfortably we will get to sub4 seconds proving we are going to do it 100%. Uh so I'm not worried about that.

I'm more worried about uh how we are going to incentivize people generating proofs. That's what I'm mostly concerned about, but that's I think it's gonna get figured out.

Interesting. So, you think it's just a matter of time?

Yeah, it's a matter of time. Um, I the only thing I'm worried is mostly that there are different stakeholders here that might have differing opinions. I It's very similar to the issuance change. I don't know if any of you really So, there was a proposal to change the issuance of Ethereum to be more efficient, right? it wasn't the most you know it's not like our main priority anyways but there was this proposals made from some EF researchers and when it was published uh it there was like a huge outrage about it people were mad about it because you know homestakers would get less um so you know big staker big sting pools will get would get less because right now Ethereum is overpaying for security the analysis their analysis showed that and they wanted to reduce the issance because you you you don't want overpaying for security by orders of magnitude.

You want to pay what what is what is what is worth, right? Um so yeah, I feel like we're going to end up in that same position potentially because it's a change to because it is a change to issuance.

Interesting. Thank you. Um uh any more questions folks? I have a follow up on this because I'm curious how do you see the future with like this uh proving everything uh in ZK and having a high gas limit like um uh what is the future of archive nodes in this world like are we going to keep the same format are we able yeah

I I guarantee you that that if this becomes the case you will not run archive node at all because archive nodes don't have the shortcuts archive nodes in archive nodes you don't have a shortcut archive node is you keep everything right and and you and you need and you need perfect latency you need you need perfect everything and the trade-off space is very small you can't the zk uh world for full nodes is actually a trade-off you are trading off the you're trading off the ability to execute the block properly um in exchange for more scalability and you instead rely on verifying an external proof. That's all you're doing in ZK. That's a trade-off. In archive, this trade-off cannot exist because the archive node by definition um by definition is um is keeping everything. You can't prune it.

You can't you can't verify it. You can't you can't do zk on it. You can't do anything. Uh actually portal network wanted to try to push you know an archive no an archive node that is sharded in the network but that probably would have not worked because it's too slow. Um so and also portal network is now dead.

So I think that there is no I mean that's the only trade-off I can think of but that's not a good trade-off because if you if you try to do something like portal network wanted to do you would have probably got gotten an RPC answer say you want to get a balance for an account you would probably wait two seconds to get it and that's not acceptable for wallets like if you if you open metam mask it cannot take one minute to load your balance your your your wallet it's not acceptable of course um so archive nodes cannot take any trade-off and they will grow exponentially and you will not be able to run it 100%. This is not a question for me. I know it sounds very bad but that's what archive node will end up being in in our defense. Agon has a very scalable architecture for for big infrastructure providers but it will not help you. It will not save you from from the fate of of not running your own archive node at home.

Unfortunately, this is something that will happen.

Yeah, interesting. Thanks, Julio. Yeah, because the I mean, yeah, the history of the state growth is uh is just getting crazier and uh I was I was being a bit optimistic because I saw um some changes in the latest I don't know, it's four or five or something. Uh honestly, don't remember the number, but the archive actually got smaller. uh less.

Yeah, it's actually smaller than Yeah. archive notes today are smaller than uh five years ago. We went backwards. We went backwards. As history grows, archive not shrinks.

Yeah.

Yeah. Yeah. So, I saw you guys uh always have some cool research on this and uh I was hoping to see something which might be sustainable and uh Yeah. Uh either way. Yeah.

Thank you so much. Um we still have uh 15 minutes to go. folks please use the time to to learn something to get inspired to ask some questions from Julio. So Le go ahead.

Yeah, thanks. If you don't mind, I want to uh get into more details about technique part. Uh I'm curious about how like uh consensus client and uh the execution client uh handles blobs and uh per messages. Oh,

and also how to like uh perform samplings because uh to my understanding now pas now like uh uh allow or requires a different client uh consensus client to uh do samplings right and also they are divided into different subnets but it's still like

so you want to know more

so big disclaimer I know Pier does work. Uh I am not really sure about the exact specifics but I will I will try to I I but I I think I think I can answer your question. So to be honest um blob handling in Ethereum is kind of a Frankenstein monster even if you I think Justin Drake even said that so during the Berlin interup I that it's that the blob endling is just uh needs to be destroyed and changed completely. Uh basically what happens is this right someone make a blob. So after pas there will be a new transaction time I think where you are forwarding uh the cell um which where you're actually forwarding the cells of the blob instead of the blob itself.

So you're going to send a this transaction to the to the Y then the then the Y will then when the Y is building the block for the CL to broadcast it will essentially then send over these set these the cells to the CL when the when the when the when the CL is trying to get the execution payload and what's going to happen is essentially the CL then does the check itself the doesn't do almost any of the checking the CL and then it they get broadcasted, right? So, as you can see, it's a bit messy because uh the flow is not self-contained. It goes from the EL to the CL. It's a bit of a mess. Um but it is what it is.

And also, you know, the CL is what where the blobs actually live, not the EL, although that's where they're born. The blobs are born on the EL, but not on the CL. And yet, the CL um is the one that keeps them, archives them, etc., etc. So regarding different nodes actually it's very deep to me because um I because there are two different type of nodes in pieras for blobs there are normal nodes full nodes and super nodes right we care I care about super nodes because I am an infra provider friend right I I work with infrastructure providers and they need all the blobs so so I can tell you that that's kind of definitely made a little bit worse it's very hard with the with Pieras to to archive the blobs, but we were able to make it work.

Anyways, I'm a bit uh sad that now it's a bit messier and it's uh and it's a bit harder, but it is what it is. Everything should I mean, we shouldn't all I mean, it's good enough, right? Even if you scale it up to 70, so it's fine.

Wow. Thanks. Thanks. Uh I mean I understand it's a mess but after your explanation I understand why and where it's a mess.

It's a mess. Yes. Because it you you it's not self-contained essentially. And uh yeah.

Yeah. If you I think again I think Justin Drake also complained about it during the Berlin interrupt. I don't remember if it was him or Alex or or Alex. I don't want to misquote people. I think someone from the Y definitely said so.

Yeah. I mean like I I I started from the crypto side. It's very clear and elegant way. So I didn't think too much. I thought okay the protocol itself is self-contained.

It's like very nice secure.

Uh but uh then like once uh we like we touch the real implementation it looks like it totally not like the paper described. It's much more complex to all this stuff. Fortunately, some design decisions were taken uh that were probably not the best. And well, I mean, it's good enough. Okay, it's good enough.

You know, you cannot have the best. You can perfection. Perfection is the enemy of good. Good is the Yeah, perfection is the enemy of good, you know. So,

yeah, I agree. Thanks a lot.

Thanks a lot for the realistic insight, Julio. Yeah. Nice. Nice. Awesome.

Uh, do we have some more questions, folks? Please use the last few minutes to to learn something from Julio. Use the opportunity anyone. Yeah.

Yeah. Hi, Julio. Um I was thinking based on an answer that you gave for example how would you uh suggest that we pass like EIP which uh for example affect issuance and uh as a result they affect like big stakers for the good of the network. I mean how do you counter this pressure by that is coming from big stakers and that could harm their interest but overall as a as a network Ethereum would profit.

Um this is a very hard question. This is a a very hard question because uh staker control the networks you know they they run the validators um as a whole. So yeah um it's not a very easy question to answer. Uh it's very hard if stakers push back and they don't upgrade their nodes there is no upgrade happening. Uh I think uh I I I think unfort I think that we are going to have to have this discussion with stakers probably in the next few years.

It we we will have to and we will see what's going to happen. There will probably be some people that are going to get mad. Some people are going to take it to hacks to complain to spin up some drama. It's not going to be probably it's going to be the messy but it will have to happen right and I I lean to think that this the changes to the staking curve would have to happen eventually. So yeah.

Yeah. I think I it's a very hard question to be fair. It's a very hard question to be fair. I don't know really how to answer it. What I can say is that I hope we're going to change the issuance.

That's but I but I mean stickers they control the network. So ultimately so

yeah I see not so probably something like a compromise or yeah some kind of this kind of solution. Yeah thanks.

Well yeah maybe I don't know potentially I I I am I didn't really really thought up deeply about this. This is actually a kind of a good question. Thank you for asking it. I'm going to think about it later probably because I never thought about it this. Yeah.

Um, cool. Awesome. Hey, glad we can inspire each other, but it's a tough one. It's something that I think nobody in the ecosystem is really clear on. Um, okay.

Last five minutes. We have space for maybe one more one or two more questions, folks. So, please seize the opportunity. So uh I'm going to go for a less technical question and uh I see that you joined the space around the age of 17 and you're you were the uh CTO of Erican by the age of 22. Uh that's a short amount of time.

I'm wondering what do you think was the driving force towards such uh such a leap in uh in your in your effort in your like contribution? Um, so, um, I don't know if I should, um, okay, I I'm trying to find a good answer to this. So, ultimately, I just add good ideas, right? I don't want to get too much into detail this because it's a bit because these things are usually imply usually involve different people and they involve different you know when people take a decision like that there is there is usually a need to take a decision like that and uh you know it's probably more of a private reasons that they should probably not say out loud but the main reason is that I had good ideas um people had the impression that I had good ideas and that I knew what was needed for uh for Ethereum infra provider and that's it. I think that's the the reason I that that's the part of the reason I can give you I think I think if I have to give an advice is to not which is something I used to see of in the past because I came in very young and I happened to you know get a bit arrogant at some point in my life.

Don't become arrogant. That's all that's the only advice I can give you because, you know, yes, you might think you're better than some people, but it's but you should never be arrogant because, you know, you never know who's going to to to be your next uh to to to deal your hand next, right? That's the only advice I can advice I can give you.

That's beautiful advice. Thank you, Julia.

Yeah. Yeah, thank you so much Julio. Uh that would be my last question actually. I wanted to ask you about some advice for the fellows and I think this is a great one and um yeah.

Yeah. I mean I guess I guess I guess I can give you I mean being I mean I guess that don't be arrogant happens when because when I was 17, right? Like imagine being 17 years old, right? And you get a lot of u you know and and your co and your and your other people around you are like still in high school. they they they they you know they don't know they haven't figured out what to do with their life and then there is you who get all of this attention and you know you get a bit out of it gets out of your hand.

Another advice I would like to give that probably would apply more is to always be proactive. There are a lot of people I think that are kind of you know that that just experience life and they don't really impact it. I I would say you should uh try to be active. I don't think I think talking to people and be confident is it doesn't hurt anyone. Um I think ultimately um I know it's going to sound a bit weird, but nobody really cares about you in general.

You should always be direct. You should not think of people like they are gods or anything like that. You should just think like they are you're equal and talk them like they they're your equal because that's how realistically people expect you to treat them. So that's what I'm going to say. So just because you don't have any of experience, you shouldn't be you shouldn't feel like scared, I guess.

Awesome. Yeah, thank you so much, Julia. That's really nice. Oh, that's a great I think it's a perfect way to finish the the AMA here. Um and uh I would just say I think this is exactly what the EPF uh tries to push people to do to just start contributing to be active not to wait for mentors or somebody to tell them what to do but uh just be there and start building because the door are open and as you can see Julio managed to also be self-taught and uh build all his career himself.

So uh huge respect for that and yeah thank you so much for being here today. We are right at the hour. So, it was uh yeah, perfect way to finish it. Awesome.

Awesome. Yeah. Thank you so much everyone for attending. Um and uh thanks again Julio for all your wonderful uh answers, all your great advice and um I hope to hope to see you again soon. Yeah.

Yeah. Okay.

Awesome.

Thank you.

Um are you coming to Dev Connect?

Maybe. I don't know. Um I need because it's far and I need to think about it. Uh yeah, I I probably will. I hope I will.

I hope.

Okay. So hopefully we'll see you there and uh fellows can meet you maybe there uh afterwards. Yeah. Awesome.

Very well. So thanks again and uh have a great day everyone.