EPF7 AMA | Thomas

Recording is on.

Awesome. So once again, welcome everyone. Uh thank you so much for joining our office hours today. Uh we have another uh special guest today, Thomas Poke of Ethereum Foundation's uh protocol architecture team. Um so yeah, thank you so much for joining us for accepting our invitation.

Um yeah, I wanted you to meet our fellows. uh have them to learn more about you and your your work your vision about the protocol and um yeah I mean I'll just start with that like so people understand who are you where you're coming from I'm curious more about your background like uh what did you do before Ethereum how did you join uh what was the your first steps in the protocol

yeah um yeah thanks for having me first uh and hello everyone it's always nice to see fellows Um yeah uh I so my background is in academia and then I was doing something completely different. Um I was doing computational neuroscience. So I did my PhD in that um and there I studied the neural correlates of decision making in humans. So basically you have a bunch of people in in doing brain scans when when they are like sort of like making decisions and you you analyze those scans and you try to sort of like understand what's going on in their brains when they make uh decisions and so I was doing that for a long time. Um, and at the end of my PhD, uh, yeah, I guess I wanted to try something different just because I had been in academia for my whole life.

Um, and I liked it, but like I yeah, I just like wanted to see if there was anything else I would prefer or like a bit more. Uh, and so yeah, I decided to go full on like full-time crypto by then, like right after my PhD, my PhD. Um, and at the time I took like more of I guess like research engineering job at Dune, Dune Analytics. Um, yeah, before I was very involved in like crypto, but like more as a hobby um in Dows and when this it was still a thing. Uh, and so like I don't know doing some some freelance jobs but like very yeah it was it was not taking a lot of my time.

I was just like a sort of like an enthusiast I guess. Um, but then I started full-time at Dune and it uh it was really good. I learned a lot. I I really liked also like the pace. It was very it was like in the beginning of of Dune.

So it was like very you know new startup. uh the pace was quite high but the result is like that I ended up doing a lot more engineering than I wanted to like I thought there was going to be still a bit of research and I missed doing like actual research um and so yeah after basically a year at Dune I decided to you know either go back to academia and do research uh in my sort of like original field or like try to find a way to do research uh in crypto And that's that's basically when I joined the the DF. Um for me it was great because I I did like the space a lot. Um and um you know it was an actual place to do almost pure research. I feel like my work is a lot more applied these days, but at the time it was still um very um yeah, it could we could have like some academic papers and some like pure sort of like fundamental research and um and still be in the space.

So that's basically it.

Nice. Thank you so much. That's very interesting. I had no idea you had neuroscience PhD background. It's really cool.

Um and yeah um glad that you made it to the the ecosystem. So um obviously a lot of work you have done since then. So I'm curious about um um some of the highlights that you worked on or maybe what is also like the transition from the rig. What is your role these days? What are some um uh yeah what is what is it like to be a researcher at the Ethereum Foundation basically?

Yeah. Um so I mean as I said I started I guess more on the sort of like research academic side a bit like I joined DF and I was working in the beginning at DF I was actually working a lot on me. So I was part of that that that team that was called the robust incentives group that was really focused on incentives. Um and so I would say it was a bit half and half like it was between suggesting ideas for the protocol. Um but also like looking a lot at at actually publishing academic papers and and having like really strong ties with academia.

Um and yeah so it was very fun. um very focused on me on like the build on the market and the the builders and and then like gradually it's kind of nice because if you think about it like incentives are basically everywhere uh on on Ethereum like in every sort of like aspects of the protocol or like you you will get some some kind of incentives in there um because in the end you need to coordinate this sort of like ecosystem of of nodes that have like individual interest and you need to make sure like they don't sort of like abuse the system um by sort of like taking advantage of like an incentive structure that is just like not optimal or um so yeah I was very focused on this. It was very interesting and I think I got to touch a lot of parts of the protocol because of this like because like incentives are everywhere like you get questions from like a lot of people working on on very different things and so it's cool because it allows you to to sort of like um touch a lot of different parts. Um I would say gradually my research has become a lot more applied like even these days like I have some interns um and I I like sometimes like give give feedback on papers or whatever but like I'd rather sort of like I don't know write blog posts and and implement things and sort of like get things moving. Um, and yeah, I focused quite a lot after my sort of like me period.

I guess I re I focused a lot on on on Fossil and on censorship resistance, but uh is still part of my everyday everyday life even though I'm now also co-ch championing with with with Jun, which is good by the way. Like I think co- championing is actually very nice because it you know championing like a especially like a headliner or something like it takes a long time and and um yeah it's a bit of a grind. So um being like more than one people I think is good and yeah fossil um I don't know if everyone knows what it is but like um and I'm happy to take questions but like it's basically like um it actually stems from like stems from my me work because I saw how the builder market was centralized and what sort of like impact it had on the protocol um and for is just like a way to force include transactions um in the protocol um and some like yeah give better inclusion guarantees for for transactions that come in by sort of like for forcing the hand uh of of builders. So they are sort of like they have to include the transactions that everyone else sees. Um and so that that is still quite present and now it's scheduled as a headliner for Hegota.

So the I guess the political grind is over. Like the research grind, the research is over. The the politics of of governance are are pretty much over. Now it's really about like making sure it ships on time and and is safe and and you have a bunch of coordination. Um and now I'm also will say I spend most of most of my time on privacy.

So I'm leading the protocol privacy uh project. Um and that means mostly over trying to like oversee what changes we need to ship to the protocol to get sort of like you know better privacy on on Ethereum one. Um and almost surprisingly like the changes were not um so crazy especially after we have like some big changes like like frame transactions or or like native account abstraction and like you need actually like a very very few changes to actually sort of like completely improve um privacy on on on the other one. And by completely improve is basically most of the crux crux of is like crux of it is removing um relayers and broadcasters. So you remove this sort of like third party that is that has always been there um uh on for privacy on Ethereum and and so by shipping a few EIPs in the protocol you can basically get rid of them which is really good because then you actually get trustless privacy which wasn't the case before.

Um and so working on this a lot. Um, and it's interesting to work on privacy because yes, you can ship like EIPs um, and sort of like change the protocol. And I feel like that is mostly figured out right now, at least conceptually, but like you need to integrate it with like everything else. Like you you need to think about the apps, like the privacy apps, you need to think about the wallet. Um, because like if one of these things ends up like leaking your privacy and then you you failed.

uh like even if you have like shipped the right tip. So it's been very interesting in that aspect because you you get to talk to a lot of like different parts that are that also sit outside of the protocol sometimes but that that are also obviously extremely interesting. Um so yeah I've been working on this quite a lot [snorts] and um yeah I guess another big part of my uh well yeah it depends on the day I guess but like uh I'm part of the architecture team and that also means maintaining the the stro map um and so you know it's I would say it's still like driven by Justin quite a lot but um still I get a lot of questions and people that have what Usually usually people are unhappy about like some parts of the store map and want to sort of like give suggestions or remove things or add things and everything and that takes quite a lot of uh time to sort of like try to align people um around the single sort of like document. uh that is of course like not uh perfect or not uh the source of truth. But um but that takes quite a bit of time to get like input and feedback from everyone and trying to sort of like um uh make sure we have a coherent road map and and view for for what the protocol will look like in the upcoming years.

So yeah,

nice nice awesome. Thank you Thomas. Um yeah, lots lots of great stuff. you worked on and I just want to say I mean huge respect for like uh championing fussil for pushing it so hard. I mean uh it was such a fight to get it adopted as a headliner moved to uh to Hegotaa and that was just yeah amazing that you never gave up on it and um uh yeah so like huge respect for this and I'm curious my question is like what keeps you going like what motivates you uh to work on the protocol like there's obviously um uh this problem in research where you just discover so many uh paths that don't lead any way you need to backtrack you need to uh deal with so much politics in Ethereum.

So yeah, how do you keep yourself motivated and uh yeah, what keeps you going?

Uh yeah, thanks. That's very appreciated. I think yeah, I think part of it is obviously like a big part of it is it's very missiondriven. Like I I I I basically believe Ethereum should be um a if not D like because I'm not sure that like exists right now but like the way we can um have like unstoppable privacy [laughter] uh for for financial transactions. I feel like that's uh something that is um you know there in in some aspects in for for other projects but like that I think we can really do a really great job at at it and and until it's done I'm like [laughter] I'm not very uh satisfied and I think another thing is personality wise I'm quite stubborn I guess so I that also plays a role I think I um and everyone needs to be um but No, I think it it at the end of the day it like the research part is always probably like the most exciting and interesting part for me in terms of like how yeah enthusiastic I can get about like how I understood something new and it's it's super interesting.

But the the pushing part is like almost the only thing that matters because um that's how you get like to have like an actual impact in in the world and and that um that uh that needs a bit more stubbornness I guess and and and grind but um but we need to do it otherwise um especially given like how relatively big Ethereum is and like how many divergent opinions there are like if you strongly push for what you believe in. It's probably just like never going going to get there.

Yeah. Awesome. Thank you so much. I mean, that's Yeah, that's so good to have like missiondriven people with the right values here and uh um also like to see your work being adopted properly, especially now also like the privacy stuff getting more attention. Uh like it's I'm very glad to glad to see all of that.

And well, we have a bunch of followers here who work on some of the stuff coming from your work. Somebody here implementing fossil or uh statelessness adjacent stuff or frame transaction adjacent stuff. So folks, now I'm curious uh about your input. Please uh raise your hand. Feel free to write your question in the chat if you don't have a mic.

Um I'm I'm curious if you have any questions for Thomas here. Um please feel free to ask. um whether it's from your experience working on something related or uh just uh curious about the the research happening the past the future. Okay, we have we have oh so many hands. Awesome.

I think Mo was first. Go ahead.

Yeah. Uh hi Thomas. Uh you have an interesting background like a PhD in computational neuroscience. So I wanted to ask like uh by learning the neural mechanisms of decision making uh have you applied that to your current work like in the MEB part where you design incentives or like because uh like the architecture and research is a combination of different aspects like game theory, economics and computer science. So I wanted to ask your take on that.

Uh yeah that's a very good question. I think um initially how I entered it is um but then you know yeah I don't know uh how I entered it first is like I was basically analyzing networks um using mostly graph theory and and sort of like tools like this um and those were like brain networks and at first when I was like working like freelancing a bit or like I was I really approached it through the data lens. So I was just like applying some of the same techniques to understand blockchain networks instead of braid networks and that was sort of like my door into the whole thing. Um but of course like yeah it's very different and so like that's that's stuck for like a bit you know but like you don't you don't do that many interesting like that much interesting stuff if you stick to just like analyzing blood blockchain network like can yeah uncover patterns and whatever but like it's it's quite far from the protocol work um I think you know the neuroscience sort of like department I was involved in was also very close to psychology. Um, and I had some some people in my lab working on um quite a lot on on game theory as well.

Um, so for me there was also like a nice sort of like segue that I can I could use to sort of like understand some of the the protocol things. But ultimately to be super honest like yeah my background is like was easy in that I sort of like was coding a bit. Um I think the Dune experience was actually good because I learned to actually code for people in in produ which wasn't the case when I was a researcher. I was just like having my little Python scripts and it uh you know it wasn't like super useful but um but yeah apart from coding some game theory and some understanding on like how networks work um you have to learn most of it from scratch like uh and and it's doable like because the you know it's not like a sort of like a [clears throat] established field that has been going on for like decades or whatever. So you can come in and sort of like get get it and focus on some some some small parts and then yeah um grow um by by learning about more topics and sort of like I think organically it actually still works.

Great. Thank you.

Awesome. Thank you guys. Um Charles, go ahead.

Uh hi Thomas. Uh coming from Zerox, I'm pretty familiar with the impact me and dexes have on end users. Uh when you think about the ME endame for Ethereum, like what are we actually aiming for? Is the goal to minimize ME itself or is some amount of me inevitable and just something we have to learn to manage? And what approaches do you think are the most promising for helping us get there?

Yeah. Um yeah. Well, my the [clears throat] thing I I'm sort of like I was sort of very bullish on and and a bit disappointed because it it's actually hard to ship is encrypted me pools. I feel like that's a very obvious thing we should do at some point u because you know it's right now we have like two options. either you get into the public man poolool and you get sandwiched and and and and you know front run and and it's just like not the greatest user experience or you use centralized parties and you just like send your data to them.

Um I think you know the encrypted me is basically to just to to get the best of both worlds. Um the problem is like we've I've worked on this a bit and like people have worked on it a lot more than I have but I've been following it and like um you need you know you need some quite advanced uh and not very efficient um threshold [clears throat] cryptography scheme that is also PQ because we don't really want to ship anything that's not postquantum secure on the protocol. Uh, and yeah, we don't really have like amazing solutions for it yet. Um, all the other solutions are sort of like make, you know, either R&P PQ and it's just like not not going to happen. Um, or you have, um, you know, proposals like Lucid, but you have this sort of like trade-offs that you have to make with like people that can, you know, submit their transaction and then not choose not to reveal and that gives them an option and it's it's a bit messy.

Um, so I feel like it's very I'm very hopeful we can get there. Uh, and I think it's a very crucial component of of how to mitigate ME. Um, but unfortunately I don't think we'll get there in the next, you know, one, two years. Uh,

so the problem right now you see is too many constraints like we need to make sure that we're postquantum in addition to solving this particular problem and right now we just don't know how to do it or or it's a sequencing problem. Yeah, we don't the the blocker right now is is uh is annoying because it's a hard one, but it's the cryptography and we don't have like a good PQ threshold threshold encryption scheme that would sort of like allow us to do that. Um, you know, it's never good to be bearish on cryptography. I feel like it it it moves faster than sometimes we even expect. So hopefully there will be uh things that allow it allow us to do it but um but for now it's just not the case.

So we don't have like a ready sort of like solution that we know will work and and that we can ship. Um so there are tons of like workar around and people trying to find ways not to to have to do it but it's uh yeah it's it's hard.

Cool. Thanks.

Thank you.

Nice. Uh thank you guys. And we have Rahul.

Uh thank you Mario. So hey Thomas, it's uh nice to meet you here. Uh Christian and myself both are working on implementing Fossil for Lighthouse. Um there's something I've been wondering about recently with frames and fossil and the conversations that that's going on there. Uh so my question to you would be do you think uh fossil should eventually you know support arbitrary frames or do you think like there's always going to be this class of frames that you know fossil just can't really guarantee for inclusion.

Yeah. Hi thanks. Uh [snorts] yeah uh I linked anformational EIP that I wrote that is basically Yeah. It's a terrible name. I don't know why I named it like this, but it's called BS profiles for for fossil or whatever.

But like the the the crux of the thisformational EIP is to um explain how frames and fossil work uh together. And I I feel well in my in my opinion it's pretty obvious that fossil cannot support arbitrary frames uh in the sort of like completely arbitrary frame uh thing. And now having said that I think you have a trade-off between um supporting more frames and sort of like more categories of transactions let's say or like broader a broader set of transactions um and eventually I think there is a trade-off with you know okay in a future in which we have partial partial stateless nodes um because you know state grows a lot grows a lot and it's very likely that we'll have some nodes that just want to keep a minimal part of of the state and that state could be what we call VOPS which is minimal validity only uh partial statelessness and it's basically the only state you need to make sure transactions are valid in the me pool um and so there is a trade-off here between um allowing more frame transactions to get to benefit from growing that state that everyone will need to keep to make sure transactions are valid. And so that that's the trade-off we have to deal with. And you know, either we want nodes to be very light and we don't want them to store a lot of state because we care about the centralization and the ease of sort of like spinning up nodes or we want like the maximum amount of frame transactions to be uh to benefit from.

But we can't really have both. So we have to think about what makes sense. Um and I think opinions will diver diverge a bit. Uh I don't think but but you know for me what should be inv for example is like the the key nons and the recent roots. So like I think privacy private transactions for example like should definitely uh benefit from fossil.

Um so yeah that's there's that that dimension and then there there is the you know the dimension of what can even be accepted in the public meool which is a bit different from what can be part of fossil but um and then all the discussions we have about the max u verify gas we can allow uh in the public man poolool and you know basically how much uh resources can this the validation part of of uh the transaction consume. Um and I think that's something else we have to work on. Um and it it somehow affects falsehood because uh you know if it can't be broadcast in the public man pool then you have a lot less like the pro probabilities of like an including that transaction is is going to be a lot less. you have people working on alt me pools like custom me pools um for you know maybe nodes that would want to sort of like see these transactions but that that need maybe additional sort of like bandwidth requirements or something like this um so you have like all these discussions I think they are really good discussions to have um um but uh but yeah that's how I see it

awesome thank you so much for answering Awesome. Thank you. Uh, yeah, very interesting. Ask Kasa the question. Go ahead.

Oh, can you hear me? Right.

Yeah.

All right. Yeah. Hey, um, regarding privacy and kind of the recent situation about frames and account obstruction as well. Do you think we will have some separation of the privacy guarantees for L1 and L2 in the end of can L2s being less private and more compliant and Ethereum staying uh kind of with its own values? Uh so yeah and if it cause some more issues on the development as well because like more transaction types potentially that's like kind of just more things to maintain uh and so on.

What do you think? [snorts]

Um yeah. Um so weirdly, you know, we have like right now we have an L2 that's Aztec and that is like very focused on privacy and that actually covers more than we are covering in on Ethereum because they also, you know, have private contracts and and things we just like don't have on the other one. Um generally I feel like yeah um I'm not sure exactly how like privacy will evolve on on L2 to be honest like I do feel like compliance and regulations will will play obviously like a role for the ones that are that are in the US for example and that seem to be like the dominant ones today. So so I'm not exactly sure. Um what I'm sure about is like um you know we privacy on L1 um the plan is for it to be um you know yeah I guess like the the the privacy spectrum is so broad and you also have like two dimensions at least that's sort of like how I see it.

say I see like the privacy dimension and what you you cover in terms of what you hide basically and then you have the trust dimension and like how how how trustless your your system really is. I think for for Ethereum what we are trying to hide um with sort of like the privacy pool or protocol model is like who and sort of like trying to break the link between who's behind the transaction but still having sort of like the amounts and and the balances and everything sort of like visible. um some some um and and for it to be completely trustless. Um so that's like the first goal. I think for example L2s like Aztec have like cover a bit more like they have private contracts and we don't have it on on Ethereum but of course like it's on L2.

So I think there is a a bit more trust uh because you need to for example trust the bridge. Um and then yeah you have like Zcash that takes like a completely different approach and that just like enshrines the pool for for for transfers. Um but so yeah I what I'm I'm trying to say is like I do feel like there will be many approaches and and we need to probably get it right in in like the sort of like point in the trade-off space we want to be at. I think the the the obvious thing is like we want it uh to be completely you know trustless censorship resistant um on on the L1 uh and that's a good place to start with not sure how it will evolve on on all the L2s.

Okay. Uh yeah thanks very much for an answer. Also um I wanted to again if I can also hijack and ask one more thing um also about something that you mentioned within the stack which is the RPC which basically gets uh too much of the user data. There is not that much options we have at the moment and there is also like a third way to run your own node and uh use um kind of self-hosting to to solve the this problem. Um yes and where I wanted to to come with this yan.

Oh yeah and the question would be do you think in the future something would also change on the like kind of this network protocol level that would allow us to like kind of have a completely different this mindset of broadcasting the transactions. Um yeah so the I mean two things right like first is like I guess private reads or like calls to RPCs without sort of like revealing what state you need um for your transaction. Um right so there are like a bunch of approaches um I think it's it's a bit more on the sort of like haku wallet side of things but if I remember correctly there is basically pir that can be actually used to retrieve state without sort of like revealing what state you are actually interested in. Um so that's nice. I I think it's getting to a point that it's sort of like efficient enough for it to work.

Uh and then there is um what is it called Tbased Oram systems um so they are they are based on T so they don't um they don't like offer the sort of like same guarantees but I think at least they are very efficient and so I think the cohack team is basically testing both options and trying to see which one makes the most the most sense in the very short term and maybe in the sort of like longer time after for for network you know metadata I think the sort of like practical solution would be is to go through to tour right now um there are you know I I I hope we we will get to a sort of like Ethereum custom based um mixnet or sort of like an anonymous sort of like bro broadcast technique that is more suited than be faster on everything down the road, but like right now I think the idea is to use to and to route traffic for private transactions directly through to um as in it would it wouldn't wouldn't even be like something users choose or whatever. It would just be part of the wallet uh and by default all your your transactions and your traffic will be rooted through through the door. Okay. Yeah, thank you very much. There is also interesting tool I discovered called Unknown RPC.

It's on the Ethereum um repo also addressing the specific problem. Thank you.

Yeah, thanks. Um yeah, so the thing is like it's very hard to to spin up a new anonymous broadcast channel from scratch, but um but yeah, I think I I will check out again and see if it has evolved. Thanks, Awesome. Uh, great discussion folks. Thank you so much for keeping it going.

Um, yeah, I'm curious to hear more from you. If you guys have any more questions, uh, please keep it coming. Um, we got a couple more people who joined since we started. So, if you have any questions about the research, about the future, about um, something uh, that's Ethereum is going to turn into Um, okay. So, I'm going to fill the Okay, Charles, also go ahead.

Yeah, I apologize if this has been partially answered already, but uh my question is really about privacy in ERC20s. Uh, when you think about privacy for for tokens other than ETH? Um, how do you see that working for for ERC20s? Can existing tokens become private through some protocol mechanism or would we need a new token standard? Uh yeah, that's a good uh question.

Um I think uh the the problem I see with ERC20s generally is that you know on the sort of like I'm really looking at it from the protocol side and and we are trying to sort of like design around if I guess because if you want like trustless private transactions then you don't want to add this sort of like trust in the sort of like stable coin um owner that can just like freeze funds and everything. So I feel like you know that sort of like added trust you that is baked in ERC20s is a bit of a of a reason the why we are not like super focused on it. Um right now there are like privacy protocols like you know privacy pools and and rail gun and everything that have like um private pools for ERC20s and I think that model sort of works like it's uh yeah I think it it's it's a it's a good one like uh and you don't like they don't need any sort of like other standards right they they just like sort of like break the the link between the owner and and the and uh and they have like like basically the the only thing is like you still are vulnerable to me stuff and whatever because like all the amounts and the swaps are public um but at least you are like breaking the link between who you are and who's making that that trade on unis swap or who's sort of like transferring this stable coin. So, so that exists that is like sort of like live on Ethereum and it doesn't need another token standard, right? Like it's it's working on rail gun.

It's working on privacy pools. It's just like not something I really want to to focus on right now because I'm I'm focused on rem removing uh sort of like the trust you put in in relays and and broadcasters and and there is one um there is one um I mean the trust in sort of like the stable provider you just can't remove. So So yeah, I'm I'm just less a bit less focused on this. Do you know is there any uh has there been any discussion to enshrine a token standard to make it part of the protocol or will this always be um treated as just a you know community norm use this contract?

Um so a token standard or or a pool because those are different.

Uh I mean I mean basically make ERC20s more like ETH part natively part of the protocol or something like ERC20s,

right? I think I actually think I saw proposals but I'm not deep into it. I'm I'm way deeper in um sorry I guess my question is really is do you think enshrining something like ERC20s is a good idea or a bad idea? Um, no. I think it's a bad idea.

Just um, so yeah, maybe maybe I I can explain also a bit more why, but like the white listing of it is just like really hard to deal with. like um you know I don't think we want to necessarily sort of like uh accept any sort of like kind of ERC20 uh like all the RC20s cuz like a lot of them are malicious and like it's just like hard to sort of like um trust all all of these sort of like ERC20s that are created basically every day. But at the same time when if you start whit listing then it's a nightmare in terms of

I I'm actually thinking about having some sort of generalized ERT20 that everyone just uses. Um not so much everyone's able to contribute custom code that we would have to like audit the code because that then you get into the problem we've recently had with like unis swap uh before.

Uh I'm not I'm not talking about that. I'm talking about is is there like is there would there be a benefit to just having a a generic like token standard that's part of the protocol that everyone just can use not not you know extend in their own ways.

Yeah. Um I don't know would they like what what would be the benefit?

You think they wouldn't use it because someone like Circle would always want to have their own custom add-ons like I want to be able to freeze accounts and that would never become part of the protocol.

Yeah, I think so. Um yeah I mean I haven't like thought about this question super deeply but u maybe one design philosophy that is very key to Etherum in terms of like the protocol part of thing is like not not like enshrine the minimal amount of things uh and don't use white lists and stuff like this if you don't have to. So I feel like there will need to be an extremely good reason for for for us to do it and um um and yeah maybe there is but I haven't really come around like that.

Thanks.

Okay. Awesome. uh sc

oh yeah with this discussion of everything that we mentioned about the native protocol level privacy and arc 20s actually like that reminded me of the zk wormholes when one of the use cases can actually be kind of burning the tokens and modifying the contract in a way that if you can prove the fact that the token was burnt you can claim that anywhere it also grants this kind of nice u privacy on the protocol level as well but I think like there there was nothing much uh new kind of towards that and I'm not sure where things are left off on this one and what do you think like kind of um from the user perspective maybe also uh like what would ideal uh way of using privacy would look for you on Ethereum kind of I I personally use rail gun and sometimes it's just I'll say challenging to be honest and uh

yes

uh yeah um yeah maybe I can give I talked a bit about privacy but I didn't really give a ton of like context or like a good overview of what my my thinking is. Um so yeah, wormholes to start with like maybe let's get this out of the way because there is um it's not secure like there's a collision attack that you can actually do with wormholes and and and I think there is a blog post about it. um I can send it to you but like basically we we found an attack um and it's uh so you can't really use one poses is like you'll you'll have to sort of like tweak it uh and um it's not ideal also um you know maybe to give you a bit more context I think my take is that if you if we manage to ship like the few EAPs I'm trying to sort of like push which are um on top of frames But there are like hidden nons and recent roots. What you do is like you make any private transactions coming from any privacy app on Ethereum native um because the transactions can be broken in the public me. They can benefit from fossil um and so they are trustless um and um uh and yeah and they get rid of the sort of like trusted third party.

Um, I think that's like the first step and I I think I like the idea that you can still have many apps uh on Ethereum that are competing but also like composing with each other to have like an ecosystem of apps. Um, I think like it's a feature. It's not a bug. So that's why I'm like um I'm kind of happy that we can sort of like give these guarantees and these protocol guarantees uh two private transactions on Ethereum without having to enshrine a pool. I think the the trade-off here obviously is that if you don't try a pool um you the anonymity set is uh is not as good right like users will use different apps um and that's um that frag fragments the anonymity anonymity set so I think you know there are like a couple of things that can be done but but um the idea is that there will probably be you one really good h app hopefully that comes out by by the time he also comes out.

Um and that will be fully trustless, no governance, no tokens, uh no, you know, sort of like PPI or or ASP or or like anything like this. um and that will just like come out um like at the same time as Hegota and that will also obviously take advantage of like all the Hegota EIPs and I think that could become like the very sort of like minimal almost tornado cash like app that is really focused on ETH transfers um and then you you can have like an ecosystem of apps around it um that um develop and and for for example like for for stable coins or for I don't know more advanced use cases like private DeFi and stuff like this. Um so that's a bit uh that's my hope like that's the plan. Um and so I don't think you need to enshrine something which is what you know ZK wormholes does like you you sort of like by default enshrine a privacy pool. I actually don't think we we need to do it because if we had like have this sort of like completely immutable app that almost that almost plays the role of this entrant pool um and that also gives space for other private apps to flourish on on Ethereum which I think is a good trade-off.

Very insightful. Thank you very much.

Yeah. Awesome. Mohit, go ahead again.

Yeah. Uh, so, uh, Thomas, you've worked on EIP7805, uh, which is like a proposal designed to preserve censorship resistance. So, so I wanted to ask like in EIP 7805, uh, Fossil relies on multiple inclusion lists committee members to force constraints on block builders. So like how does this architecture handle the unique mempool validation and invalidation challenges from the ERC 4337 account abstraction transactions compared to the standard EOA transactions.

Uh yeah uh you don't you don't need uh much more um for 4337 you actually don't need anything else for anything that is like you know uh not 7702 and not 8141. You just need the nonsense and and and the balance and that's all you need. You need the nonsense balance checks and you have the payload. You take the transaction, you um you append it at the end of the payload and you see whether it's valid or not by just checking its nonsense balance when it it does become a bit more tricky is for for 8141. Um in that you need to maybe you know check the kidances, check the recent roots, uh check maybe like one or two storage slots.

Um um so that is like a bit harder. Um for 7702 you just need to check if there is delegation code associated to it or not. Um but it's can just be a flag basically it's very easy as well. Um but for 4337 like checking the nons balance actually is is enough.

Okay that's interesting. Yeah, cool. Awesome. Uh, any more similar technical question from anyone working on related stuff or something maybe more higher level folks? Feel free to ask.

Um anyway, I'm um yeah, I'm still uh fascinated by all the uh great like censorship related but all the privacy stuff you did. So I'm curious like uh finally of course like a lot of it is being adopted but generally like what do you find as more underrated part of your work? something that maybe deserve more recognition to be humble.

Yeah. Um I think the the thing that doesn't always sort of like like [clears throat] isn't always it sort of like always goes under the radar radar a bit is um the coordination part of it. I yeah, I think there's a ton of work uh that that is just like uh me talking to people and and explaining stuff uh to try to um get them to, you know, work on the same things or or or like be excited about like one one an idea or or like a design I've been working on and stuff like this. um or you know be convinced that an EIP should be prioritized over like another one or stuff like this. And so I feel like there is a ton of work that actually goes into um yeah making sure your ideas are really well communicated and clear and rallying people behind it.

And that's that just doesn't come as naturally as people think. Like yes, you have like the power of ideas and and who proposes it also like plays a role I guess but like there is quite a lot of work of like making making sure like everything is coherent. um you can explain things like from like the very high level to down to like um a very sort of like like more detailed view and and I think that that part is underrated and some sometimes that is like what people miss like a lot of people are convinced they have a good idea and it will just like go through because it's good but like the reality is like it's it's almost never the case like you you need some people to actually push and maintain sort of like morale and enthusiasm around the topic for especially when that what that works like takes months you know like u you need some some some kind of coordination for so I would say that's probably underrated and um I think we need more people sort of like coordinating things

yeah it's one of the biggest challenges in Ethereum uh just uh being able to coordinate on stuff, talking to people. It's stuff that is mostly not seen in the repos, but happens more than more than we can handle. Um yeah. Okay. If there are no more questions coming, I have at least one last I would like to ask you and just u something simple.

uh whether you have any advice for the fellows or I mean putting yourself sort of in these shoes like you know starting in the protocol what is something that you wish you knew at the time or yeah any advice for the fellows

uh yeah I think well it really depends where you start from like uh yeah if you're already very familiar with like the protocol and and and the work there like it's it's obviously a bit different. But I think like if I was to start again, it's usually very nice to stick to one project and sort of like try to push it end to end. doesn't not necessarily like up to my net or whatever because that takes longer than than your your fellowship usually, but you know like I I'd rather have like a small project that is very focused and very clear and very concrete and that actually gets through and achieves like its objectives uh rather than sort of like explorations that keep sort of like going on for for months. uh like I feel like it's basically my advice would be to that it's it's usually better for me at least but it's also personality wise like it's usually better to pick something and go through it even if it um you know if it it seems minor maybe or like just a bit niche um I think it's it's always going to be more useful for me to do this um rather than sort of like have this generalist view or like this sort of like physical philosophical takes on like a bunch of broad topics. Um especially when you start I think for me like now I've I feel like I've become a bit of a generalist but it's by going very deep on like different projects over the years.

It's not by starting trying to sort of like getting all the protocol like the protocol is way too complex for anyone to get in its entirety now unfortunately I guess but like uh you know um I mean even Vitalik like uh it has like a lot of blind spots like it's too there's too too much stuff going on so so I feel like uh you you need to start by finding the niche you really like and you are like really passionate about go deep and then you know once you finish this find another one and go as deep and then over time you'll actually get to know more of the protocol. Uh and also to sort of like adjust based on what you like and what you want to drive or what you are passionate about but um yeah maybe that would be my advice.

Yeah thank you so much. I wish more people are here to hear this because uh that's yeah uh that's great advice. Exactly what we try to achieve in the fellowship like to get people to dig into like one domain and sort of become experts and yeah totally this experience like some um like it's um it's uh also my experience like I remember my first years in EF and uh I was looking into layer 2 when it was still a new thing and I remember correcting Denny Ryan when we had some assumption and I was like in this call with Denny and like you know big name or something and like well he also wasn't you know expert at everything so gave me the the opportunity to uh to shine there to uh uh show that I learned something specific so I hope that you guys uh can take that to the heart uh we have two minutes left maybe if you guys have any more thoughts or some some last remaining questions or or anything else to share um otherwise we can slowly wrap up here um I think We discussed a lot. Um, yeah, I believe we learned a lot today and got inspired. I I find it very motivating.

Uh, great to uh listen to you speaking, Thomas. It was really nice. Um, yeah, it was really good and yeah, I hope you guys learned something, you get inspired and uh uh hopefully maybe you will have a chance to meet Thomas in person at Defcon. Uh so uh there might be an opportunity to chat more later and yeah we'll have the uh EPF day there and dinner and stuff so you all get invited. Yeah.

Awesome. Yeah. Thanks uh thanks for for having me and and yeah feel free to to reach out if you have like any questions and anything. Thanks a lot. Thanks Mario.

Awesome. Of course. Yeah. Thanks so much again everyone and yeah uh huge thanks for uh for joining us today humas. Yeah it was very inspiring.

Uh have a great day everyone. Bye

bye. Thanks.