PQ Ethereum Workshop - Day 2 Kickoff

Transcript

It's been our normal. >> Okay, let's get started. Welcome to day two. Um, today we actually have three uh different rooms. So um we have full bird which is actually uh upstairs and this room is going to be dedicated uh for for the devs. We have another room called the borden room. And here we're mostly going to have uh for more verification topics except possibly for uh a fun last minute uh presentation on Bitcoin post quantum signatures from my pending uh

confirmation. And uh in this room the the Franken room we're basically going to be alternating uh theory implementation theory uh implementation. So, um, I guess what I'd like to do is invite the various technical team leads to say a few words about what they're hoping to achieve today, some of the open problems, talk about the the the agenda. Um, and I guess we could start with the dev team for exampling, but probably not easy in terms of the

work. Um so we'll start often talking about the bindings uh between the uh existing research uh libraries into different implementations with having like C++ and and some rust and then there's also Z. Um and then we'll follow up uh at 11:00 a.m. with metrics. Um this is basically going to be like open discussions regarding like what metrics are we going to measure and also following the networking session we had yesterday. site. We also want to uh embed all the metrics like requirements measurements in terms of networking as well which requires a lower level of of of metrics and measurements and then basically the afternoon is going to be breakout sessions. So we are trying to achieve um the net zero drop today. Um probably a bit quirky but at least like

you know the three clients that we have here should be able to be dropping each other and finalized um on what sort of like infrastructure we're not so sure yet but we're trying to get it to talk to each other and finalize. Um I think that's basically the day. So uh hopefully finishing this demon here today so that tomorrow is going to be about dep one talking about post signatures and also know whether how much of the networking stuff that we can actually networking improvements we're going to try out in de one and yeah perfect um I mean one of the goals of the binding session from my perspective is to try and get the implementers of the reference implementation so Tom and Emil to talk to the devs. Um so a bit of a forcing social forcing function here. Um any

questions for O any questions about the session? Uh and again you know feel free to modify the the agenda. It's yours. If you want to you swap out metrics for ptoin networking that's that's fine. That's fine. Um next up I guess we could do formal verification with Alex. Yes. So we have a pretty packed day with every session back to back. Uh so we'll start with encoding with Fry who's been looking at the uh postm signatures uh which is obviously a nice separate work stream from our clip and and the rest of things. I believe he's alo been doing a lot of work with an agent on this which is quite exciting in terms of workflow.

So we'll we'll get to see a bit what that about what that's about and it's pretty exciting to be doing for verification of this at the same time as people are starting to implement it for the first time as well. Uh then we'll have a session on end to end for verification with Julian. So this session will mostly be about looking at how to expand the scope of the current foration experts to the broader stack. So we're able to cover the entire trust space that would be used for ZKM1 for example. Uh and so that should be pretty interesting and involved I think hopefully for many devs or who is not just a crytographer which has been the main focus of a lot of the form of campus so far and then after that we'll have our clip which should be quite exciting um because this is a library of

snarks in ling and so we're for this one I think mainly hoping to get a lot of photographers in the group and to get some feedback on what we're doing and hope to interest some of them into what we're doing and potentially find new contributors as well. So I think we'll primarily go over the interactive reduction framework that we've been developing and we should have a spec in lean to show off which should be exciting as well. >> Alex, would you say that arclave is the the primary library that's looking at the mathematical foundations? >> It's the only one. >> The only one. Okay. And it's it's part of the film foundation organization. Is that right?

>> Uh yes. Oh yes. >> Okay. I'm the only EF contributor to it, but I guess it's probably funed by EF. >> And how many contributors do we have today? >> Uh 10 to 15. We have about probably six to eight. I would say four contributors and we've had about 15 total including >> and how many uh contributors are you hoping to have tonight? >> More. >> Uh there's never at this point there's never enough. >> Okay. Because with Leo, we can have like this massively collaborative.

>> Yeah. >> Work. Okay. Come on. >> Yes. If you send yours and you have proofs, I will merge them because you can prove the correct. >> Okay. Excellent. >> Very easy. >> Perfect. Very exciting. Um, any questions for Alex? Nope. Okay. Uh, so next up, um, we have, uh, I guess Tom on the, uh, Yes. So first I think that we will have a first binding session that is not related to the implementation but uh I hope that we will fix the point about uh what's need to be done with the various

clients how we can link the work that we have done with all of the primitives and all of the VM stuff with the work that is done by the client team. This would be a very important session especially we discussed that yesterday with Gon how to play with the various languages with z with rest and maybe with others. Um and then we have like two um other session about lin VM and memory argument I think. Um so I think that during the first um the first session we will more talk maybe about the architecture of the VM. We talked about this during the VM day a lot. But now here we are much more expert in the room and much more people.

So maybe after the the quick introduction of yesterday by Emmy, it could be good to have like a a session talking about like the choice of the architecture, why we have this choice or this choice and have a debate with other people in the room that are familiar with with the KVM and maybe a more specific topic at the at the end of the day about memory argument stuff. uh what type of memory are you using in the VM? Maybe it will be like a good moment to talk more about like the new logo stop because a lot of people are not that familiar with this and yet why we did all of those choice and yeah with the

experience of a so it would be really valuable to have like a second point of view that [clears throat] is basically for today. >> Yes. So I think as you said Emil will be presenting a lot of stuff which is very new and seemingly very powerful and Aki has this very I guess exotic memory model that has been implemented in in knock me up that should be a very interesting presentation um I'm also hoping Gish can join the memory argument because he he came up with log up uh in the first place and I think from our perspective the memory argument is very important because we're spending a lot of proof of time on So any any optimization will help and

everyone is using memory argument. So it will also lift boats if you can find optimizations. Okay great. Uh thank you so next up we have okay so uh we have two theory sessions today. The first is uh led by Albert about the toy problem. So just a reminder the toy problem is sort of very strong con conjecture that is used to analyze concrete soundness levels for yeah snarks and starks and Albert has some exciting new results which I'm looking forward to hear about they are also linked the over is also

linked in the hackmd if you want to look at it um and yeah I'm looking forward to that session to learn about these results and then discuss these results and the second session about mutual correlated agreements. So we continue with conjectures. This is sort of a form of of yesterday said orthogonal. Uh another form of conjecture that gives you complete security levels for for snarks and starks and alone and issue that session. And we sort of continue the discussions from yesterday from the proximity price maybe on a more technical [clears throat] level whereas yesterday we talked a lot about you know logistics and that stuff. Yeah that's

it. Yes my understanding is that the toy problem conjecture is false. Uh Albert will present about that and uh for the mutual credit agreement uh Elon and Ori believe also be doing short presentations. Um yeah I think my understanding of maybe what Rich did is that he was able to prove the soundness of where without using mutual we'll see I guess if that's correct. Um, great. Any questions for Okay, great. Um, so today we don't have

any organized um [snorts] group dinner. You can just uh break out in smaller groups, have dinner wherever you want. Um, I believe that a bunch of you have had late night conversations. You're welcome to do that. Um, a cafe or in in the college or around Cambridge. Um and uh yeah, that's pretty much it. We'll be meeting again uh at at 5:00 and um >> um for conversations. Is there a place where we can chill and work and like uh like where should that happen?

>> Yes. So, we have access to the lounge um which is [clears throat] place where we had lunch yesterday. So there's like comfy area that just chat. >> Okay, great. Um, so we have uh seven minutes of of break time and then we'll be starting the session. Thank you.