Ethereum Sequencing and Preconfirmations Call #5

Transcript

okay uh welcome everyone this is the ethereum uh sequencing and pre-confirmation call number five uh technically our sixth call and today I'm super excited to have Aztec Labs uh discuss base rollups and considerations regarding layer one censorship and I think this is uh potentially a a call that's going to discuss some of the breakthroughs that we did in in London during the the sequencing call there so without further Ado I'll hand the floor to Aztec Labs hi everyone um hopefully you can

see this okay if you can't Justin please feel free to yell at me um all good yeah nice uh so I'm Cooper I'm on the product team at Aztec Labs uh historically we've called me the head of decentralization my job has been making sure that we launch kind of fully decentralized from uh day one of of Aztec and what that looks like for us is kind of Designing our sequencing selection software and how the provs will talk to each other um very excited to share today you know kind of our current thoughts on based rollups we've been talking about it with you know Justin and all the people on this call for for many weeks now and uh yeah super excited to be here um yeah so many people have asked in the past week uh just to be super clear Aztec is currently not Aztec is

currently not a base rollup but it may become one eventually I think that this is kind of important to kind of reconcile we think about this as a potential future road map item for Aztec and potentially a very important one um and so just you know little meeme for good measure this is where I'd put my base roll up if I had one and we definitely might have one uh in the future um and so yeah just kick it off with some kind of disclaimers a lot of these are my current weekly held opinions I think a lot of why we're here is to get other people's opinions um many of them are kind of consistent throughout Aztec Labs we're a growing in a big organization um and we're trying to build a very diverse Community where everyone has you know very conflicting opinions if if everyone on at Aztec was

on the same page uh that's kind of a less uh fruitful world to live in I think than one in which we're kind of driving ourselves forward and pushing ourselves to get to the best design space um we do and will continue to disagree with each other a lot so if you talk to someone at Aztec labs they might not agree with me they might not agree with the project I think that that's actually healthy if you're trying to build the decentralized blockchain ecosystem um and yeah key point is eventually azex road map will be in the hands of its Community we really mean this when we try to launch day one with a fully sequence decentralized sequencing improving set we uh want it to be in their hands um and we'll talk a little bit more about what this means um and yeah and so regardless of

what we say what I say what my takes are uh it's important to know that the community can become based later right like this is not uh a decision that we want to make in isolation this is not a decision we want to make on behalf of our community we want to make a decision that sets the community up to be most likely to be successful over time and make decisions that they think are in the best interest of Aztec um cool sorry one second uh part two of the disclaimers we're not super here to debate any specific definitions philosophies Etc uh we would love if you debate these publicly uh especially in our Research Forum I'll share a specific link later um we're not here to sell you on anything or encourage people to use

Aztec that's not the goal of why we're here uh we're here to share some ideas in a neutral format with our peers um yeah thank you to Justin and you know this is probably the most stacked set of calls I've ever been on in my life and so uh the amount of energy here is just you know truly inspiring for us and so we're really excited about that and we're here because we think those people can actually help us improve our designs and we think that these people can actually help Aztec be a better Network um and we're here because we think that these might just help other layer tws in general like you know our research is not asz seex specific um and yeah so I'll talk this is kind of the end of my preface here but the most important takeaway from this is like we'd love for anyone on

this call anyone interested to be part of azex community like our goal is just to incredibly neutral bring privacy to ethereum and we're going to have a series and an end number of difficult decisions throughout as seex life cycle and we'd love for the people on this call to help us uh make difficult decisions um and so yeah azex Community is really the one at the end of the day that I think should be making this decision on whether to be based or not I think we can try to set up the community to be as successful as possible and that's kind of how we see ourselves stewarding uh the Aztec Network so uh thank you for listening to my preface uh enough of that let's kind of get started here um what are the benefits for Aztec becoming base um it can actually provide

faster block times than Aztec's because current MVP designs I think a lot of people say that you might need you know espresso or some of these other shared sequencers or an another mechanism to get finality faster than L1 block times uh which is very true in Aztec Current Designs are block times are way slower than ethereum's block times and so if we do do this you know we can actually meaningfully improve uh asex block times it's less engineering work in theory I think that this has always been one of the key value propositions of reusing L1 PBS infrastructure etc etc all the L1 consensus uh synchronous composability with other Ro ofs in theory is something that is a huge you know value proposition and we you know cannot ignore

that um and it's more ethereum it is more in spirit with Aztec goals of bring privacy you know as credibly neutral to ethereum as possible so I think like everyone in Aztec would agree with these I think you know this is this is very straightforward uh what are kind of like the high level counterarguments that we we're currently seeing for becoming based um our road mapap more explicitly becomes ethereum's road map right if you become based you're kind of inheriting the changes coming to the aking layer to proposing to inclusion list all these other things at least in a more meaningful and direct way than you know other layeres that might have their own sequencing infrastructure or something like that uh and this means that you have a reduction of sovereignty right like you do have to inherent ethereum's road map

and you know effectively the social consensus on what should go into ethereum's road map and I think that there's like a fundamental question that we ask ourselves is like do we want to internalize this uncertainty or kind of externalize it you know for a lot of teams who aren't super opinionated it probably makes a lot of sense to externalize this uncertainty and kind of ride with ethereum's road map um but these are like really you know kind of important questions to to the project and to the community um like what if we want to deploy immutably and things just kind of drastically change right like do we have as much confidence in in this design uh as as something that we can kind of internalize um and like this is more of like a a Project Specific philosophy and Aztec generally likes to build stuff we

don't love Outsourcing uh we don't want to Outsource our risk or our road map in a lot of cases and so I think you know this kind of sovereignty of our specific road map is you know one of the key arguments or one of the key reasons we would consider not becoming Bas is we can come up with a road map that's ideal in our mind not ideal in the mind of Le the entire ethereum ecosystem in a way it is less well defined or understood economic impact um it could definitely have just as good economics you know Justin and others have talked a lot about how you can SL have your own token you can SL have me you can have fee shares you can have all those things but there's like a 40 IQ question in our mind or at least my mind of like are

projects with large staking networks just kind of worth more um just from the Market's perspective like you know not like in Practical terms of Economic Security or otherwise but like Will the Market perceive something with their own large distributed staking set as as a more valuable project to potentially invest in than something that doesn't have one we do think that L1 censorship could impact the value proposition uh we'll talk about this you know throughout the presentation um specifically ad seex ability to get significantly faster block times are kind of predicated on a minimal amount of soft censorship right if people are delaying Aztec transactions and the goal is to try to produce Aztec transactions or blocks as quickly as ethereum can produce blocks

soft censorship can kind of eat into that value proposition in a way uh but we'll talk more about that throughout the presentation and similarly L1 congestion could impact the value propositions uh like if blog blobs are backlog for inclusion due to high demand or otherwise uh those are not things that are unique to becoming based right all roll up to deal with soft censorship or blob you know gas prices spiking and stuff like that um we'll talk a little bit about how they might be worse though uh in in a based Aztec design so these are like super super high level counterarguments and you'll know that most of these are like in a sense kind of like a business decision or like a project decision right they're not inherently just technical um they're kind of of like what do you want out of

out of your project um yeah cost could go up with that one and so to be based or not to be based really you know this is a multi-billion dollar question for Aztec for all the other l2s that are considering it and it's not something that you can kind of take lightly you know it's it's a it's a really important thing to think about and this is like kind of how Aztec is thinking about its red map just to set set the stage for for people to have you know some context uh we currently have a local developer environment you can write no R programs you can deploy them to an Aztec mode and we're currently building a test net I would say it's coming soon uh everyone says that but hopefully our actually coming soon if you really want you can deploy the local developer environment

to a persistent node and have like a one node test net but that's that's not as interesting I think and this is my opinion is we should build a very basic fully decentralized sequencer protocol uh this is azte articulated on our Research Forum if you go to forum. a. network uh we did a request for comments on it many of the wonderful people on this call uh submitted other designs and things like that to it but I think we should build like a very very basic fully decentralized protocol and this should be like the v0 kind of of Aztec in way um it might not even support contract deployments like it might just be for sequencers and in this case we could actually set up the sequencer set themselves to choose to upgrade the

network to enable contractor deployment so you know choose to enable application deployments and I think that after we have this kind of MVP of a fully decentralized credibly neutral privacy focused rooll up on ethereum right like we hadit a huge milestone like we would be the you know first meaningful rollup that has fully programmable privacy on ethereum even if there's like obvious ways that it could be improved like I think that that's just like a really important milestone for the industry and there's a lot of incentives to try to get there as quickly as possible but kind of as part of this we we do think that the community the Aztec Community IE the sequencers and the token holders and everyone on this call that cares about privacy on a ium uh should consider performance

optimizations to be you know the priority for the first or second upgrade to Aztec uh we see two kind of different paths forward here and obviously one of them is BAS Roll-Ups and Bas sequencing and so we'll talk about that and so you know thinking through this model of our road map like if if the first upgrade second upgrade is going to be specific for performance improvements there's there's kind of two categorical ways as can have faster block times one is actually going opposed to Bas FS one's kind of becoming more Sovereign is in a way uh building an L2 based consensus protocol and uh one of them is becoming oidium if we can just throw like an ungodly amount of Da somewhere we can kind of become an optimistic uh ZK rollup uh we'll talk about this in a bit the other option is becoming

based and uh we can have a very very fast- based R by having faster cryptography uh I don't know if this actually means real time proving in the same way that other people on on the call does um Aztec has clients side proving even in V1 uh so you send as a user as Z knowledge proof to the network um but generally Aztec can have faster block times in a base setting by having faster cryptography obviously faster cryptography will make Aztec faster across the board that does you know kind of goes without saying uh another option is having significant economic collateral on block proposals in a base design um so we'll talk a little bit about about the

trade-offs here um the Counterpoint to building some of these things so an L2 based consensus protocol is like this isn't hard a lot of people have built L2 consensus protocols this is just engineering scope if we want to invest in this direction we kind of you know it's a pretty clear path forward becoming a validium that's not very ethereum aligned of us that's just you know that's not kind of what we're doing yeah Justin yeah I just wanted to highlight that um becoming a validium and becoming a rollup is somehow a fogal to the sequencing you can be a based validium or base roll up and I also wouldn't say that it's not necessarily very E Line like if you do choose to use eum

sequencing then sry it's eum settlement alone then you you're kind of inheriting you know maybe e as money to an extent um and if you choose to do sequencing then you're even more aligned even though you might choose to not use the da layer totally agree yeah it's a it's definitely a spectrum yeah I had a question on it Justin I guess like if you have like a normal case of 10minute block times but you're you're kind of using CH of Da to um enable based sequencing for your kind of uh normal kind of like rollup data can you hear me hear you I can hear you the connection is not very good but I can

just about hear you yeah basically like if you end up creating a lot of applications where the security of transactions B some kind of like pending State UPS um on an alternative da and yeah like it feels a little bit of a weird position because you have a lot of transactions that would not happen um uh if you were just using ethereum they would happen if you're using Alterna so those don't inherit security because like you Takeo um your trading on state that doesn't exist um and only exists in in the kind of alternative da universe so still a

statement right so when you choose an alternative da layer like you do lose some things like you do lose a sensorship resistance because this other da layer can potentially censor you um you do lose like liveness and you know maybe there's a a data availability failure where n haacker kind of basically manages to to fool the da layer to make it think data is available when actually it's not aail available and then your your validum like completely stalls it just can't move forward which is a a really big failure um but on a assuming that there is no failure on a slot bys slot basis then you still can get the full synchronous

composability um and the other advantages that uh that base sequencing provide yeah that makes sense I also don't think we're talking about any of these things as like mutually exclusive right like you could you would use multiple these techniques to make block times faster um I appreciate it um yeah faster cryptography counterpoints this hard like obviously everyone's trying to make cryptography faster we're working on a new proving system called honk um and with significant economic collateral this could potentially centralize around the entity willing to post collateral it could also push out solo proposers and increase potential censorship concerns and so like very high level like we kind of see this as as what's

you know favorable and not favorable if we had to top you know consolidate it down to top two three things in each category base rollups you get sequence and synchronous composability you're more ethereum align you do have some censorship risks and you do have a little bit of an uncertain road map uh in Sovereign you're less sensitive to those sensorship risks you can kind of internalize them and people in your validator set kind of opting into the risk that may come uh with proposing or sequencing for a privacy pokus Slayer 2 a lot of engineering it might be better for the Aztec ecosystem we're not convinced there's there's a question Mark there for a reason and it does have weaker composability um I'd say that we're mostly concerned on the uncertainty on the road map and censorship risks which is what we're here to continue talking

about and I'm excited today we've done a lot of really nice research in my opinion on becoming based um we're excited to share most of that research today uh it was posted maybe an hour ago on the Aztec Research Forum um and yeah shout out to lass our lead smart contract engineer for This research uh I think he articulated a nice idea about collaborative based rollups which might be a new articulation there and so how base ZK in general could work and so basically to be a based rollup all the data needed to prove the correctness of the state transition function needs to be made available somewhere uh as Justin kind of mentioned it could be made on validium or an Alda layer it could be made

available on ethereum and like in a very simple case it would be possible for Aztec to become a base rup today like our research very conclusively says if you know this is the direction that the community wants to go we could do this today um however it has some issues and we would argue that many of them are unique to privacy protocols um and so in a simple naive based Aztec Aztec would send all the data needed to construct a block to L1 as part of the block proposal transactions uh otherwise client side proof withholding attacks are pretty easily possible you could put up a block proposal and never reveal the proofs behind the proposal and it could never get proven so you'd have to have some kind of for Choice rule or something like that so this is the most simple model you could possibly come up with

right if you had effectively uh enough da you could put up the block proposal and you could put up all the proofs behind the block proposal again Aztec is a privacy roll up you have clients that generated zero knowledge proofs you're not just sending a transaction and because you're sending clients that generated proofs these are massive proofs like very very large and so this alone would use the current three blobs on ethereum uh and Aztec would be caped out point8 transactions per second um if you're questioning this I highly encourage you to go look at the Aztec Forum L articulated some really really good stuff um but we know for a fact that this would work um it would be quite cool it would be very

based and another idea um potentially one that we think is more more feasible and could get you beyond uh some of those initial limitations is kind of an economically based version of Aztec right so we have like an na based Aztec and we have an economically based Aztec uh and so you could treat Aztec data similar to how pre-confirmation work in a shared sequencer you could require like very large economic bonds that effectively act as a guarantee that the proposal and the proofs behind that lock proposal will eventually be revealed uh and will be made available and If not you'd have some pork Choice rule that determines kind of slashing conditions um maybe a mental model is that the per block Bond could be more

more than the max ethereum Mev ever extracted to prevent proof withholding attacks you know just taking it to the extreme uh there would never be an incentive to put up a block proposal that would get kind of reorg and you'd get slashed because you can just make the Bon so arbitrary High arbitrarily high that they will always you know put up a proposal and reveal the data behind it um definitely just go go look at the Forum research uh and search for fork Choice rules if you're curious how that would work and uh yeah so these are like the two kind of naive options uh that we have articulated so far and we know like we're 100% confident that either of these would work and we know that there's a clear path to to kind of

scaling these things right through internal cryptography improvements through compression through you know more da on ethereum uh and in the Forum we actually say that if if you aren't like Aztec and you don't have all these privacy zkps from your client side proofs you probably should be a base rollup is actually our current take I think we we clearly State this in in the research um and so yeah this is kind of where we got at least on the on the two design options right now currently available um and we're going to talk a little bit more about censorship and and we'll talk about some other concerns um and so yeah what kind of L1 censorship may exist and how does this kind of impact Bas sequencing so you have soft censorship right delaying inclusion of transactions

or blobs and then you have like strong censorship right is like if they just simply never got included I think that these are kind of well agreed upon definitions and so what can be done about L1 blob censorship um like the naive idea is like we're worried about L1 transaction censorship you could have blob inclusion lists um I don't think that there's a huge amount of support here and so another idea that we talked about is an idea of like a shared encrypted blob registry I think that this is kind of a cool idea um and how it would work is you would encrypt all the blobs you would have all the rollups or all rollups that wanted to opt into the system submit blobs the same registry from effectively EAS or one-off accounts um and you'd have rups actually

decrypt these blobs as part of their proofs right as part of your your proof verification you would reference this version registry um it's kind of a cool idea um and it's completely supported right now on ethereum no changes necessary or anything like that uh you'd have to you know build some smart contracts and update your circuits and all those things but you could do this it'd be pretty cool and so like I guess a question from for from our side is like this would work like we would no longer really be worried about blob censorship but does this make economically make sense economically and from like a risk management perspective right like any you know if Aztec is using a encrypted blob you know registry with a bunch of other rollups they're kind of opting

into sharing some of Aztec's concerns in a way um it might make sense for all the Privacy protocols or all the Privacy Focus l2s to go in this direction and kind of work together on on solutions that you know could solve our issues collectively um and yeah shout out to Ben Fish and dank Rod I think that they originally talked about this encrypted blob registry idea with Joe and L A couple weeks ago in London um it's kind of a cool idea we're not super sure if it makes sense if like other people would actually use it either economically or just from risk management like kind of a cool idea uh what can be done about L1 soft transaction censorship um this is kind of you know going against the the the based role of design in a way um and is

like in option right it's not mutually exclusive or or otherwise is you you could settle infrequently right you could settle every 24 hours or you know n hours and you can basically mitigate uh and acknowledge that soft censorship can and will occur and you just lean into it right you try to make those soft censorship concerns less real- time concerns for for your users um and yeah so you know in this logic decoupling Aztec and leaning in could actually you know uh make more sense than trying to get closer to the entities knowing to be censoring you try to bifurcate like user pre-confirmation and and what they're worried about versus kind of settling onto ethereum and what can be done about strong L1 censorship um I mean we like inclusion lists and uh I would say

having like guarantees on one inclusion list will ship would actually mitigate a lot of the concerns here um like a worst case fallback right if Aztec just like can't use ethereum in some way like it's actually causing liveness failures or otherwise is you're kind of of only option at some point is to be an L1 um and uh obviously we don't want to do that otherwise we would have been building an L1 a while ago um but we would you know again encourage people to consider adding inclusion list to to Electra um Justin you have another question yeah so um I actually don't think that inclusion list helps at all with strong censorship so basically the strong censorship is when there's more

than 51% of the ATT testers who want to censor something and uh they will just ignore the inclusion list if they if they uh if they want to do that so inclusion lists only help with weak censorship makes sense thank you for clarifying I've updated the slide uh yeah and if that happens I guess the only option is really to become an L1 um but I I don't think that we see this is like a realistic you know possibility um and we are still excited about inclusion list um uh yeah so main findings uh Bas rollups are much easier to implement for optimistic rollups than ZK rollups um I think that there's just it's a simpler

design space uh Losi articulates this nicely in our Forum research post uh base RS May provide like a very viable path forward to decreasing ASX block times um if concerns about censorships collectively L2 could come up with shared standards that effectively mitigate this and kind of share this risk you know through this encrypted registry idea otherwise um if ethereum wants to have serious privacy focused solution we think it needs to retain its credible neutrality it's the most credible neutral platform that exists literally in the whole world which is why we're all excited about it um we do advocate for inclusion listen in Electra uh finding new more neutral Builders stuff like that we think could help and

increasing da limitations specifically because privacy protocols have a lot of Da and we would prefer not to use alt da um and yeah so these are the main findings of our research um all of these things are pretty practical and they could be implemented like right now like the the short-term time Horizon is is quite promising um and yeah so let's let's take a poll like we're serious when we actually want azex red map to be fully driven by the community um I don't know if we can do this in Zoom Justin but i' I'd be curious if we could take a poll of people on the call and see if they think that Aztec should become a based R yeah so we can actually do this on telegram I will create a poll uh um we can do this now whoever is online so

should Aztec this is fun become a based rollup and then we have yes no show results yeah and like I I would say like we're very serious we you know want to set the community up in the best way we want to ship something neutral that has privacy to ethereum and we don't want to control its road map we don't want to force the project into our direction or direction that we think is the best and and so not saying we'll abide by the results of this poll but I'd be curious to see how it tracks over time and uh you know maybe at some point some some version of voting like this is actually

going to make the decision to upgrade Aztec to becoming a base roller so um so here are some preliminary results we have 17 who voted yes and one person only who voted no so maybe worth uh hearing out who who voted no I think it's Noah yeah uh and last slide I have here is like we really want all these discussions debates to live longterm uh at least in our research on our Research Forum we think it's much more approachable accessible it's you know better for information preservation um so please please we would encourage you guys to to take a look at this um and yeah follow us on Twitter um thank you for the time the the poll's not legally binding to uh

odyss uh and chat um Aztec team members are allowed to vote Santiago uh I actually think Santiago would be in the based Camp uh he would he would be in favor of Aztec becoming a based roller uh Ben yeah what's up yeah if you're um okay doing a bit of Q&A I wasn't sure what the agenda is but uh I had a question about some things um so if you could go back to the the part where you were talking about sort of the tradeoff between um being a base rollup and having sovereignty I was

curious to understand a bit more there uh yeah I mean high level we think that uh I mean like there there's a couple takes there right like at a high level we don't love token voting like for upgrades and governance and we trying to design a world in which the sequencers of Aztec the provs of Aztec the node operators of Aztec have the ability to dictate the Aztec road map in a more meaningful way trying to manifest or reflect the social consensus that you see on ethereum where you actually require and you know need the node operators to upgrade their software um and so we you know really in this kind of model or if we go with the based approach really the only way that

we see viable to upgrade the network is through some form of token voting um we're not going to have like a centralized upgrade Council the foundation themselves otherwise isn't going to kind of dictate Network upgrades um and so like maybe that's hopefully just like a practical example of where we see a trade-off with sovereignty but why is that incompatible with being a base rollup that's what I didn't understand it's not incompatible we're saying it could reduce your sovereignty like it limits the design space of what our upgrades could be yeah I didn't understand that yeah so how does the how does the aspect of involving L1 uh validators um or allowing them to

participate as sequencers reduce your ability to do that I didn't understand so you could imagine up go ahead it's whether there's a staking set or not so if you have proposers also St I think Joe's at the airport uh we should probably just mute him and um definition of like any e any propos can propos then uh would reduce kind of the set of people who are a line who care about operating as see you're you're wored that it will reduce the number of people who are um like staking and participating in azteq

because if all the L1 proposers are like you know have have the ability to participate as sequencers then it may just reduce the overall size of your community of external sequencers who care about Aztec and who would be participating in this decision did I understand correctly I sort of tried to piece together yeah yeah Joe is cutting out there um I I think that you you somewhat summarized it um generally in this model we would be less likely to entirely rely on sequencers to facilitate Aztec upgrades because they would care about you know so many other things besides us versus a sequencing set who exclusively somewhat

dedicated to our community and what's in our best interest yeah okay well anyways I won't take more time it just it seems like uh to me personally just my it seems a bit of a stretch like it seems like you could engender a lot of interest in the Aztec from the Aztec community and and decision- making that without you know necessarily having everyone participate in in sequencing but let's table this and talk about another time super interesting thank you Hooper yeah of course I mean I think it's fair and we're not saying it completely like gets rid of your sovereignty or anything at least through our current lens it does change how we would upgrade the network and we'd probably be more reliant on

token voting in this model uh is at least our our current thinking I think Noah you are next thanks for add me up I guess I'm the one person that voted no on this I think the main reasons why is I believe that synchronous composability may be easier if you are using a sovereign bft style consensus um for that synchronous composability as well as using I wanted to hear kind of what are you guys kind of thoughts about how you're using alt da with based sequencing because wouldn't you just be posting kind of a commitment hash for a bundle of transactions let's say we're using slj you're let's say we're posting five

transactions on slj wouldn't you just post that kind of transaction hash on ethereum then so then technically the ordering is still happening on the alternative da layer at point I just kind of want to get your thoughts on that that was kind of why would knew knows because I think the base sequencing with Al gets a bit confusing as well as synchronous composability might just be easier with a sovereign layer because you don't have to worry about theum itself at that point and it can also be cheaper then I think uh we would generally agree um not sure if there was like a a specific question in there but yeah the design space for base sequencing who is also using ethereum da seems much more clean simple and kind of elegant in a

sense also Aztec as a pro as a project who very much believes in like governance minimization and doesn't really want to have any formal governance you know Celestia is pretty good about this like they you know mostly do social consensus but I'd imagine some of the other da layers have their own governance and Aztec would be hesitant to introduce them into mix thank you for voting uh contrary to the the herd I guess glad too yeah yeah uh J I think you are next yeah I just wanted to comment a little bit on on Ben's question like there is a loss of sovereignty in a very specific case which is when you've chosen to be a base roll up and you've

completely given up on your governance so you're fully mutable and that's something you mentioned in your presentation then basically you become subservient to whatever sequencing dl1 does having said that my guess is that Aztec will have governance for some years to come and the other guess that I would make is that uh L1 sequencing you know will will aify over time and so there might be a point in time where if you do want to become immutable well actually the L1 itself will also be very close to mutal yeah that totally makes sense um I've been advocating for Aztec to try to deploy IM mutably for a while now I think it would be really cool uh but

there's obviously a lot of practical reasons not to uh and I agree it's much more pragmatic on a a long-term time Horizon if ethereum is more aifi so um right yeah and then to go back to better now just on on that point I mean even if the even if the rollup instances are immutable which is our Cur plan so you have like v0 V1 V2 with the seate deployments and you don't upgrade between them the community can um put for Bridges and and we the portals as it's coming into the the network um the community could decide to use an Aztec Sovereign secret inet to move assets between mutal versions um and you lose that in a base model because that that

right would be conferred to now random ethereum deaners who may not have the same uh interests um so you have to have token voting as a way to kind of stop malicious malicious moving of assets is our current yeah Joe's going on a rabbit hole of our governance plans uh the current plans are high level have like an mutable version registry a mutable deployment so you could add new things to the version registry and so uh the first deployment may or may not be immutable and then if the community wants to deploy something with formal governance they could do so um I do think it would be really cool to also get to a point in a world where there might be multiple versions of Aztec deployed simultaneously some of them

might be immutable some of them might not be some of them might be based some them might not be based to give you different characteristics out of the Privacy protocol you choose to use uh Daniel I believe you were next yeah thanks I wanted to add one small comment related to the sovereignty that Ben brought about uh it's it's similar to what Justin mentioned but on the technical side so there is a specific reason why you would lose sovereignty over uh when when using base robes and that's practically due to the restriction on who is eligible to sequence on A1 that's purely the uh one

of the requirements of being a based Dr so that's a and you mention it briefly a couple of minutes ago so you are losing sovereignty in a way because course you are forced to sequence in some specific way so we have limits on the let's say uh technical upgrades that you are um able to do but that's um that may or may not be uh something that's small enough for you to consider a good tradeoff totally Mak sense um sorry to like just jump in here but because I feel like this is something that people are getting confused on a lot so like I don't I think we've discussed this a lot

but base base sequencing doesn't mean that you are you necessarily are restricting the set of parties who are allowed to participate in sequencing to the L1 validators you are giving the L1 validators the option to participate and you are giving them a privileged partic privileged ability as well to participate but there's like a range of designs that we discuss and so maybe for Aztec for example it may make sense for to um have the control in the hands of a uh a sovereign sequencing set that only uh sells the right to propose to L1 validators when it makes sense because the whole point of involving L1

validators is for synchronous composability with the L1 um and uh you don't like I I think we've discussed this times but like base rollups doesn't necessarily mean that you you give up the sovereignty to L1 validators you can also if you have a way of including the L1 validators in a way that enables um synchronous composability at least part of the time when it economically makes sense for your rollup that is a form of Bas sequencing and that may mitigate the concerns that you have around giving up sovereignty but in that case the the the main benefit with universal synchronous composability becomes blurred because if you allow for

option Ally transferring the sequencing to L1 proposers or the L1 coinbase at a given SWOT then you're reducing Universal synchronous composability by default so unless everyone decides to sequence in a very specific way uh you uh lower the chances of universal synchronous composability yeah it's a design trade-off but like as we discussed right then Aztec may be somewhere on the design tradeoff I think the point that I'm making is that it's not completely binary right you know could choose I think we've talked about this where we've had like you know in The Current Designs we have a random leader election that specifically randomly elected leader could choose to sell pre confirmations through espresso

or something like that um and so any any any other way I mean whether it's through expressor or not is is orthogonal but like you could yeah you could you could sell it you could sell it to the L1 validators and there could be a tradeoff or half the time you could give it to the L1 validators half the time you could do it yourself um I think something that we went over in London is that like we want to have sort of a more sort of inclusive uh idea of what Bas sequencing is because it would be a shame to create this completely you know binary iocation between like what base sequencing is and what it isn't for the exact reasons that you're saying here like it would be a shame to have Aztec like never do base sequencing never

involve L inv validators just because that means complete loss of sovereignty so this is a great example of where having something in the middle on the on the on the trade-off space would make sense yeah that that that makes sense to us and like The Current Designs where anyone can run an Aztec sequencer you know we would never restrict them from any based activities or you know otherwise that they may choose to do and make money off of I would also just say on the on the point we did try to articulate at least um our approach or our kind of current thing thinking and and so this is just like the preview of the thing I've been teasing in the Forum and so we're kind of going at it with this definition of a collaborative based

rollup and I know we all need more random definitions and flavors of based rollups but um hopefully this gives you at least a little sense of of how we're thinking about it if you choose to read this and Ben we would love your opinions in the Forum as always uh Ben you still have your hand up do you have other questions no I'm Sor sorry coer I I just gotta lower my hand thank you for very much yeah no worries we're definitely getting wrecked in this poll right now everyone is is choosing Bas sequencing but I guess uh is the right crowd I supp sequencing call so what do you expect there are a lot of very smart

people on this call and we also trust their uh aligned with our goals of wanting to see you know privacy on ethereum so uh Justin yeah I just wanted to go back to Noah's question about how would it work with LDA and base sequencing and especially with respect to synchronous composability so one of the requirements for synchronous composability is that you have shared sequencing and so if you have Alda you're basically introducing a different sequencer at the da level and so what you can still do is basically have synchronous composability for all the validium that have opted in to using Celestia as as a as a DA

layer now that there might be another approach where you can have full synchronous composability for all the based rollups and all the based validium which is a bit more optimistic where basically um the the sequencing happens by the L1 sequencer and they promise to eventually publish on Celestia you know maybe a few seconds later and if they don't then they stand to get slashed uh but it's it's not something that I I've thought of much uh but uh yeah I'm there might be Solutions basically yeah that makes sense um I'd be curious back to the kind of specific question of like an encrypted blob registry if anyone has takes on that

like it's kind of a cool idea economically it seems questionable and from risk management it seems questionable but I'm curious if anyone has thoughts about that specifically um yeah also I liked your inclusion list post thank you um yeah it solves the economics problem and the censorship one but um we still need to propagate blobs faster um so regarding encrypted uh blobs you know uh has this encrypted meol at least for the bids they're encrypted and the and the the bid the bid encryption is important because now you can't really um basically like see what the other blobs are bidding for and and that can be kind of a um you know I

guess a an economic censorship kind of uh angle um so at least like that's that's what we've been thinking on but what can be added is the whole blob in there as well which which then becomes pretty interesting um but you know we would have to kind of really articulate what like the registry looks like and I don't think I was at that um particular discussion in London uh but you know like we could curate a registry through through this mechanism I think um but you know then it then it becomes an A A contention on like who curates the registry is it like a protocol level registry um then is it enforced is it like an inclusion list um does everyone see it uh you know can there be griefing because of that because of the size of

blobs so there's kind of a number of uh problems that um comes with the registry but I generally really like the idea and I think more uh kind of encryption when it comes to blobs where which are not really conten contending on State uh is generally a good thing that makes a lot of sense I'll have to read the rest of your post I know you dropped it like right before this call it was unintentional by the way it just happen to be very kind of you know I guess synergetic at the same time no it's good stuff um Odus I saw you just posted Ian plus 4844 seems more trust minimized than Celestia I would say technically it might be but socially we don't think so Ian DA has like four multi sigs in it um and otherwise um so

kind of a different set of assumptions there um Matthew you were next in the question Q I think so how do you avoid censorship of the entire shared registry uh you couldn't it would still be subject to soft or strong censorship but the idea is it' be less hard to censor specific actors in like the encrypted like the whole encrypted registry could still get pretty much so there's some like critical mass of um like laob buyers who would have to participate to make it useful because even when it's a small industry like you wouldn't be able to censor specific

rollups without censoring everyone so at that point it just might be more worth it to censor everyone is that what you're saying yeah the it would only work if you got significant Buy in from you know pretty popular rollups that are not uh maybe not even doing privacy things right that you wouldn't have you know reasons to to potentially censor um uh Lynn I think you were next yeah I also had a question about that registry part so um so when we say censorship resistance there's like the censorship from like the next sequence there an next proposer it's also see censorship from like the modory the 01 Network and this encrypting the blobs

this protects against the ladra not the formula correct uh it would Pro protect against soft censorship it would help protect against soft sensorship as Justin correctly identified earlier strong censorship would be if like a majority of the validators just kind of never let you touch that registry I guess in a sense but if there's soft censorship for example if the next sequencer said that they do not want to sequence azc transactions so they won't even create blob then that's a form of like how does the encrypting the blobs help that in that case there would be two different concerns you would be worried about

proposals and block like block proposals being censored and separately you would be worried about ta getting uh censored they would they're not super related like even in the current designs where Aztec has its own sequencer set we still might be interested in this idea of an encrypted blob registry uh Justin you can chime in yeah so the idea basically is to to change censorship from a Black List approach to to a whit list approach so right now if you want to censor there's a black list for example the ofac sdn list um but if you have this encryption then the only way to meaningfully sensor is to have a white list of approved contracts and that might still be

possible and everything that's not part of this white list would kind of all be bundled together uh but the the hope is that there's just so much activity on ethereum that the what is not part of the white list would be very economically uh meaningful I guess another comment is that there doesn't have to be one registry there can be an arbitrary number of Registries and actually there could be one registry per BL per encrypted blob so like keeping like one one of the questions is like can you just add the registry to The Blacklist well that doesn't really work because you can just always create new Registries and then another thing is that we already kind of have Registries for example we have in encryptions like just people just post art and like you

know it could just be encrypted White Noise art and no one would ever know um so I mean other things you can do for example is you can you can uh push data in the in the lower bits of a cat image and no no one would know the wiser and then your proof would would would prove that hey actually this cat image is is contains Aztec transactions so I really do think that this encrypted blogger idea is very robust thank you I cannot take credit for it at all yeah uh uh Lyn did that answer your questions maybe probably not just this one um so so you said that the

proposer censorship and da censorship is separate so this is the case where the proposer in honest but the D is but in the Bas case didn't it most the same like still didn't see that so it so for settlement um that would be specific to the AET contract so the settlement transaction could be soft censored and so it could be delayed by the few blocks but the blobs there's no reference to Aztec at the point when they're published no one knows that it's an Aztec blob other than the person who published and so the L1 proposer can't meaningfully censor other than if they're if they have some sort of

whitelist base censorship which in my opinion is not really meaning uh not really workable for them but in the based case doesn't the proposer also create the blob so even um the the blobs are created by the Builder market and the Searcher Market we have proposer Builder separation the L1 proposer only signs off on a header they have no knowledge of the of of the block content and actually that is a security requirement because otherwise the builders and the Searchers can get unbundled I see Lynn and I guess anyone else I'm always happy to get on a call and chat

or you know talk about whatever um especially we didn't answer your question happy to get on a call um George I'll go to you next I also want to thank George in the lime chain folks for uh they give us a good dock and we've had some calls about vanilla based sequencing so so shout out to them thank you my question is not about encrypted blobs uh and this is why I was constantly bumping myself backwards so if anyone has you can yeah yeah I think it's safe to change the subject we are running out of time I don't know if we're going to do the 30 minute overview or you know kind of hang out after but um yeah feel free to change the subject I hate to put you on the spot and if you tell us that maybe we should discuss this in the non-recorded section please

tell me uh because I kind of know your opinion but what's your opinion as you said that your opinions are not atic opinions per per se uh around the specific part of Bas sequencing idea that is uh the soft requirement or the suggestions of usage of reaking I'm not sure I fully understand your question what like what are my thoughts on reaking as a requirement or would would you consider this favorable or unfavorable part if at St is to go based uh I have not seen a governance minimized restating platform and so my opinion is you know I

would encourage the Aztec Community to not adopt third party software that is not you know sharing that same philosophy okay I don't know if that answered your question but maybe it did your opinion answers my question I just wanted to know what your opinion is on that yeah I mean my opinion is we're trying to design Aztec without any type of like upgradeable multisig uh specifically for you know ensuring sufficient decentralization and not necessarily like individual liability or accountability and I think that it you know we're preserving that within our own protocol it would be kind

of strange to adopt third party Solutions who have significant amounts of kind of antithetical design choices obviously I you know love Ian L and all the the reaking folks but Aztec has different concerns than others Cooper thank you so much we're at the hour um so I'm going to wrap this up I do want to very quickly um give the listeners who can't who are not part of the telegram group and update on the poll so uh we've had 31 people vote yes we've had four people vote no so that 89% voted yes of those who voted yes we have Joe Andrews from Aztec and Santiago

also from uh Aztec and of those who voted no uh we have quintus who voted no just uh because he wanted to be contrarian um on that note thanks thanks Cooper again for this wonderful presentation and now we have the 30 minute open discussion uh that's off the Record thank you