The Case for Monolithic Rollups | Jonathan Wu - Aztec Labs

really excited to be your MC for the right now and 2: p.m. we have a jam-packed agenda with some speakers presentations great stuff going on I'm really excited for the very first kickoff here with Jonathan woo head of growth at astec Labs he's going to be diving into the case for monolithic rollups let's have a big round for Jonathan woo am I good okay uh thanks for coming out eth Denver uh big conversation right now between monism very hard word to say and modularity I wanted to talk through some of the trade-offs and how Aztec thinks about it um and we're going to do a lot of memes and we're going to talk about this analogy of an ice cream shop so let's get going um for those who aren't aware um if you've been living under a rock modularity is this promise of decomposing blockchains into kind of single purpose components that can be remixed and the idea is each of these components is very good at the one thing they do and so thank you Celestia for this uh great diagram there are execution layers settlement layers consensus and data Avail availability um and each of those kind of does one thing very well and the notion is you can construct new blockchains and rollups very rapidly by purchasing these off-the-shelf components seems awesome um so I want to kind of try to steal man the case uh for modularity I think the number one most important thing is speed right that's what they're advertising they're saying you know rather than build all this stuff yourself rolling your own uh infrastructure get to Market quickly buy the off-the-shelf components um the second is specialization like maybe you don't want to build all the levels of the stack maybe your focus is just on execution and you want the fastest VM why don't you spend all your resources doing that take the the other components off the shelf and the last one is security kind of like you know why are you investing so much in bootstrap trapping validation when you could just rely on other people to do that stuff for you I think it's a very very compelling argument um and all that leads to this kind of selling case like the more you externalize the faster you launch um I think that's pretty awesome and the vision here is you know there's going to be a thousand rollups there's going to be a trillion rollups there's going to be infinite block space okay so what are the premises like what premises do you have to accept in order to kind of like go into a modular universe I mean the first is you have to be like the security assumptions are like same same you know like you're getting a lot of convenience and like you're not making many security trade-offs um the second is just integration is easier and faster than rolling your own right you could build all these pieces but it's a lot easier to just pick them off the shelf and kind of stitch them together and then finally kind of picking the right combo of these off-the-shelf components is going to be a differentiator right I think you have to accept all these premises to be like modularity is awesome okay so let me talk about why maybe it's not not so awesome um you know one thing is is hidden security bottlenecks like you're relying on all of these vendors to kind of like provide the different components for you but you're really only as secure as your least secure Lego block like you're putting all these things together um and you're you're hoping that everyone's doing a great job in the back end and that risk compounds across all of your modular components and so there's some failure risk that could happen across your componentry you're kind of not in control um and you're hoping that uh you know none of them have a catastrophic failure and create a liveness risk for your rollup and because the infrastructure is shared right the aggregate prize for attacking just one of these shared components grows bigger and bigger as we go to like millions of rollups like imagine if there's like an Infinity rollups and they secure all the money in the world and they're all riing on like a couple small infrastructure providers and each of them is carrying the weight of like the entire crypto Universe um well then maybe you might think well the security budget or the security Capital across any one of these components might not be big enough to support that ecosystem the second is like integration isn't painless and the reason why I know integration isn't painless is because there's an emerging category called rollups as a service they're basically systems integrators they're telling you hey you know you want to pick some stuff off the shelf let me tell you what to build and not only that we'll do all the stitching for you and so maybe you don't need to hire devs internally to do all the stitching but you have to now pay some kind of SAS fee to kind of integrate and maintain all these pieces kind of takes work to figure that out um and finally like is selecting the right combo sufficient to win you know you talk to Founders they're taking these Lego pieces they're put putting them together okay well anyone can do that so is that a sufficient differentiator open question in my mind so I I think this all leads the what I call kind of like sfic of rollups where in order to coordinate the different pieces right you've got you know different government governance requirements maybe your da layer or your execution provider or um you know validation provider like they all have you know some of them are going to be governed by you know weighted token voting some of them by by multisig and you're kind of at the whims of all these decisions from your vendors like very very similar to SAS right you've got all these SAS software products they're able to unilaterally change the terms on you and so what you kind of need is you need a way to like coordinate all these people and how are you going to coordinate like governance internalized governance is already really hard but now you've got governance across all these vendors that you have to keep track of if you're going to build a modularized rollup so kind of like here come the lawyers like if you're a modular customer you know you're kind of subject to the whims of your provider and if you're a provider you can create liveness failures by unilaterally implementing changes and so you kind of need like a fallback you need to be like okay in case of a failure I'll go back to ethereum for instance for da um and you kind of need some version of like an SLA you kind of just need a please don't screw me type contract and then that how does that get enforced well you could try to enforce it socially you could you know have some kind kind of uh token governance where you you kind of have a say and Upstream to what your suppliers are doing um but really what it is is like it has to be governed by some kind of legal contract in a court where if your service provider doesn't do right by you you don't re you can't really reach over there and fix the problem on your own behalf and so all this means is kind of like a reduction in sovereignty for rollups and so like how do we think about this at Aztec we're privacy first layer 2 on ethereum and our goal is to be a fully credibly neutral decentralized privacy protocol with minimized governance and so when I think when we think about externalization right internal versus external like does externalization get us closer to our goals like we're very well resourced we have money we need governance minimization we require deep integration because we have some novel blockchain features um we don't want privacy set fragmentation in the end modularis kind of just doesn't make sense for us that doesn't mean it doesn't make sense for everyone but it means in my opinion that modularity is is not a maximalist position to take so some questions to ask if you are a rollup founder kind of like is it worth it to me to internalize most of these Services if you're like a100 million rollup like you're securing $100 million I me make maybe it makes a lot of sense to trust a $10 billion modular solution as as part of your infrastructure stack if you're a hundred billion do rollup and you're heavily resourced maybe it makes sense to roll your own and kind of buy back your sovereignty so I have kind of this assertion that the rollups getting going with modularity once they get very very big are going to start thinking about how to buy their own sovereignty back they're going to be like wait a second I have infinite money now it kind of makes sense for me to re internalize all of these different pieces um and if anyone can launch a rollup then like if you're a rollup founder you know what makes me different if anyone can just click a few buttons and do it well then what makes me different and so my analogy here is kind of like Baskin Robbins for the Americans here you know they got 31 Flavors and so like you're a rollup founder you're going to Baskin Robbins you're going to you're trying to make an ice cream Sunday and like you're like I'll have a chocolate and Manilla and I'll get some sprinkles on top and everyone can choose to make their own custom sunde right but out of these kind of like prepackaged ingredients like there are only so many flavors and there only so many sprinkle sprinkles and only so many cones and you're looking at your Sunday and you're going my Sunday is really really special cuz it's got all of these exotic components that I bought off the shelf and I put together but like anyone can put those things together anyone can go into Baskin Robins pick those flavors put them into their own Sunday so where where exactly are you trying to create distance technical distance between your project and others and to make a Marvel reference this is like the doctor strange from Avengers where uh you know he's imagining all possible future worlds and there's like 14 million combinations it's like is there any one where we win and he's like there's only one there's only one outcome where we win I think this is the logic the endstate logic of off-the-shelf modularis which is like it is a combinatorics problem there's n degrees of dimensionality you can make the most exotic Sunday because there's going to be so many different providers and the hope there is like I'm going to pick the one combo that's the tastiest combo and that's going to make me win and I think that's a very interesting hypothesis and you have to think about as a rollup founder um is that is that the the path forward is my ice cream sunday special or unique in any way or is it something that anyone can come in and order so my assertion and I'm biased because I'm a marketer is that modularity leads to commodification and commodification leads to this apotheosis of marketing anyone can make the same Sunday so what happens when everyone can make the same Sundays and am I okay externalizing my road map you know now my road map is not really up to me and my Dev team it's up to all my vendors and my vendors are going to upgrade as they're going to want to upgrade so my overall conclusion is modularity is a tradeoff of sovereignty for Speed and I won't go super deep on this but you know when we thought about this problem it's like okay monopolism fun word it's well understood it is very costly it's very costly to invest in your own security Capital right to kind of go and get people out there to commit money and funds and effort to kind of securing your network um but it gets integrated economics we have hand and glove compatibility seamlessness we don't have to rely on third party vendors and sfic we don't really have to rely on slas and courts to kind of like resolve disputes between the different parts of the system it allows for Less coordination and it's more governance light and there's consistency of vision and control and for us being a privacy first rollup um having a very very different complicated ultimately challenging infrastructure to manage we there are no real off the shelf components like we're hand making our own little sprinkles we're like hand churning the vanilla ice cream and so for us it it's very challenging to think about okay we're going to custom make all of these pieces that are very idiosyncratic how how is that going to fit with these offto shelf components for us it doesn't really make a lot of sense on the other hand for modularity you have this fantastic speed to Market you can launch very quickly um there is some integration overhead but at the end of the day you know you get to ship way faster than we could I mean we've been at it for many many years and we still haven't shipped the thing and so someone could look at that and go no speed to Market is my priority so what do I what do I think happens like next in this modular Universe um the first and most obvious thing is there's going to be a proliferation of ice cream flavors like right now there's like a couple da layers and there's a couple VMS you can choose from and you know couple settlement layers Etc we're going to see just a vast proliferation of these ice cream flavor like it's not going to be 31 it's going to be a trillion flavors and it's there's a strong incentive there to be a winning flavor like if you invent the vanilla like the flavor that everybody wants in their Sunday that's awesome so we're going to but we're going to see a huge proliferation right modular service providers but then if you go up one level like think about the chess game one move deeper I think we're also going to see a lot of ice cream shops we're going to see a lot of roll up as a service providers these are kind of like Channel Distributors right every ice cream shop is going to say we carry our own little special set of flavors that you can't get anywhere else and if you come to our ice cream shop we're able to offer you the most unique Sunday and so now you're going to see competition between Distributors being like I need exclusivity for your modular piece because I don't want to serve the same ice cream flavors as everyone else that does not make me as a distributor and an integrator competitive so the sfic means like there's going to be Preferred channel partners and that means there's going to be a creation of the these exclusive flavors these Channel agreements right and ultimately we're going to get this highly industrialized systematized rollup Creation in trillions of rollups and again I'm biased because I'm a marketer but I think that in that world marketers inherit the earth in that world nobody really has technical differentiation um and and marketers inherit the earth so I'll just I'll just end it here I know I'm a little bit early on time um I'll just end it here with a little diagram of like how I think that it plays out kind of like modularity is forcing Founders into thinking about how important is my engineering team how important is internalization how important is um owning the secret sauce and why is it is it more important to just get something out there very rapidly and with minimal overhead and we're going to have this externalization and Outsourcing of everything that is hard on a technical perspective both from the componentry to the integration um security and ultimately post posters will rule the world right like the tech ultimately won't be owned inside the tech will be completely outsourced and essentially the rarest thing right now and it's the hardest thing to get will become even rarer which is narrative creation marketing those really hard to find CMOS so that's kind of how I see the world evolving and the case that I would make for monopolism which allows you to kind of internalize some of these differentiators uh thank you for your time how's it going hey all right thank you Jonathan another round of applause for Jonathan that's all you got come on eth Denver let's go um so we have about 5 to 10 minutes before our next presentation so go take a restroom break go grab some grub come back here in about five or 10 minutes and we're going to kick things off so stick around for