The Platform Dilemma Facing Every Web3 Developer Today

[Music] yeah you say we're live hello everyone thank you so much for joining us we have an extraordinary panel here today to talk about a very important topic called platform dilemma facing every web3 developer today before we get started I just want to get a quick PLL in the audience how many of you are here developers can you raise your hand everyone a handful of you how many of you are founders four of you how many of you are thinking about becoming a developer in a web through space one of you okay great so you came to the right panel we're in the right place with industry experts from igen layer AWS and chain link so what platform dilemma are we really talking about I first started coming to East Denver in 2019 and when I came to e Denver in 2019 developers didn't have a lot of choice there hadn't been a lot of platforms a lot of L ons you had ethereum a couple of ethereum killers uh little middleware was built at the time so you didn't really have abundance of choice that you have today fast forward to 2024 you have more than 400 blockchains you have dozens of twos you have a lot of middleware tools now you have modular architecture to choose from so how do you make sense of it all how do you find product Market fit in the context of this abundance of choice this is what we're going to cover today so without further Ado I would like to start with quick introductions um I'll go first um I'm Kate Lawrence I'm the CE founder of black accelerate we are web3 investors have been uh investing in this space since 2018 personally I spent 10 years at a Global Research firm Gartner and then I became an ethereum minor in 2016 uh we're proud backers of uh companies like Makow Quant stamp and igen Lair among others so if you're a developer if you're a builder uh we're actively deploying Capital come talk to me after this panel now I would love to pass it on to churro Paul and said thanks Kate uh yeah it's great to be here my name is churro I'm the head of product indust marketing at chain link clabs uh I joined chain link clabs in 2021 and uh it's it's been an amazing ride prior to that I spent about 5 years at Amazon uh working alongside Paul on the Amazon manag blockchain uh Enterprise blockchain product and prior to that I was at Microsoft uh launching uh developer products and AI products Paul yeah hi everyone I'm Paul ferton I'm the GM for uh manage blockchain and web3 services at AWS I've been leading that team for about four years when I jumped into the blockchain space following a 20 plus year career in web 2 building scalable systems at Amazon where I was there for 10 years uh managed SNS and launched pinpoint and a number of other services and I was in Telecom and wireless and unified Communications before that and I saw what was happening in blockchain and I thought this is a a challenge like no other I've done in that 20-year career and and so it continues to perplex me and excite me and I'm excited to be here hello everyone uh I'm Sid s SEL I'm the VP engineering head of engineering at ion Labs uh formerly known as lay Labs uh I joined the team when we were just around seven engineers and have been working for the last year and a month to build out the team uh a team of uh people composed from the the the web3 crypto space from the web 2 space so it's a pretty Dynamic place that we have created to to Really reflect what we trying to build uh in in Igan L uh prior to ig I was with AWS for close to 5 years with AWS EC ec2 ec2 networking no SQL databases engineering teams launched a whole bunch of bunch of foundational products for ews I also spent some time at Facebook uh building out the data infrastructure for implementing privacy controls for all the products uh then of course uh I got super attracted to the whole proposition of the decentralized web uh worked at protocol labs and and on uh decentralized data storage protocols like ipfs and filecoin um and and prior to my uh time and in the US with AWS when I joined them I was in Australia working for traditional Finance big Banks uh as a as a as an architect trying to make sense of all the things that we we are trying to build in for in the financial services space uh but on right from day one I've always been in the P2P middle integration like middleware development space and uh this journey up until this point in time has been a reflection of how technology has evolved to put the user and the developer at the center of everything so super stroke to on the mission we have a Dagon layer so yeah glad to be here so one of the unique perspectives that we're going to hear about today is all three of you come from a very strong web two background and two of you made a transition into web 3 um let's start with defining what platform dilemma means specifically what platform what dilemma and to an extent that you can provide the bridge or the context of how that platform dilemma is different in web 2 versus web 3 that would be super helpful yeah um the way I think about it and internal internalize it is from the context of uh one monolithic versus microservices architectures um so monolithic applications uh have existed from the Advent of software right so uh you think of mainframes I think that was a monolithic uh application close system um it is uh solely uh centralized uh maintain and built by one team and then over time that grows into many teams they're all working on maintaining that one application many interdependencies between all those teams to kind of build that application and launch it very conservative but also very secure like if you still think about it today monolithic applications like mainframes still power a lot of the financial infrastructure uh my first job was as a mainframes developer a long time ago did not enjoy that at all um but then when you think about like as software modernized software best practices modernized uh there was this concept of service oriented architecture and software design uh which pretty much meant that you could break up a monolithic stack into multiple smaller independent software competents and that gives the developers a lot of freedom to uh architect those software components in their own way also take them to Market in their own own way and also that allows those uh different components to be very very composable and also that all those components can come together to really achieve the business function that you are trying to achieve with that monolithic application as well so it still provides the business uh aspects of providing the intent for achieving the business application they needed to but now you have this aspect of composability and uh speed by which you can really move fast you can innovate and then you can quickly leverage it as a building block so you can use those building blocks as literally Lego blocks and you can use them and piece meal together and piece them together into one application now I think the Dilemma kind of comes into the picture here is the interesting aspect as well is that you want to innovate fast you want to move fast uh from a defi perspective you talk about the concept of Def comp composability where you have multiple different pieces coming together as building blocks you could use uh a bridge an L2 roll up um you a stable coin you could all inte integrate them into a new kind of a financial application you could introduce new def Concepts as well into it so so it really helps Innovation you know where defi is today it it would not have been without this concept of microservices and uh moving fast Innovation Etc but I think in my in my uh view where the Dilemma comes is that hey you want to move fast but you also want to move securely because you are securing a lot of value and when you have all these different microservices talking to each other you want to really make sure that uh all those interdependencies uh you are not compromising your security because guess what your overall security of your system is as good as its weakest link so that is why to me uh where the platform dilemma comes from is going monolithic versus microservices yeah interesting when I when I think about sort of coming from a long history of of of sort of traditional software development I think about like system design and software development and monolithic versus versus modular are sort of really traditional design best practices and Concepts and and I think those Concepts when you think about it from Just software and system design perspective um you know they translate from Web Two to web 3 right and if you think about you know I want to start writing software if it's easier for me to get started by just writing all the software in one package having a single environment where you deploy it and you can just keep going but there's limits to how that will scale right how many developers can you contributing to that how scalable can that get and that's really where sort of a modular design principle makes sense in terms of being able to scale that if you think about like just general system and software design you know at AWS S3 is one of our biggest Services it service it it serves you know on an average 100 million requests per second and we've got hundreds of teams working on that um and the only way that would scale it's bu built out of hundreds of different microservices right and so that's sort of just from a pure software design perspective and if you translate that to web 3 we we're seeing starting to see that Evolution where to create Simplicity a lot of blockchains are sort of you know a monolithic client you have a single client everybody runs the same client you can create a decentralized environment really quickly and easily but what we're seeing is having that Allin one stack where everybody has to run the same thing and everything and all the data has to sit on one node we're seeing that that it's having scaling problems it's also having privacy problems and things like this right and so this is now as we're seeing the layer new pop up we're starting to see modularization of that and there's different layers of that modularization where we have modular blockchain design but we also have modularization of the clients themselves where you can pull the storage out of the actual server and Store the store the data elsewhere um if you take that to another layer that's where I think it really becomes a little bit about decentralization versus centralization and I think that's a little bit more of the Dilemma it's related to modularization but I kind of think of it from a system design perspective modularization so when I think about going at another layer and now you want to decentralize all of these functions I want to um make sure that I can trust the function that I'm calling at a given layer and I want to be able to not be have that function controlled by a single entity that's still modularization but it requires a lot more capabilities to do it in a decentralized way and so we're seeing that challenge where you have all of these layer ones to choose from which which one do I build on well I want them to all interoperate oh but they don't scale and so we're seeing all of the effort in towards solving these problems in a decentralized world and and and and Sid will talk about the very interesting things that they're doing in Iron layer to sort of create a decentralized modular infrastructure sort of environment and system design but that's sort of where I see the The Dilemma is that that next layer up where we're really trying to decentralize things yeah sure so I think uh the term platform means different things to different people for example so um the goal of ioner is essentially I I'll actually back take a step back and try to explain the the the main ethos here the he thoughts here here is anybody who wants to build an an application let's start right at the stack at the top of the stack you want to build something for your user ultimately right we are building fancy stuff fancy protocols complicated stuff fine that's all good but ultimately there's a customer there's a user you want to serve so if you start from the top right from the top I want to build a Dap right uh a decentralized app because that's what you believe is the right thing to do for your user the the the properties of decentralized of web 3 are important for your user base absolutely build a appap if those properties of decentralization are not important for your user base don't build a app simple as that so but if you so my premise is that if you want to build a appap a decentralized app you then are realizing that the the very nature of a decentralized app requires a blockchain it requires trustless so which means I need block space um and then you look at let's say Bitcoin or ethereum any of these platforms let's say ethereum a programmable platform and you say my block space is super severely limited so what do I do so I don't really have real world infrastructure Services I can use to build real world daps so I can is saying that okay you know what uh there's a lot of innovation that's going to happen on the DAP layer there a lot of Builders out there who are super well connected to the ecosystems they know what to build let's provide the the infrastru Services let's provide the ability for developers to build infrastructure Services A permissionless and B reflecting the properties of scalability security robustness I mean you're only as good as the as the thing that you use for your application so if I'm deploying on a particular infrastructure layer that infrastructure layer needs to be probably even more resilient than your app itself so I Lay's F goal is to create um U positive some some games and an Open permissionless Marketplace where infrastructure service developers ABS actively validated Services we call them they can actually be secured by the stakers who have restake their their their eth to secure applic infrastructure Services other than the ethereum codee network so once you have this sort of a economic game playing out uh I is essentially providing you this open platform to create these developers to this open platform to go and don't have to bootstrap your own Network anymore you just do what you do best build your infoservice build your Bridge build your data availability layer build your uh sequencer so build all these foundational things while they are being built in they have an offchain implementation so you're not limited by the block space and you make use of The onchain Primitives and leverage the cryptoeconomic security of the ethereum proof of stake blockchain and you get the best of both worlds it's a complex thing to do I understand and that's why we are doing everything in open source but we we approach this from an open Innovation access accelerate open Innovation and let us build real world scalable performant infrastructure services and that will that will create the symbiotic relationship between the infra service and then the dab develop devel developers will see we do actually have Services we can use like sequencers and data availability and and and and rollups that can serve the needs of real world users thousands of transactions per second we can achieve it so it can't be achieved entirely onchain hence a is opening up the entire world of offchain compute to allow traditional web2 developers to come and say you know what to my users these things are important uh uh decentralization sensorship resistance uh user sovereignty uh trustless all these things are important and hence I have real world infrastructure Services I can use so that's the approach we're taking for creating an an open in permissionless uh marketplace where all the builders and capital providers can come and re create this next era of uh real world scalable performance applications so just I want to make sure that we what you talked about is the benefit of decentralization the benefit of using igen layer and we'll cover the tradeoffs in a second but I want to make sure that we have a very clear definition of what platform dilemma is so in your definition or in your opinion is there a dilemma at all yes what is it absolutely so for example some of the benefits that you get typically used to get or and rightly so you want to provide the best quality of service for your users you want to have the your your application should be robust you should have a predictable performance you should be able to predict what your performance is going to be my my P90 latency under a certain amount of load should be this I need to have a degree of predictability in my system I need to be able to choose um uh for example in in web 2 for example you not always able to choose openly freely because you actually have to go through certain providers but what you do get though in in as a tradeoff is predictable performance so similarly here the The Dilemma is that if I go to web 3 am I going to get the same degree of resilience and predictability and that's exactly what I is trying to do is saying that you know what it won't happen on Day One automatically what we need to do is remove the barriers to creating these Services which is basically oh if everybody has to bootstrap their own network blockchain network it's not going to happen because you're not focusing on the core problem so yeah so the so we are trying to solve The Dilemma of uh providing high quality services that developers can build on top of the foundational infrastructure so that's our mission to make that a reality so now let's dive into the tradeoffs because obviously we talked about a few different architectural approaches we talked about the concept of scaling Economic Security of ethereum um specifically let's start with security so how do you balance security while enabling developers to move fast yeah uh that's a great question so um you know um in the web2 space there are a lot of these kind of Concepts that um move fast and break things um build your plane while you fly it and um but the core kind of uh Concept in web three is that you are securing value and the one thing that every developer out there should always prioritize uh is having a Security First mindset and that is the the the way that we can all ensure that more and more value flows into def so from that perspective um when you think about it and some of those principles that people talk about hey you can build your plane while flying it so what does that translate into web 3 so that could translate into hey you could launch a semi Ready or Not Ready or an MVP product into the market you could then continue to add features onto it but guess what when you're trying to add features and you have a very modular approach you have interdependencies now you need to make sure you are testing and auditing all those interdependencies uh then you also need to make sure you are passing all the governance proposals so what is that doing is is that it is counterintuitive to Innovation so why not start with the security first mindset and build a product and have the airplane ready when it leaves the hanger so that it is uh secure for the users and it's secure for the whole ecosystem and is secure for the trillions of dollars that the tradire folks are going to bring to web 3 so uh from a chaining perspective when we think about the approach we think about it in in um in a best of breed way and the way I talk about it is when you use more than one chain link service you are not adding any additional trusted assumptions uh you have the chain link Oracle decentralized Network that is powering all the products and services that leverage it so when you go from data feeds to chaining functions to chaining ccip you're all leveraging those decentralized IED node operators their experience uh who have experiened in devops running Mission critical applications running them 24/7 who um uh are going to ensure your services are resilient robust and and you're getting that whole benefit across all these different products and services now as a developer the benefit that provides to you is you don't have to do the undifferentiated heavy lifting you have a great IDE idea you are now going to implement that idea and bring it to your users guess what you don't have to build your own Oracle you you don't have to reinvent the wheel you can leverage the industry standard uh security focused reliable resilient Oracle Network and leave that leave that work on the Oracle and focus on the business Logic on your creative idea on your Innovative idea so now what you're looking at is you're being modular you're being innovative you're you're you're launching quickly but with that Security First approach with the security first mindset and and leveraging that to build your app and launching in the market yeah look I mean there's always trade-offs in terms of different dimensions and security is one of the big trade-offs you have to make individually as you're as you're looking at what you're trying to build and what you're trying to deliver to customers um you have to assess the risk and the impact of you know relaxing on some aspect of security so that you you can move faster um you know at AWS where we're at you know by default no questions asked security is job number zero right like we don't launch any service without satisfying a very long list of Baseline security controls and operational requirements um and and that's just where we're at right we we we don't want to afford losing that customer trust that's the most important thing for where we're at and I agree with sharog that in the web 3 space this is where maybe we've as a whole on average indexed on the Move Fast Break Stuff and not enough on the security side given the value that's at stake here and and in many of the many apps there may be some apps where that value is is different but we've seen you know a lot of value lost because security probably wasn't treated with the right level of priority and you know Security in the web 3 world certainly poses a different set of challenges right you you're deploying code to this open transparent system the surface area of attack and the threat models are are are very different and more challenging than say in a web web 2 world like we're building at at AWS largely um but a lot of these hacks actually were not just um you know complicated you know bugs in in exploits in complex contract code a lot of them were traditional um lack in security where keys were compromised because somebody broke into a server right and this is this is you know long-standing traditional security and I think we just need to kind of up the game in the space like that particular area is something that you know AWS certainly can help any Builder with right consult with art Security Experts essay and so on to help sort of secure keys and the servers you're running and so on um and then and then as we think about like the challenge of of of that SEC those security layers this is where really thinking about you know and a lot of innovation is happening here uh with KP know Zer knowledge proofs and um you know automated reasoning around smart contract security so there's a lot of innovation that can happen to evolve that but I think we just need to as a whole really up the game in terms of how we think about uh security when we're dealing with um you know customer assets that that that have value to them right but it it sounds to me what you're both are saying is that security is high stakes so you might as well delegate it to best-in class providers who think about security at all times versus trying to develop it yourself internally uh sit for you I have a slightly different twist to the same question on security because obviously the whole premise of aan layer is to lend the economic security of ethereum across other uh service providers so a devil's advocate would say wouldn't you diminish the economic security of ethereum by uh making it more expensive expandable what would you say to that no I mean uh there how you how you use the security that's available it it comes with certain responsibilities of course you can't just do it in a random way you have to think through it's a it's a pretty big area of research for us and that's why we not doing it alone so security is something that you can't go into a room of you know five people and you figure out i' I've soled security you need to basically engage with the with your community and understand how how what are the different way in which security manifests right so Economic Security is just one type of security right but if I have but the other reality of the of the situation is that if my capital is just sitting there doing not much whereas potential could be a lot more so so so the whole point of innovative solutions that you know internet and you know slice bread stuff everybody says that oh there are certain thing or the cloud things are not possible so you always have to take a step forward and think about how can I make how can I make most efficient use of resources and of course yes uh you have to be very careful about the the risks that are associated with implementing innovative solutions but it's not to say that uh those cannot be solved so ethereum already provides ways in which you can have something called objective slashing so it's nothing new to the protocol you're just making it accessible to and useful for other infrastructure Services just beyond the ethereum validator set so as long as you have onchain observable objective slashing so that's an an obvious security you can provide to other service providers so it's just extending what we already have and then you actually able to provide programmatic access to the the uh to the underlying um um security offered by the ethereum network yes those Primitives are being developed not just as a a software update it's actually being developed through governance mechanism so we are setting up V2 committees and and we are setting up uh what similar to Escrow right like you have an unbonding period for abss where you can just randomly withdraw stake and just walk off so there is both process governance as well as the technology pieces that we putting together so and that's how you achieve security you just don't achieve security entirely through software you need to do it through both software and through the governance processes so and and the other part is trust like um because we are working so closely with the ethereum foundation everything is being done in in conjunction with them so it's all open it's all happening in the open and with the right people outside of wagon Labs I think so that is how you you have those discussions about security you don't ignore them you tackle them and you are very honest like there are certain things we should not be doing and certain things we should be doing so let's get together and solve for those so maybe the question to all of you then if you are a developer and you're just about to start building a appap or an app what are the properties that you need to up optimize for in order to go with one or the other type of architecture that you're talking about decentralized versus centralized modular versus monolithic vertically integrated versus horizontally integrated what is that decision tree looking like for you as a developer yeah um it's a good question um I think um you know um having optionality from an infrastructure standpoint um with rollups with the um interoperability with um the D5 stack um I think developers now have the right tool for the right job the way I think about it is you can leverage ofch computation for example with the the chain link article not infrastructure you could leverage uh computation across rollups you could do sharding of your compute of your of your data across different kind of layers of uh of the stack uh so you have all these options available that were not available uh probably one two or three years ago obviously you have wallet abstraction as well which is which is a whole another concept so so all these optionality what that means is it it's allowing you to kind of use that those Lego blocks and build really um application the right application that that serves your needs um um and and provides the value that you're looking for it to provide so if you think from that perspective I I feel um you can start from what is the user that you're looking to serve what are their core needs from a business perspective and from that you can start putting the Lego blocks in place um and then figure out okay what are the business logic pieces that I need to focus on to uh build that dap and then obviously audit audit audit test test test um uh and then bring it to Market I love it starting with the user we often times start with a blockchain which blockchain is going to give me more tokens we should start with the user um Paul for you since you're coming from a centralized provider maybe you can shed the light on does everything need to be decentralized or what you know what is okay to be centralized yeah absolutely and I think that gets to sort of this this these tradeoffs in this decision tree it's not sort of a static thing right it's not like you okay you make your decisions and you're done and you're going to succeed forever you know have your journey Forever on out it's a constant thing that you're you're evaluating ating and so at the at the beginning right like like sh said is you you you work back from your customer and you understand well what is the most important thing about my my proposition to solve customer needs that I need to test out to see if I going to get product Market fit right like if if customers are going to tell me that they really need this and so from that perspective you kind of focus on the things that are important so if it's important that um that I have self sovereign in custody so like I I need to have a self- custody wallet right that's really important to my use case and I'm catering to a certain customer base that needs that then that's something that you need to have decentralized but does it mean you need to use a decentralized data infrastructure to do your your analysis and your metrics and your reporting or a decentralized storage layer to to to to to store actual data for your business logic right and and and and and and no not necessarily to get started maybe there's a vision where you want to move some of those things to that way and that's where you know using man service at AWS for example can get you going faster right like we've done some of that a lot of that heavy lift and and and and some of that heavy lift for forever can be decentralized like you know I I I kind of tend to think um you know I think there's a vision certainly that that Igan leer has in terms of creating a decentralized infrastructure ecosystem and and that's super interesting and I I think you know AWS has a role in that we have to change how we offer our services we have to integrate with a different sort of Commerce layer and control plane but S3 absolutely with the standard interface on top could participate in a decentralized infrastructure economy um you know at the end of the day um but what I think really at least for the mid and foreseeable term I feel like there's a couple of things that where decentralization is really hitting what is bubbled up to to to to sort of drive this um web 3 um evangelization which is ownership value exchange and identity and those are key things where you know them being tied up in the silos of the big tech fch companies is where that um frustration has come about in terms of all of the value of my data is C occuring to those companies and I don't have control over my ability to transact my ability to be in control of my assets and ID identity um and and I think focusing on decentralization of those things and and and not NE necessarily trying to take one Leap to the decentralization of everything I think will help us find the path to that product Market Fit For A lot of these apps so ownership value exchange and identity yeah those are the three properties yeah and uh censorship resistance I want to make sure that you you are you are the so you um I I didn't user sovereignty data ownership sensorship resistance I don't want all my all my data all my my computer all my everything to be with one controlled by one entity who that could be a government that could be a company I don't I don't know but I would actually answer the question from why can't I create a decentralized app why can't I so we are working towards the goal of this question doesn't even arise like you have all the tools available and you're not this dilemma that we discussed I'm not worried about my I'm not worried about my performance and my up time because I can select without having any intermediary in place I can see all the data that's being published either on chain or or you know on chain or or with the with with the decentralized infrastructure Services I can access that information openly I know that there is actually a massive network of operators who are running this underlying infrastructure I know that there is no single control from anybody and I know that there are some high value Premier tier operators who are underpinning this particular service not just one 20 of them 30 of them 3,000 of them in the future so this question doesn't even arise like I when I was at Facebook I was like why can't I have something like farcaster why can't I have something in lens it should it should be like that so our go of L layer is to open up the world of offchain compute access the entire cryptoeconomic security of the 100 plus billion dollars of ethereum stop this fragmentation that's happening in this you know different liquidity pools across the different blockchain networks let's band together create a base substrate yeah where I can actually access this security and and provide the same economic Primitives to every single ABS to every single actively vated service and all of them interoperate on the same basis so you just focus on building your actual service and don't worry about these bootstrapping networks and stuff so the question of why can't we have decentralized Facebook is opening a new kind of can of worms that we don't have time for perhaps another time but I would like to go since we have about a minute and a half left with a quick lightning Round And if we have time after that I'll ask one more question so lightning round quick question yes or no answer uh or a very short answer afterwards um first thing that comes to mind first question do we have too much infrastructure in the web 3 space no not should I explain quickly we don't have the right not enough of the right INF structure not the right Primitives what's missing off expand offchain compute access same the same underlying security Primitives and build on a common substrate and then you will see the magic happen Paul no not enough there's a lot but we need that so that we can get the parallel Innovation to find the right product Market fit find the killer apps that are going to get the next you know million 100 million billion users yeah same not enough you know with every boom and bus cycle the infrastructure is improving uh and that is really helping to make D5 more secure so that more value can flow in because I think that's what everybody wants so basically don't build a house of cards build a solid infrastructure service L yeah that's a controversial answer most would say that we have too much infrastructure next question what will be the next most important primitive for 2024 2025 and you cannot say the name of your company or your space so so so we I exists not for for ourselves we are here to create this network of abss network of Builders including things like AI for example so so AI is in everybody's lives right now and AI really really needs like companies like uh ritual for example doing some amazing work on bringing AI to the masses in a decentralized manner on igon layer so AI is the answer what about I think I think making Ai and and what is important for the for AI is is the actual data that trains the AI models I think that's where I think decentralization and I layer can play a pretty big role in that that's a good one I'd say identity might be a little further out but I identity I think is the yeah um identity privacy the whole concept of ZK because it all goes back to Bringing more value and for trafi to feel safer uh in transacting on defi I think you need uh you need that identity last question we're out of time but do you think by 2030 Amazon is going to launch its own token yes or no or abstain abstain I'll give you my PR person's name all right without further Ado I'd like to give a round of applause to this amazing panel