Vaclav Pavlin - Let's put things off chain

Hello. Uh thanks for coming. It's amazing that I don't have to speak to an empty room which I was kind of expecting as uh being a first speaker other than the massive uh opening keynote. Um so my talk as uh Z said is let's put things offchain and I would like to tell you a story about how there is more to web 3 than just a blockchain because now and the at the opening presentation you also heard that like we want to talk a lot about blockchain and I don't well not not completely ignore blockchain but I want to talk about other things as well. So my name is Vata Pavin and I work at a company called Institute of Free Technology.

uh there is a project called Logos and under that project there is a project called Waku. We can talk about these more and I'll talk about them in the presentation as an examples of uh these off-chain peer-to-peer technologies. Uh my official role is not a chaos agent but that's how I'm perceived usually. I just go somewhere, break things and then leave. Um but I'm a solution engineer for Waku.

So I help I should help people understand how these things work. Uh so if if if you don't understand it after this presentation I'm not good at my job. Don't tell my boss please. Um can I get next slide please because uh clicker doesn't work. This is you this is done in files.

If you're friends with file um clicker doesn't work. Tell them. I told them uh they need to hear it from more people to fix it. Um so what is this talk about? I'll talk about history of web 3.

I'll talk about what I perceive as a pre present state of web 3 and what I think could or should be the future of web 3. Next slide. Um, so the history, once upon a time, I walked into a blockchain/crypto telegram group and it was full of people that were trading shitcoins obviously. Um, and I was not really interested in that, but I was curious how it looks inside. Um, and they started asking if there is a developer around so that they could look at the smart contracts and so that they don't get wrecked.

And I'm like, I'm a developer, but I have no clue about solidity. Um, and they said, well, you're a developer. you can learn. I'm like probably. So I learned solidity and I started doing this kind of auditing for them.

That's probably four three four years ago. Um and that was a lot of fun. Uh don't get me wrong. Next slide. Um and I have learned or I have twisted my brain into thinking everything has to go on chain.

It's going to be so awesome. Like we will live on chain. Everything is going to be on chain. DeFi, decentralized finance, games, anything to earn. You remember step to earn or step in or like walk to earn.

I saw a wape wave to earn recently. Um I think there was something like masturbate to earn as well. So like uh there is anything to earn. Uh that's why the asterisk art will be on chain. Our social graphs and social networks will be on chain.

It has to be also is it on this chain or on onchain or on space chain? Do you have an opinion on that? Uh I don't but just like nobody knows really. So that was the idea like let's put everything on chain. Everything has to live on blockchain.

Next slide. And this was also um this was also kind of um what's the word uh uh emphasized for me through an article that I read which uh yeah it's not this slide but that doesn't matter. um called Hyperstructures. It was put up by somebody from Zora, probably familiar with Zora protocol and Zora marketplace and now chain and all that stuff. But the guy from Zora put up a uh article about hypersstructures and hypersstructure is kind of like how you think about blockchain.

It is something that is just there and you just deploy there and it's there. It works like you don't have to pay for infrastructure. it exists forever for as long as the blockchain is there and the application lives without your kind of need to maintain it to uh put it in front of people to pay AWS for the bills and all that stuff. So that's kind of the the definition of hypersstructure there and that's just like amazing architectural shift for me but then also that tree the history of the tree wanted us to build something else right so it was not about putting everything on chain and doing everything on chain and every single thing that we do does not have to go on chain because initially this is this is uh I took this from the blog post uh that vitalik put out a couple maybe it's even years or months uh ago called Um let's make Ethereum cippy punk again. And this is from the original uh Ethereum white paper where Ethereum is not just Ethereum blockchain but Ethereum was supposed to be three things, three pillars.

It's supposed to be consensus so the compute of the world storage which was swarm I mean still exists but it's not part of Ethereum really that much. Uh so storage decentralized truly decentralized durable storage where you can upload stuff and it's just going to be there for as long as you need it to be there. And then communication which was called whisper uh which no longer exists. And like that was for for the cases where if you want to build a deb uh a decentralized application these days you would probably go to uh AWS and deploy there and then you every single interaction goes through those servers kind of right there is some backend there and then it goes to blockchain. U but we should not have that.

It should be there should be a peer-to-peer network that that just works and we communicate directly. That's what this whole narrative was about initially. The bed three, the three pillars of the future of the internet. Uh, next slide please. So from my perspective, a truly decentralized application needs exactly what was shown there.

So there has to be communication. We need components and users to be able to exchange the state of an application of or or some interaction quickly, permissionlessly, and in a way that cannot be censored. We also need to store the data because some some some state cannot be just communicated between people. It needs to be stored somewhere off the devices. Maybe it's a backup to to prevent loss or just to prevent just to improve access to the data.

Right? Uh if I have to ask every one of you to send me your latest change to some application, it will be slow. But if we can all collectively put it somewhere um that is cheap and fast then then we can we can have that scalability and then registry which is for me the consensus the blockchain the ledger uh some pieces of the information need to be persisted like with the data storage but they are smaller and they get updated often or quickly and there has to be consensus about this. So I have if I change something there has to be a trace and provenence of who updated after me and before me and that's what blockchains are for and then encryption uh every decentralized app needs a good encryption and but it's not just that end to end encrypted between two users but everything has to have some kind of uh keys and they need to manage and the access has to be controlled to the data and all these things. So it's not very straightforward.

But if you look at current state of Ethereum, that's not where we are, right? We have everything publicly on the blockchain and then bunch of L2s that are cheaper and faster, right? So next slide, please, if I have the person there. So uh this is how you can build a decentralized application and it's I am very biased. This is there's a disclaimer.

I'm biased of this stack because that's what I work with. That's what I develop and I'm paid to do that. uh as well. So that's why I talk about these things. But maybe there are better and other options for these pillars of bet free and you are free to choose whatever you want and you're free to contribute and change these and you're also free to disagree with me.

Right? I am happy to argue about about these things with you. But basically what I want to show is the original picture of the web tree with the three pillars. We have them here as well. So Nomos as a consensus, Codex as a storage and waku as communication.

So and this is logos. This is kind of the umbrella of these uh trying to put to the front the idea of network states how we think they should be done. Uh can I get the next slide please? Uh but that's the future version of the present. The current version is different because the nomos part the blockchain is not done yet and then also it's missing the encryption.

So I've created this new version uh sorry this this more current version where we have status network which is a a a gasless layer 2 that we are going to launch next year. It's in test net right now. Um so that's going to work there. Uh and this is a logo of taco which is a threshold access control. Um very cool project.

Um I'm I'm not biased because I'm not affiliated with them. This is unbiased opinion. um very cool project for access control in in general but also very much uh in bet three uh and you can use it with storage or you can use it with communication and I'll explain more of these so uh can I get next slide I don't remember what is what is there sorry I forgot this uh this calling the next slide uh is is hard so um so right now the the challenges or one of the challenges that I perceive in web3 space and that's kind of uh another part of my pet peeve with how we do things in battery. We have trained the next generation of battery developers on accessing a peer-to-peer network through a JSON RPC endpoint. So like a rest API or like an HTTP endpoint.

What does that to people is that they assume that there is a database that they can query for some state. But that's not how peer-to-peer networks work. you connect to a peer-to-peer network and then you can integrate interact with other peers in that network and there is usually not a single source of truth uh and you need to kind of do a lot of things locally but we have abstracted blockchain through infura and other node providers that there is a node that knows it all right and that's not that's not usually the case so that's I think very uh um very problematic and that's also what I am trying to explain in the future slides so uh next slide So let's take a look at an DAB that I have built as an example. Uh and it's a DAB with the capital D where the D actually is decentralized. Um and there is no central server and there is no there is no um kind of a way to prevent you from using the app.

So next slide please. Um so the application is called Quaku. It's a Q&A. So you were supposed to scan the Q QR code to ask questions. This is very similar type of app.

Uh you can go to quaku.app. I was I wanted to show you a demo but apparently I did not read the emails right so I can't show you the demo. Um but normally I would click on this and there will be a UI where you could go and create a Q&A. So I'll just talk about it.

Uh can I get next slide? So this is how it will look like because this is a new UI that uh that we made and uh there is still an old ugly UI that I made uh on that URL. So don't don't get confused. But basically the idea is very simple. You create a Q&A uh when people can ask questions.

You have a unique link where you can go and it will pull some historical questions and up votes and all these states all these information and display the state. Now how it works is that you connect to a waku network. WKU is successor of whisper in that in that diagram of the three projects the the three pillars communication storage and consensus that's the communication. So you open the browser and it connects to waku network and if there are other people using quaku u the Q&A app they will basically talk to each other directly over the network rather than going through some centralized server. And wu is a peer-to-peer network.

it uh just cares about sending bytes around. It does not care about what the bytes are. There are some some uh features that prevent spam that prevent that make sure that you can disconnect and connect back and query some history and all these things, but I'm not going to go too deep too deep into the into those details. So that's one part. We have Waku as a transport layer for this application where you can communicate directly basically with other participants in the Q&A rather than going through some centralized server and database.

Second part of that is codeex which is a decentralized storage. Okay, I have eight minutes. That's what you're showing me. Perfect. Thank you.

Uh I was not sure it's quite quite small. Um so um codeex is a decentralized storage. So Waku is ephemeral. Vaku will you will send a message it will be there in the network for a while but then it will disappear. Codeex is a persistent storage.

So if I want to access some old Q&A it has to be in a persistent storage because the ephemeral storage will lose it. So I will push the state I will snapshot the state and push it to codeex and then basically just announce the ID CD of that state in the application so that the future users can come and see that CD and load it from the storage network. So that's the second pillar and then we have third pillar. So for example, we have connect wallet here. It's not very well visible but uh connect wallet.

So you can be you can uh reclaim your identity through like ENS or potentially other identity systems if you want to show who you are or you can stay completely anonymous and that's kind of where the blockchain comes in because blockchains are pretty good for identity. You can have ENS, you can have uh zero knowledge proof of of humanity uh like Rimo is working on and all these things and then when you combine that you can have some trust networks like uh circles and we combine that you can have some kind of unique identity that you can use uh for your applications. Uh next slide please. Um so this is the architecture. I have uh decided to do a very professional architecture diagram.

Um, so this is Quaku. Uh, obviously, um, there is some swag in the bag when you're when you will be leaving. There are some like Quaku earrings and coasters and stuff like that. So, please take them. This is codeex as the storage node.

Uh, and this is Waku uh the snowflake kind of shaped thing. Um, and how it works is that a user accesses Quaku and it can be hosted somewhere uh on Versel or whatever is the hosting platform you want to use or you can just run it locally in your laptop and then you communicate over Vaku with other users and you you kind of collect the state locally. uh and if you if you what I described if you want to uh snapshot the state you can push it to codeex and then they can communicate directly or there is a caching codeex node um provided which can make sure that the data stay around longer and are accessible to other people connecting in the future. Next slide please. So that is basically how I think the decentralized application should be developed with actual peer-to-peer networks in mind.

Uh but there is more kind of uh so what I think about future is that the future has to be local first. We have started local with the origins of internet like everything was running on your laptop sorry probably not on your laptop when the starts of internet but on some machine under your desk on or on your desk and then we progressively went towards running everything on someone else's computer called cloud. Um but now we are realizing that maybe that was not the best thing for privacy and for for censorship resistant. We are pushing back to going back to local first going back to our computers doing the actual compute. We also have much more powerful computers.

So we should be we should be very easy to do it unless all the compute is consumed by browsers and stuff like that. So we should focus on local first. We can if you Google it up there is a website about local first. There is conference there are conferences and all these things. Now, future is peer-to-peer in my mind.

And it's not just about DeFi and like doing peer-to-peer transactions, but it's about every single thing I do online, I should have it locally, and then I should just exchange with people that care that I want to show this to uh what they need to see and that's where privacy comes in as well. The future is encrypted as well. Uh we know cippher punks, we're all about encryption and that that is how it solves a lot of privacy issues, but it's not the only thing. The future should be permissionless. Nobody should be able to prevent you from using infrastructure and from using services and applications, right?

If the application we we we recorded a podcast with Alina uh today and big chunk of that was discussing like so if I upload data to Dropbox, is it my data or is it Dropbox's data or can Dropbox prevent me from downloading it back? Right? I mean they can obviously so then it's not permissionless and it's not censorship resistant and the future should be permissionless and it should be censorship resistant. So you should have control over your data and over accessing the infrastructure. Future should be private by default.

We talk about privacy and I realize that like nobody cares from the perspective of nobody knows what it means. If you say like you need privacy like it was very hard to imagine what does it actually mean like it's it's intangible. If I tell you you want to use censorship resistant software, you'll be like, "Oh yeah, I probably do want to use software that cannot prevent me from using it, right?" Like if I post on Facebook, I want to post on Facebook. If I want if I post on Facebook, I want others to see my post.

And if they can't see it, if Facebook tells them no, then it's not it's it's censorship, right? It's not censorship resistant. So, it's much more tangible for people. If I say you should you should have privacy, they'd be like, "Yeah, sure. Uh but what about the bad guys?

Uh yeah, well they also want privacy but from different reasons. But it's very hard to grasp why it's important and the future should be provable now with like all the ZK stuff, zero knowledge uh zero knowledge technologies being pushed to the front and also being much more efficient. Uh we should be able to prove a lot of information and a lot of state about ourselves. Yeah, I think I have one one last slide. Uh so so actually two slides um you can go to forum.

logos.co where we post our work and our ideas and our stuff and our discussions. So please go and take a look at what we have built and please let us know what you want to see built. If you want to see something built uh that should be decentralized or if you have an application and want to discuss with us like how can you decentralize it further and better please come talk to us. And next slide has actually a QR code because we love QR codes.

Uh and that's for our discord. I know I know we should not be using discord. Discord is not decentralized and permissionless. We're working on that. Uh but right now that's where we are.

Uh so please join our discord. It would be amazing to have a chat with you if you have thoughts about this presentation in our discord or anywhere else. Uh just find me. Uh and uh yeah that's it. If you have questions I'll maybe answer them.

[Applause] Do you see Quaku as a model for other big D apps? Uh what kinds of applications are best suited for this architecture? Uh very nice question. Uh yes, I do see Quaku as a model. We have actually we have actually uh modeled some changes in other applications like uh if you heard about status uh as a messenger and a wallet that it uses waku as a communication layer but if you set if you send a big file then it's hard to like propagate it over waku because waku is not designed for large data.

Uh so we're basically using the similar model for quaku uh uh for qu sorry for status as in quaku to offload stuff to codeex and then and then send the c around but there is there is more what kinds of applications um we are also experimenting with like something like google forums something like a 4chan uh I've built very very simplified p for like a streaming platform on top of wacu codex uh it's all on that forum link that I've showed. Um so there is a lot of applications that you can build in this with this infrastructure. Um top three things that should absolutely not put on chain go. Um porn um I think that's one uh I actually I was actually part of a porn based porn related project and nobody bought any any of those things. Uh and I think it was because of the privacy issues of blockchain because nobody wanted to be seen buying porn NFDs.

Um so what should not be put on chain? Um well ideally your financial stuff uh shouldn't be publicly visible on chain or not the case right now. Um big data I don't think that big data should be put on chain because we don't want every single known having any big data. Uh we want to keep that small and succinct. Um yeah I can't think about third thing.

I'll think about it later. And then usually fully decentralized app has worse UX compared to more centralized ones. Do you think regular users care about the centralization enough or so you think the value they value UX way more? I don't think that we should compromise on UX. Uh I think it's uh obviously always there is a trade-off.

Uh there are always trade-offs, but I think we are relatively close to be able to go with reasonably good UX for decentralized apps. And the question, do regular users care about decentralization? I have a problem with the word decentralization. I use it a lot, but I have a problem with that because it means different things to different people. And so what I talk about is permissionless, censorship resistant, and private.

And I think that that's what people should care about. [Applause]