Building the decentralized datalake with Fabian Riewe of KYVE

Transcript

we are live my name is Sebastian cuccio and this is the interop this is a show where we dive deep into the decentralized economic networks that make up the interchain I'm here today with Fabian Reva who is the CEO and co-founder of kive kive is a decentralized data Lake protocol we're going to get in all of that and understand what it means uh how it integrates with the cosmos ecosystem and Beyond and also we're going to dive deep into our weave hi Fabian thanks for joining me well thanks for having me we have quite some agenda for the day I see yeah it's going to be interesting because I mean like our weave is something that I kind of you know passively pay attention to I've got lots of friends who are like just

all about our weave we've also have some content here on the show before about are we but um yeah I'm still like struggling to understand how it all fits into Cosmos I think you guys are probably the project that you know are in the cosmos but are also like mostly uh close to our weave uh before we get started though if you like the videos that we're producing if you like this content please make sure to subscribe hit the like button hit the notification Bell because we usually do live streams every Thursday around this time so yeah uh tell me what about your background and you know how you came to be working on uh on Kai absolutely it was it's a very organic story so to say so I'm a software developer um and I was just working as a

freelancer um and then got invited by the sap which is a big German um development company I think it's pretty International actually it's and this was the first time I was in contact with with crypto like some dude would introduce me there I don't know how he got into crypto but um that said and we said you know what for the second one let's build our own um years of 2012. this was like 2019 yeah 2019. um we're like okay cool two hours later we had an erc22 and we're like okay it's a three-day hacker so what we're gonna do with it now right like okay and so we looked a bit deeper into all the crypto stuff but you couldn't really see use case and so on and then I was chatting to the guy next to me and he was an sap intern and he was working on the CLA

assistant and the CLA assistant is an open source tool from the sap used by like Amazon PayPal big Enterprises that they have the problem that when they have external developers doing contributions to the open source code that they have to then sign online an agreement a contributorial license agreement CLA that says that basically they're giving away the rights on the code that they are contributing to the project and they were actually storing um those signatures on like some small hard drive uh sitting on the fridge in like the office Department because it was like a work in progress tool for them I hope it's fixed by now but I guess that's what I was supposed to do was working on I got back to the hotel and I was thinking about actually that's a perfect use case for blockchain right

like if I would store like the proof of signature there that no one could ever say actually no I have never signed the CLA because it's all very happened on GitHub right I don't think any big legal substance obviously so I thought there might be a cool like use case and then the other two days of the hackathon I developed that and always thought well nice proof of concept it was deployed to well ethereum test net even back then gas was already expensive um and then uh yeah that that was my crypto journey and then have you later uh the dude from the sap hit me up and said hey have a look at are we if it might be super cool for your signature use case basically you can build a decentralized DocuSign I was like oh actually it's very cool because now what

Riff allows me compared to other blockchains so that I can upload much larger files so I could actually you know attach the real contract and then let this one basically be signed by my party or my counterparty whatever so that's that's nice so reach out to Sam the CEO of our weave um so we just chat about him what he thinks about it and so on and then we actually both kind of got turned into this Rabbit Hole of actually like how do we even know if an address is a real human how do we know if my counterpart is legit because it's very cryptic it's not very user friendly we're like okay that's that's interesting and then this was my first ever like big crypto project called a verify um it's fully based on our weave and basically we are analyzing the rif traffic and by this are able to compute

a score for net risk seeing if it's most likely human or not so think about it a bit like on chain kyc but it's not through an external like kyc provider um and then this happened this was then um late 2020 very early 2021 um and I was reaching out on the Discord was Community projects hey does someone want to help me build this and then John um who's now the co-founder of kaiif and reached out to me and said hey actually yeah I can help you I would love to work on this I'll say awesome that's great so many on Discord and then we started working on our verify um launched it in December um but actually quite some some good revenue on it it was like a little tipping system and this worked out super nicely and then we're just browsing through Bitcoin one day and saw a bounty

from parity that's a development company right behind polka dot um and our weave and it was a shared Bounty about bridging polkadot data to our with seven and a half k or something like oh that's easy easy weekend money let's do it so we did it and I got some amazing feedback for it and then we're like oh actually that's quite a system that we can scale now right there's this archiving kind of solution I can go much deeper into it maybe later why this is such a cool thing um and then from there on like Kai for spawn right like we started integrating more and more chains uh the team got bigger and bigger um I think I went around 16 to 20 people now now switched over to Cosmos and everything happy to talk about this uh this more in a minute and yeah well now

I'm here I would say uh this is how the whole types thing happened in my backstory cool and how long ago did you start working on kive and sort of what's been the Arc of you know the project like I don't think you guys are live yet you have an incentivized testnet but you're going to be live like shortly yeah exactly exactly so maybe let me just now step back and talk a bit about what kive actually is so with Calif we're building a fully decentralized data Lake and this means basically that we are getting data from literally any data source blockchains off-chain data financial data whatever into this data Lake and then make this data accessible for

developers right think about it as like an oracle on steroids um so it's a little in Oracle for for any type of way validating it it's sure it's correct and this idea fully started out um on on our weave and working with are we smart contracts um so they were pretty new back then and weren't production ready to the extent that we needed them we needed like contract to contract calls that wasn't supported um and so on and in general like just solidity was more of a standard that we needed to to work and to scale with and uh yeah so we switched away from using a full-on RV stack to a solidity contract are with storage Stacks so we're storing we still do by the way store all the data on our weave and all the logic is handed by or was handled by a smart

contract then we realized oh actually also evm also gives doesn't give us the flexibility from a tech perspective but because you always have to share your transactions like the transaction space with other projects right it was bull market so gas prices went up the roof and we're like oh no this is not what we want we want something where we have our own space and we are sure that like Kaif in itself stays economical and this is when we looked at either this was February then February 2022 so eight months ago um we were like okay should we now switch to Cosmos or to polka dot um and we as personally we weren't big fans of the polkadot like shared security um system and we thought well if we're gonna build something that's independent

we want to have something truly independent right I mean this is 10 and how we decided to go with Cosmos and no one on the team know neither rust or gold so this was a pure decision of kind of like the philosophy behind it and this is um yeah how that went and then I think in April or something when released that we switched to Cosmos deployed that had the incentivized testing going on now for quite a while 40 000 participants absolutely insane um yeah and they're now really just yeah going towards mainnet in big big steps and we're very excited for it and yeah I really cannot wait to to launch it um so yeah that's I would say the the current status quo cool well I I'd like to talk about permanent storage

um because I I think it's one of the most interesting use cases in crypto so you know on epicenter we've interviewed like tons of projects over the years that wanted to do either decentralized storage or permanent storage and I've always felt like this was the thing that got me kind of really excited about crypto um in terms of non-d5 use cases and you know you know I I was really impressed by projects like Saya even back in the day like uh store storage you know these projects but then our weave came along and they had this totally new um model and this this this new kind of proposition of not just decentralized storage but permanent storage

um I I have yet to really have confidence I think in the tokenomics or fully understands the tokenomics to to be able to have confidence in the fact that like this storage is actually permanent um so you know I have friends who you know just kind of are big are we fans and think it's super cool but I'd like you to maybe explain in your view like what what about our weave makes it uh so interesting and you know how does it ensure its permanence because when you say permanence like that's that you're talking in absolute terms here you're really talking about like something that will exist forever there's you know it's not a binary thing uh or it is it is forever or it doesn't and like to have

absolute certainty about this permanence is like a big statement so yeah how do you describe how it works and then also the tokenomics that make sure that this these this the storage remains permanent sure so I'm happy to break it down because that's actually also like a big reason why Kaif exists right and I think also gives like a good base on like how Kai really works basically the problem you have right now on blockchains like on the traditional once right ethereum Solana and so on is that the pure incentive the Toca incentive is set on block production right and this means of course that you only get rewards for actually creating blogs but not for keeping the data I mean this is what's now happening for example to Solana is a

famous example where a lot of nodes are pruning the data mean that they're leading it out of the database just because it's more flexible and of course you maximize your profit by investing um well the least in the incentive to to not storing data because they're really not incentivized through that um and then alif has the complete opposite so RF is highly incentivized on um well data access and that accessibility so to say or data permanency I really don't know what the correct term for that is but basically the way they've designed it to economics is um basically by by saying listen we're gonna give you tokens for the exchange of data right wide and the more

rare data is the more tones you're gonna get for discovering that and keeping this in your network which is a very smart approach right and then they also say listen we're not a blockchain where every node has to have a full history where your block weave that's the name right R weave um where actually one node can have as little data as it wants but of course maximizes its profit by having all of the weave like all of the data right because then they can always retrieve any data meaning they get the most rewards for that and I think if you understand this concept then it makes more sense why you would say okay this is why it's permanent storage and more of a block weave and then a data blockchain and the reason why it's permanent that's then a more like

in-depth design choice of our weave and actually I'm a big fan of also challenging are we for permanent structure but actually saying there is no need for every piece of data to be stored permanently right accounting data is a very good example I think 10 years in Germany at least where I'm from law it is mandatory and an app that you can absolutely discard it because why wasting like money basically on something that we don't need to pay for and then rvf has the easiest thing just says listen you're just going to pay for 200 years of storage in advance right and they have to storage Paradigm and they say okay the cost of storage is decreasing by I think it's 30 every year right and of course by this you can just draft a mathematical curve right and

then you can follow this you can compute how much you would have to pay for storage now in order for to ensure that the node always has enough money to pay it for at least 200 years that's how they they just how they justify the the permanency yeah with regards to the to the tokenomics then so the the idea is that you're incentivized as a node to store data that exists in scarce uh and sort of scarce amounts in the network so if there's like data that is on our weave that is not um you know replicated to some to some Factor uh there's going to be an incentive for you to pick up that data and replicate it on your node

um the the thing I think that with our weave that I've always had a hard time kind of grasping and getting coping with is is this idea that that you mentioned that like all data should be stored forever and as far as I understand it or at least when I was you know when we interviewed Sam on epicenter I think like a year and a half or two years ago there wasn't a way on our weave to effectively delete information like all information on our weave was to remain permanent forever is that still the case and how then you deal with this idea that like not everything needs to be stored forever because okay if you go to some extreme like you can go to extreme like very granular kind of data storage um policies for any sort of application

right you can like store every single you know finite piece of transition transitional information uh in state change and you just have like an infinitely large data set yeah exactly so actually um I don't think there is a way or now with to ensure it's not permanent right um I think that's actually something where other storage solutions kind of look like might there might have their unique selling point right and I think it's a big question of which one's actually winning over the other right um and also like I mean for kive actually we are storage backend agnostics what we like to say meaning that it is possible for those storage pools that are like relaying the data they can relate to other backends they

don't have to use are we if they can use like falcon ipfs makina or near or whatever right and we just as a team I mean because we came out of our we fight that's our biggest expertise and so of course we are staying on there and pushing that but we do absolutely see like a viable point that it's not worth every piece of data to be stored um well permanently right um and also it comes out with a big a lot of responsibility from an ethnic ethical point of view for storing your data permanently right because what might be okay now might not be okay in 50 years 100 years 200 years right probably it means a lot like yeah right and coming from Germany like you know

you're well aware of like this whole right to be forgotten uh you know regulation that uh I think was um targeting Google you know several years ago and is still very much in the in the say German mindset right that like no information should be permanent a lot of people still pay in cash in Germany and uh and for you know for reasons that are very cultural and so yeah there is there is I think also like cultural ramifications to this um okay so so there so are we still stores everything in Forever uh yes yes exactly so I know there is I don't know if if it is in place by now but I know that in the theory notes can agree on uh not storing any

um transactions anymore so let's say you upload some illegal files to Avi right the nodes could agree on excluding this the transactions from the data exchange and then over time you know the data would extinct there but then my big like thesis on this is is if it's illegal data it's usually always has some high value attached to it right yeah um then there might be a subnet of are we force sub Fork of RV or something where you don't where they just don't use this Blacklist yeah so you will keep this data around it would be much rare to find but actually then you could write algorithms to find it like you could also encrypt the data yeah right and cryptidius is one big problem in general right like it's encrypted here whatever

you cannot do anything against it right but if it would be open data then I would guess even then like something would in that reaching with form right so it's it's questionable on how this box works yeah how does our weave deal with Spam then because like you know yeah I mean I guess you also have to pay to post data on our week but like if you had some like I don't know what the what the uh capacity of our weave is currently but let's say you had like some State actor that just wanted to spam are we for whatever reason um and had you know unlimited money I mean this is the same I think with like you know Bitcoin or any other sort of you know like yeah but decentralized Network that has some amount of scaling

issues um you know you could just Spam are we even just post like tons and tons of data to it making it practically unusable and maybe even economically unfeasible over time yeah I I can tell you it was us it was Kaif in in the early days that did this like like we outskate our we've um at one day I remember like I remember a message from Sam we were in development conflicts like hey guys I think you guys just killed our Gateway and we're like oh okay and so we're like okay no worries um and then basically so they did some optimization of course right um we did some optimization um as well and then um Bunger Butler came along in the nft hub from Josh we

are very close with I'm with him um and what Butler is I'm happy to give a quick tea on the art it's basically um like a scaling service for RV so what they do they they take all your your data um but with foreign data like basically it's like a roll up for for they take all the data from others and they just ensure you that at one point in time it's settled or not leave I mean this really takes off some load and now they drive actually 95 I think of how is traffic wow which is crazy and on the car side we've dealt with this by also bonding the data um custom like basically on our side so we're pushing much less data to our we've then we did in the early days especially the nft I really took down now before a while but

it was never Network it was always the Gateway as far as I'm aware uh which is the next thing service on top of our belief and this was the one that uh yeah didn't didn't like all that traffic as far as I were so before we get into to kive I I do want to talk about are we smart contracts briefly yeah like how does that work and how is it different from say the Paradigm of smart contracts that we're used to with ethereum or like Cosmos and like these kind of on-chain uh State transition um contracts yeah yeah I um I know I I must openly uh admit now I don't think I ever did that I'm not a not a big fan of the obvious smart contracts um because of the way the average smart

contracts are designed um it's called client-side execution right which means that there is you have the transaction happening like I'm transferring right 10 talks to you you're transferring 10 talks to me and then usually in a normal Network like a normal Network in on ethereum right for example the nodes they would do the updating State calculation right and always store in the latest state so it's fast and easy to access um and you find consensus about it right while this is now the big problem on our weave um your client does meaning that client is well your computer right so your computer actually has to download all of the transaction that that have happened and for 10 protection is the problem but for a million or a billion it gets quite exhaustive right and then you have to

compute it basically to the latest tip which also takes some time and computation always takes some time and even the the fast implementation would slow it down for an indefinitely uh large right number of transactions um and that's actually one of the reasons why we switched away from it and the other reason being that there is no consensus on a state so if if my client might be corrupted or I might not read some data and I end up with I know I have 100 tokens but when you said it says 101 tones who's correct right we don't know yeah right I mean you have to go back and like someone would have to go back and look at all the state transitions that have been posted are we even then I guess that would be the source of Truth yeah but

but it gets even more well funnier because then well you have different teams developing different clients and uh the clients had different um ways of interpreting um some data and this really completely demolish the state so like what one story my favorite story there is actually that we were using the native smartwave client and then the Redstone team was building their client um and basically what happened then is that we were executing the exact same contract with the Redstone client but I got a completely different result completely different we're like okay why did this happen and we invested investigated and we found out that very early on um an investor unlock failed on either our or on the other client and this investor unlock of course right

then uh influenced a weight on the vote on the governance that was changing the parameters of the network so it's really like this complete chaos after this right because you have different parameters suddenly your different weight shifts and so on it just got worse and worse and worse right yeah how do you unbundle that exactly and then well then the solution from the Redstone team is and uh is to build this like warp uh I think they also have the protocol warp right it's like their smart contract Network yeah it is pretty centralized um if not basically centralized because they have a caching service and then it's not trustless anymore right like how can I trust someone teams cash that's not like the blockchain spirit and not how it should

should work right um and that's actually one of many reasons why why we switched away another one is also I think that a lot of people have to keep in mind is this it's called foreign core protocols actually it was a brainstorming idea by Sam Tate and John our co-founder we just said okay how can we make contract to contract calls possible um and because everything is stateless um you have to do this weird concept where basically um if I would like to interact with you right I would basically have to hold up a sign on my side saying hey if you read this please execute this on your end right and then you would read this and say okay Fabian says I don't need to execute this but then you have two

problems one is the touring problem that means that you don't know if I may if I might make you run into an infinite Loop right and then imagine your billions of transactions in and suddenly there's some bug and I'm able to call that can exploit call it into the loop or yours and the contract is broken forever there's no way to fix it right it's broken forever which is super super dangerous that's the reason why Heath introduced guests right to solve the touring problem because they said listen at some amount of computation we're going to abort it and don't make the transaction count because something is super off here right always doesn't have this and they it's a bit of a yeah they say it's it's gasless but actually it's dangerous that it's gasless in my personal opinion right um and then the other problem with that

is you have an ever-growing State because me on my side I don't know when you have read the message you are supposed to read because it's stateless right yeah so yeah I basically I always have like to duplicate myself and create a new family and holding up a sign hey next time with this hey next time do this and this just explodes the state and now we mentioned like lots of smart contracts interacting with each other and you have suddenly no clue what's going on it's just not scalable at all are you familiar with orbit at all uh no what's that well uh so yeah Urban is this um is this kind of decentralized compute system

um I wouldn't know how to describe it uh my co-founder at epicenter is a big fan of it and and um and basically everybody has kind of like their own computer right like their own kind of virtual machine and then things settle when you communicate with others in this peer-to-peer Network so I don't know if you're familiar with secure scuttlebutt no no okay yeah this is like this is like deep in the weeks kind of stuff he's like pressing my share um yeah but yeah basically it's kind of like this stateless compute uh thing where there isn't actually a network there's just like kind of data structures that settle with other data structures welcome butchering this

explanation but it sounds kind of similar and yeah I need to do a little bit more research about orbit but um so yeah what's what's the um what's the interplay and the interaction between say like like our weave and the cosmos ecosystem and you know because I think you know being in the cosmos ecosystem I think there's I don't know very many projects that are building on our weave or leveraging our weave I know like the Redstone team are doing some stuff with IBC and maybe the cosmos SDK but there's not a whole lot of overlap between those two ecosystems I think I mean maybe I'm wrong but what what is the the kind of play here and how can like the cosmos ecosystem like Leverage our weave to to

its benefit do you think yeah I think I think we're getting more and more into this like modular um text deck right where you have some technology it's in our case right where the logic well Cosmos is very suited for like tokenologics and Logics in general so we use this for that obviously suited for storage so we use that and actually what's very crucial is that the bridge in between is important and that's basically what kive is right um and for example I would never recommend actually anyone else immediately integrating with our weave um if not your whole system is completely trustless right because once you store something on our vehicle because well it's there for ever right you really have to make sure it's actually correct because there are huge huge huge potential like attack vectors

um there um that are just really like dangerous and for for example right because you just mentioned like the Redstone team right um they've they're building this Oracle system um and they actually have this system where you can dispute a piece of data after it got uploaded right from my personal point of view it's not um a very elegant solution to a problem because let's say I'm uploading a financial data stream Euro to USD prices and some hacker is able to corrupt the incoming stream right then my application who's using the data is already using this incorrect data and then later on after I might have the damage can then say yeah raise it dispute this is incorrect but actually there's nothing reimbursing me actually

that that there has been some incorrect data being uploaded because it's like from one Central right right now at least right from one Central provider being apple and that's something we do different with kive but we say listen we are actually uploading and validating the data and after it has been found valid just then the kive chain publishes to prove and says listen developers listen apps whatever we know that this are with transaction ID has some absolute correct data that you can use in your application safely right so could you could you leverage kive to build like a roll-up for example that like in the same way that Celestia is a data availability layer for for Roll-Ups I think the Redstone team was doing

something kind of similar where they're leveraging our weave as a sort of data availability there it's just also what one can do with Kai yeah Yes actually so the kive really possibilities on that side are kind of endless because I like to say what Skype is a big of a triangle right and like we're really focusing on getting web 2 web 3 data into web 3. right and then you can do whatever you want after the data is there you can do whatever you want so you can use it for for example I should see the little region ticker um down there right it's like they have a lot of off-chain data they are working with and they have reached out to us and said hey guys how can we get this data on chain how can we make it turn into a public good right

that's for example also use case of kive that's Way Beyond just the archiving of data right also turning data into public good making it accessible making it a trustless so I can safely use it um between chains and then with IBC Excel and so now you can leverage tons of already trustless Bridges so then get my data from from A to B yeah so let's maybe talk about some of the use cases here so I think one of the use cases that you mentioned in your blog post is um these like archival notes so in ethereum you can have different types of nodes one of those nodes is an archival node it calculates all of the state transitions since the beginning of the chain those nodes are usually quite uh long to sync and take like weeks perhaps

even like years at some point when the state becomes large enough they're not a whole lot of people also have archival nodes I think like I think at some point it was estimated that maybe a handful of archival nodes existed uh maybe even the ethereum foundation didn't have one uh or something like that but yeah but but but here with kive we could say okay well the archival data is available on kive it's been it's been proven valid and so therefore you know we could maybe forego having archival nodes or this can kind of act as the the a source of truth and you could do that for other blockchains as well is that is that kind of a value proposition here yeah you absolutely

nailed it it's basically about this like data accessibility and datability right because as I said earlier there is no incentive for someone on ethereum to run like in eth of course no one Center for eth to run an archival node I mean I can sell it off to like analysts for example um or projects that need to have access to archival data and I can just charge fee at revenue on it but there's no protocol based incentive for that right and actually with Chi if this this idea switches right because once data has been stored with kindness publicly and free um available for everyone right meaning there is no need anymore to even run your own archive level because you can just get it from there and be actually building out some super exciting tooling right now which I think we will well maybe released by the end of the month

and that allows you to get your data from Kaif into any Source destination you want local database Excel sheet if you want to right or like your big snowflake a data warehouse and I think that's also something we are kind of like actively pushing there to then get rid of this problem of like archival nodes and even node operators right can use this to sync their notes even faster because well they don't have to worry about how do they get their data from this or that chain really foreign architecture like what are the different layers here and how do you Leverage The Cosmos SDK I think that's what people are really interested in sure sure um so basically it is actually a very easy concept if you really boil it down so

basically work with so-called storage ports and storage pools they run protocol nodes that are relaying data from A to B so they're relaying data for example from from Cosmos right or from polka dot whatever over to our beef and then they are also validating this data and the way it works quite simplistic populate as an example I pull down block number five from polka dot and push it onto our weave and then tell the other nodes hey nodes this is my proposal of this data can you please check if it is correct and then those nodes would go ahead download this data from our weave and maybe pull down the local copy check if it's if it equals and if it equals it's fine they vote with yes on this piece of data if the majority has found

yes then the cosmos chain picks this up and says okay developers here is this uh bundle this has been found valid or hey his respondent has been found invalid so please don't use it and then of course we only install the ballot once so it makes more sense right um so you can get along um this is that's a very fundamental process and the incentivization behind this is that the storage pool requires funding so you purchase Kaif token off the market put them in the pool as funding and then Ah that's amazing perfect and then the pool pays out um the kive Jones to the validators as a reward for relaying this data um and then basically then the validators well they can have delegates like a proof of stake system but then basically that's the easy process like a

triple down if I write it the foundation might purchase type tokens and put them in a pool as funding to make sure that they create a constant Archive of the data the tokens dribble down to the validators and the validators might restake it for security or sell it off at the market so it's a purely utility driven token so so we have funders who who are these funders practically speaking uh usually I would say foundations or projects that are dependent on the data um easy example if I'm for example the graph let's just take the graph and Solana right the Solana of course has an interest in creating the archive of Solana for well cost reasons on their end and the graph is maybe interested in get ensuring access to Solana data

because they might offer a product on top of this right like the software for Solana and actually now it's cheaper for both of them to just fund the pool 50 50 and effectively split the costs right then if someone would pay for that uh would pay for that data and then of course a lot of people actually argue well isn't there a free rider problem because if I might be a small person I don't have to pay I can just benefit from the big fish to not paying this um yes true but on the other side the big player will always make sure to refund the pool because they are building a very very successful business model on top of this data right it's a bit like using a free version of a database online and then you say well yeah I don't have ever have to pay for

it well yes of course because at some point you're going to scale and then you want to make sure you don't put your customers at risk right so you start of course paying the database provider for providing your database it's the same concept there and because then a lot of people can split the costs it saves a lot of money for everyone so we essentially have you know actors who are incentivized to have this data be stored permanently um the easiest so the easiest go-to example here would be like a foundation or the team that is behind a project they're incentivized for that data to be stored online they would fund uh this pool that pool can also be co-funded by several entities um the

the staking those funds then go to I guess here where I don't understand here is like what's the role of the validator so the validators they are the ones who are verifying the data that's coming in signing the data testing that that data is correct before it gets posted to are we so the delegators the validators get Rewards who's who's paying our weave yeah so the validators are paying out with right so so in a pool we have let's say 50 validators for example right and then we have this round system so every round every round is one block for example right so in one round I might be the

uploader so I store the data onto our beef and I pay with this in R right and then because well if my data has been found valid from the others then I get reimbursed in Kaif tokens right which of course later on I can sell for the r I've paid now and then in the next round you might be the uploader right and then you do the thing and then the next one I am the upload again so it switches um always from from A to B and D probability that I'm going to be the uploader is the terminal my on my stake so of course I've always have a high stake because I get a much higher reward if I have um basically stored the data um and so the stake the staking providers the validators they need to also take

into account like they need to also have R they need also um be liquid in R and there might also be some kind of Treasury management there where they they might be buying and selling R in order to maximize their return on uh on their staking rewards so there's probably like some game game theory and economics going on here to maximize faking rewards versus like the amount of R they're spending exactly so you can do this or either on the other side bandra offers um archive payment integration so you can store your data through bundler and then you can just pay native Kaif so you don't have to worry about that in theory but something else which is a very cool use case that Kaif can do as well let's say you are

having a web 2 API it makes you for say Network we have this use case with sports betting data and this costs you of course Fiat money right yeah what you do is you purchase your API key with Fiat money and then you just pull it into your node and then you know it uses this API key to relay the data but then of course the node still gets reimbursed in kaifton in order to pay for the API key in every pool has its own parameters on operating and storage costs Right Storage costs depending on where they saw the data it's a factor of one and a half on top of the RV price so they get a 50 profit and of course the operating costs can be set also on a per pool basis for example if the API key is a hundred dollars a month or you can do

the maths right to get a Clive price that will reimburse people for that and then through the governance those parameters can be updated whenever okay so can you explain the process by which validators um [Music] verify the data so when let's say we're talking about a web 2 uh API because I think that's like a simple example so you have an API for I don't know stock market uh uh prices right like over time and you want to have that data be in a permanent storage as a user of that data what it what are the assurances that I have that the

validators have come to consensus around this data are they all signing this data collectively and what it yeah is that is that how it works yes exactly so to to to to make the shot yes so they vulnerate they're not signing the data they are voting on it with a result either true or false right and the voting itself like how you come to true or false is completely up to the developer that is writing the implementation right so for example for deterministic blockchain data the proof is very simple is that's the copyright with equal my local my local copy yes the equal okay I'm gonna work with true right um and then for example for pricing like stock market pricing there might be a little like slippage in the pricing

right because maybe in Frankfurt the exchange rate is a different one than in New York so actually in your validation function you can just code in the slippage check and return for example true only if the slippage is less than 0.1 for example as I think what um chain link is doing so it's completely up to developers to customize this to their needs they only have to return true or false at the end that's the only thing that they have to do and then of course watch the notes how you can make sure the nodes vote correct is first of all they don't have an incentive to vote on the wrong side like either invalid or valid because then they don't get money for that because well they've been wrong right and of course they have the stake as a security meaning that if they voted

on the incorrect side their state gets slashed so they receive us like a slash for voting invalid or being offline so you can make sure that there is this Economic Security from proof of stake behind the data that has been uploaded hmm I'm trying to think I was a web 2 service I think maybe like as a web three so here the ethereum foundation and you want to store the entire state of ethereum in kive I think it makes sense yeah to have a valid intercept that is verifying that data and posting it to the chain because the ethereum network can't sign data yeah

um we don't have the cryptography yet for a blockchain to hold its own keys to sign data for something like let's say I don't know like what one example like Bloomberg API or something right and like you're you want to store the stock market data into the into the chain I could I could see where Bloomberg might just sign data and post it straight to are we like it like for web 2 Data if you trust the signature of the issuer of the data then you don't need kive you can just go straight to our weave right yes exactly so I think this is now a bit of like this this question on like how much security you want to have in a system right where you can say well I'm already

trusting the bloomwork API on web2 so I'm just going to relay this one to one into web three and then I'm going to use it there right you have kind of two well it's a perfectly fine solution first of all right but you have one problem with it first of all something on the transport could go wrong right which you don't know and then how do you know later on if it's invalid it's a bit of this Redstone problem I talked about earlier right so like relaying the data immediately but then later on maybe if you already have the damage you can say yeah but this wasn't correct and like okay cool like sorry for sorry for that um but there's nothing that's gonna happen so that's one problem you you know but if the data is signed if the data is signed then you have the

signature over the data are you talking about missing pieces of data or incorrect data because of the data design there's no there's no risk of this right no exactly well if it's signed then you just have to well check the signature on your end right yeah it's actually correct and then you would have to well also copy the Bloomberg data down right in order to create the other signature of course that's that's a valid solution but I think it gets a special interesting when talking about multiple data streams at the same time right because if you want the system to be completely trustless on some data that can be trust us right like Stock Exchange Trader can be trusted if we go to multiple providers of the stock exchange data and of course then like the game comes more and more into play

where you say okay now we have to get multiple sources in have to find consensus about this state actually yeah yeah because I mean if like if in any case where a sort of traditional centralized company wanted to have its data available on our weave they they could just say hey here's where I say here's our private our public key we're signing the data we're going to put it here for the benefit of everyone they might have some economic layer in there where you know you get to pay you have to pay for it or something I don't know how that worked but let's just say let's just assume it it wants to be the data wants to be free um then if you're using that data then you're trusting The Source you assume that you trust the source therefore you

can trust the signature in this case right so let's say you had Bloomberg and maybe some other data that you wanted to aggregate um in in this case where you mentioned say like yeah you want to have you know stock market data over time but you want to do the um the average price across you know several exchanges or several sources of beta then that's where Kai would come in handy because you'd have the validator network that is sourcing that data executing some sort of computation on the data um like like an average uh of a price putting that data in uh our weave and then signing it but effectively you have a new here it's a new data set it's not

like a one-to-one data set it's actually a new data set and this is where uh Hive has like valued exactly absolutely absolutely right and yeah also if you have to design your runtime implementation of course to solve this problem right on the other side you can always argue for example with the Bloomberg thing is is it Ethan is it even in in crypto mind to just relay they like just have a one-on-one copy from a web to like from a trust Falls like just sounds so weird but from a source that you would have to trust like Bloomberg for example right and then getting this data right into our weave and then do you trust Bloomberg and then do you want to expose especially default right do you want to

expose your users that are for sure thinking that your system is 100 trustless right we wanted to expose them to that risk and our lead developer always likes to say yeah the only reason why we're making such an effort building blockchains and making everything trusted because we just expect one single person on this whole planet to fit without with the whole system right yeah um and so I think it's really important that you then kind of like think about okay what are we optimizing for right are we optimizing for True crypto then we cannot just relay data then we have to make a trust yeah I mean this this I think like the philosophy I agree with the philosophy uh here where you know it trusting single sources of

data exposes you to risk right like uh in this case you know let's say that there were applications that were relying on this data and someone malicious at Bloomberg wanted to provide some faulty data that would um cause say like I don't know cascading liquidations or something like that like I mean you could do that um there there might be some use cases where one would trust that data and would like be content with that with trust in that data but I agree that like we need sort of like resilient data uh that perhaps comes from multiple sources and has been aggregated and um where there are like mechanisms in place to you know fish outliers and and things

like that um yeah no this is really this is really interesting uh I I really like this um what up so who who who who decides basically like okay um in terms of public goods right so coming back to this ethereum uh archival node idea can the community kind of come together and say like Hey we're going to fund this thing because we think it's a public good and it should be available to everyone like I notice in the kive app I think you guys are archiving like if you know Bitcoin and polka dot and Moonbeam and a bunch of other things

like who's funding those things especially like Bitcoin who's funding this Bitcoin yeah so so right now but because it's testnet um it's like I mean we are we are funding this right with the ecosystem wallet but actually um well we have some measures in place for mainnet it's actually that our first ever fundraise were just um just foundations from all the blockchains with convertible notes so we had the near Foundation we had uh Solana we had um all the others right um and so basically it's like this is they've purchased this these tokens early on to provide this funding for their communities right then I kind of get the initial spark um started right but of course it's a completely permissionless system so of

course if you have three people that have enough money and say well we would love to see Bitcoin being archived they can of course get together and just turn it into public good right um and uh yeah that's that works so for here for example we had like the Bitcoin blockchain being archived there's one person funding it exactly and actually if you turn on the little expert mode in the upper right um for the for the audio only uh listeners we're just exploring the the kive web page um and now um I think you should see the the funders if you scroll up there yeah see there's no thunderstop the second one if you scroll up a bit there's no there's no funders tab exactly and then you can see who's funding the pool right

now so it's just the Kaif team wallet um just funding that of course it may not we have different right things in play and it's also a cool mechanism in my opinion because it's a true validation on the business model because if you create a runtime and no one is interested in funding this well there's also apparently no value in this data being stored right now right so why would you do this yeah and so as a as a as a validator or as I guess as a maybe not a validator but as a we call it protocol notes right okay so as a protocol node if those protocol nodes also need to have clients for the chains their store

they're they're storing in uh in in are we through kive so if I'm you know if if we're backing up the the Bitcoin blockchain there has to be some client there that is relaying that information through kive exactly we're calling these clients we call them runtimes um and I think if you would go back to the app now you would see it's like every um pool is like their own little runtime of course all evm based chains are all using the evm runtime the Bitcoins using the kive Bitcoin runtime uh I see okay exactly and we made this as easy as possible for developers so the core of the runtime is already written by by Kaif and it just exposes two functions

one is an upload data function but you have to return some bytes that's something a developer has to write and the other function is the validate function and the default one is thus a equals B but if you have a different one well you can overwrite this function and then you publish this binary and then the node Runners have a pretty cool tool called Kaiser which is the homage to the cosmobilizer which makes like running your kind definitely two breezes you just point it to a pool and the Kaiser automatically downloads the right binaries and also automatically listens to updates in the governance and pulls the binaries in time so you don't miss the the update okay so I wanted to ask you about this so uh for listeners sorry Cosmo visor is a

piece of software that is run by validators and it essentially manages the binaries um right because like as a validator you're running software on a machine and whenever there's an upgrade so let's say Cosmos Hub uh as you know Cosmos Hub binary guy ID goes from version one to version two um as a validator you have to update that software and Cosmo visor is sort of a rapper that just allows you to put the software there and then when the update happens you're validator automatically switches over so it's kind of like a it's a service that runs on your on your on your validator um how does how does this kive Cosmo visor

um um what's it called Kaiser Kaiser it works the exact it's the exact same idea behind it the only cool thing is it actually supports listening to like multiple pools at the same time right because um actually we are rolling out an update now this is a complete Alpha right now for for I think everyone listening for you as well we're running out we're rolling out an update called interpool security where now one validator is active in multiple pools at the same time um and then even of course Kaiser makes even more sense because then the binary well the wrapper automatically listens to all updates happening on all pools and then you can just have all the download enabled and you're good to go and get um never miss an update so

okay very interesting very very interesting and so as a as a validator as a validator on say Cosmos if you're already validating the cosmos chain should you just start validating Hive and like get rewards for or like is is kive as a validator node are you dependent on running also a validator node on another Network or is it fetching the data through IBC yes so so it is fetching the data by your foreign API endpoint right where you're leveraging like Pock Network for that coinbase cloud for that

right now um chain stick we are in contact with right now right for that but of course you can plug in your own so if you're already running like to keep the story completely trustless right you're running your own RPC and RPC endpoint on Cosmos you're going to Archive Cosmos just point it to your own note trust so it's just a chive is just connecting to an RPC endpoint exactly so you could connect to any public library C endpoint or if you're running your own validator then you'd want to connect your to your own public RPC of course and you can code this connection thing however you want right this is just how we as a team have implemented it if we take the example of course off while storing um does the sports betting data right you might work with one provider and you

of course write a function to just fetch from this one provider or for multiple it's completely up to developers to call it okay yeah and um so you're currently in incentivized testnet uh uh how's that been going crazy honestly that's it here they are it was crazy we had um you know we had I think so the the incentivization phase ended a few weeks ago um we were like cleaning up the leaderboard and everything with 40 000 participants which is insane and participants really being like people that have like staked actively right claimed from the faucet and so on they're not people that have just created address and send some tokens right that's been absolutely mind-blowing um experience from a well-growth perspective was a great

stress test um for the current architecture it is running on the corellia sitting of the testing so that's running on the corellia Chain I think the team learned um a crazy lot of this and yeah it was very valuable for us um I must say the contributions we got from a Tech Community and meme perspective were absolutely insane really what was the biggest lesson from the inside of my assessment um good question I think we had some well I think a good lesson for us was that we are actually capable of running this stuff under load and what like real load on the blockchain looks like right and then we were really able to optimize for that so I think we're going to do one or two like one more big update now and then we're actually done then we're

like it's gonna be ready for mainnet then I hope it stays with the One update I'm pretty pretty confident though that you know it's Tech um but actually so I'm really hoping that this kind of works out the last update and I think they're really good to go give us a great ux ux insight into everything so yeah what's like beyond the storing your favorite blockchains data um in uh in our weave what are the other applications here like if you're thinking more long term and outside the box in terms of what this can be used for in the future as web3 grows and new source of businesses are built um on on using crypto and blockchains what are some of the things that you'd

be most excited about uh for in terms of the use of uh Kaif yeah I think really that Kai will become this like critical like middleware in your web 3 stack right because it allows you to well communicate well allows you to get the in communication from web 2 and other centralized web3 data into a completely trustless web3 right and then utilizing some other tools like excela um like our new like Oracle module and stuff like this to then do the IBC to IVC transfer off data right of information with exceller you even get a cross chain and stuff like this I said we're rolling out a tool by the end of the month that allows it to get it into your favorite web 2 stack so like a

whole new role for analysts there right and even like projects that might not work in web 3 but are interested in using web 3 data it's like a no code solution it's one button and they have it in their favorite Warehouse accounting is a very big piece where we see a big use case happening right it's a big thesis of mine now the pool Market is over everyone's going to look back at compliance and accounting and they obviously need archival history right or history access to data that's something I'm very bullish on and I think that's kind of the the way we're going kind of like yeah increasing our market share um there what's the roadmap like uh where where when are you launching mainnet and when can uh yeah when can we start building on on this well ASAP I I hope so uh no

so we were planning to launch the um the mainland in this quarter um we're just going through like the last stuff we have one audit in place and stuff like this and we hope that there's no major delay in that we're really looking forward kind of like getting it out um by the end of this year um and I think then we're ready to go from a tax side it's already tested the validation works of course from a mathematical point of view there is no Economic Security behind it because we don't have a token but that's like the big next thing on the roadmap getting the main it out and after that it's a lot of well community and ecosystem building um I would say right like use cases pushing for use cases granting out um tooling around it um and so on and so

on at the the road to decentralization after many that gets wait you said you don't have a token there's a kive token right yeah but it's not trading yet that's what ah okay yeah okay so in test it of course you can play around with it but it's not trading okay interesting uh where can people learn more about kive and I mean what is what is your biggest ask to the community right now like what do you need people to do basically um it would be a way more like you know find a use case like without finding use case but building out use cases on the tech side I think this runtime development is a is a big thing um we will yeah as I said push a new update there very soon um and then of course it's a burden on us like right to create some better

documentation around this right and get people um on board um if you want to learn more about Kaif like the dogs are a good place to start um and then from there to the Discord with a very very active Discord where you can ask a lot of questions and they usually always can answer it um and yeah I would say that's from there if you just want to check out some news Twitter is a good place I think we're tweeting run every day every other day maybe um it's always some some very insightful content we're putting out there cool uh well this is really exciting I'm uh I'm hopeful to seeing uh yeah all kinds of new applications being built on this and I mean I think for me the thing that um feels cool about this is that it is

you know practical use case for our weave um not that there aren't any use cases on airweave but I think it's a use case that makes sense in the broader web 3 context and you know could be uh a really good sort of like long-term incentive for our weave to to stay up and to remain permanent uh as it claims to be as it claims to do uh yeah thanks uh thanks so much Fabian for coming on look forward thank you so much I really enjoy it thank you