DePIN, AI, and Onchain Storage Unpacked

Transcript

welcome back to the rollup High Roller Community as always nothing on the rollup is to be ConEd as Financial advice we are just an Educators in the defi space toss us a quick like give us a share a repost send it to a friend who may want to watch this and finally give us a sub or a follow it means the world to help support credibly neutral educational content in def5 we're here to educate Empower and enrich you and let's jump in What welcome back with another Discord notification ping from our W Rob we had our boy Pat in the building and we are talking about or someone will take a little bit um honestly misinterpreted and like there's a lot of of understanding as to how all

these pieces kind of come together um so excited to kind of break those down and talk about some of the interesting things happen in IBC um yeah what's up that how are you that's never DL Roman that's Jack a LS that's your sh and a lot of building the uh the the last few years were pretty brutal but we're happen to be coming up on the other side of it um outside of that just building Integrations building Outpost exploring like AIML Integrations with teams like Morpheus for example um they're just crushing it right now we're having blast and kind of bringing decentralized storage to the cosos ecosystem and uh Beyond through IBC and onchain Integrations is pretty fun let's dig into the the AI ml Integrations that you guys have coming

around yeah um yeah V vitalic posted a tweet today about his like thoughts on the intersection of AI and crypto haven't had a chance to read it yet but like what are your thoughts on the intersection of AI and crypto I mean there's a lot of different intersections right and there's different ways that you can build the centralized applications to like fit a use case like that um there's different parts of the stack that are really important so you can think of a compute is a big one where uh computes expensive and by do centralized the and I me kind of St Market places where people are using like their spare compute to hook into a network and earn on it while uh while they're not using it for example you st is like a cal aome for things it's like where you're looking to either um just

like get instances so you can train AI or run different applications it's wonderful um the other piece is storage right so there's Vector databases and there's also just storage of history or weights or um kind of hitting checkpoints and making sure that the AI will like remember from like before right so it's one thing to have ai but it's another thing to like have a memory for the AI so it can kind of add continuity when you're interface and like that um that's something that we don't really have right now in blockchain um to serve a use case like that and you'd have to kind of add centralized CH points of failure um whether that's like an Amazon ATI or something like that where it's not really as quot code decentralize if

that's what you're going for at the end of the day so um early thing lowest thing for the jao protocol we've been looking at Integrations um for memory of chat history and weights um the next thing that we're looking at is rag frame March so this is like retrieval augmented generation and this is like really cool for running like lightweight purpose built AI applications that are trained on custom data so if you have like a company and you don't like want to share the proprietary companied information with a third party or there's laws preventing you from doing that so you can't like use like a chat GPT ATI for example um we give the ability for you to build like self- custodial and private um Vector

databases essentially where only the end user with the private key has access to it so that's great for people that are kind of boxed in by regulation but it's also great for people looking for decentralization um where you try to defend against things like supply chain um or sorry information and supply chain attacks where like you're cutting off AI developers from different data sets or things along those lines so that's where we see uh us spitting in but in the grand steam of things it's really early for kind of AI in general not let Al on like blockchain eets Ai and I'm just excited to be along to the ride as we continue to develop it for sure now that gives me hope because we've seen a lot of these like AI times deep in narrative 100x moonshot

coming oh really I'm just thinking like man it's so early now like we're just like gotta really set the set the foundation which super important yeah yeah it's it's super important it's uh it's one thing to Market something like that but it's another thing to do the like boring cryptography work and the boring infrastructure work to make the thing that we're marketing possible right and uh for us like guys that are kind of on the bleeding edge morphus for example I've probably mentioned them already but uh they we still have a long way to go before the infrastructure is even there to do something um like that right so I'm just happy that there's teams that are kind in it for the right

reason do the right thing and uh um Greg Greg said something on Twitter so he's the founder of a of a cach recently where if you see something where it's just like dpin and bold letters or just like Ai dpin and bold letters on a landing page I would try to stay away from that if possible but um most of the the guys that are kind of in it for the right reasons we Market what we are right Jackal stores data and that's what we do do we fall in the category like dpin absolutely but um at the end of the day it's we do something really simple and we find place for you to put your data and make sure that it's super private with a high security posture and that's pretty much it then one thing that we're realizing um in in this

interesting dynamic between like cosos and ethereum just scale is like the different approaches but they're all like like basic cosos had kind of the horizont approach each that like like like the vertical now they're trying to both kind of come together and not like merge but like you know almost like like take what the other has done and try to implement and um what's really cool is that what we've been woring about a lot about the scaring Stu is that this all comes down to data it's like State looat um data publishing processing there's all these different things that all come down to the actual like level of computer ships and compy and do size and all these different small things and so

like I'm curious how you see about like the perspective of scaling um you know the the capacity of these distributed systems with um you know with check like how does the the the the data um storage and kind of system which you to allow for further throughput further scale and kind of be your uh capacities for these systems yeah for sure so so there's kind of a different between like data availability and a grand scheme of things when people say like data availability data availability it's usually just referring to the blockchain itself and how much data you can store on a blockchain um this is kind of you'll see like Celestia I think they're on a higher end where we're looking at

like 10 megabytes and then on a smaller end you're looking at like 100 kilobytes right or less possibly depending on what chain you're building on so this is like how much data can you like fit on chain before like all the wheels falling off essentially so kind of figuring away for that to be scalable without the the validators just exploding that's a 100% really important thing to do what we kind of focus on at Jackal is for things that are bigger than that right so anything like over 10 megabytes you got to put it somewhere so right now the consensus in web 3 is okay like I guess we'll just spit up a an Amazon server an Aur server and kind of add the central Tri point of failure for um for the applications on a building take a nft

Marketplace for example the nft marketplace takes like the metadata of the nft and then they Jam that on chain saying like Okay Andy owns this nft and then the file Associated like the jpeg associated with that nft they'll put a link in there and it's an ipfs link usually which just means that they have a Amazon server that's like pinning to ipfs or a centralized server somewhere that's pinning to the ipfs Discovery layer and saying look over there that's where the nft is um this is something that like we've kind of been focusing on where how do we decentralize that storage layer in general to make sure that in the event that the nft developer uh did forgets to kind of

renew his credit card or God forbid he gets hurt or something like that he dies then you have this situation where once that Amazon server expires your nfps are not um like C danger right uh just Absolut Community implodes because of something like that so looking at ways that we can sustainably have decentralized storage where it's um you pretty much have this ability to Leverage The Jacko protocol storage layer by integrating cross chain so you don't have to kind of build your blockchain for like this specific use case of storing really really large files but you can integrate with the Jack protoc code via IBC or GMP where with a simple smart contract call on any

chain you can just get apps of the scalable storage so kind of our Niche that we focus on is like larger files you can think about PDFs or we had use cases right now we're working at a telecom company in Iowa where they're pry much storing like all of their user data and they want a way that they can have a really decentralized secure and redundant way to store that for like an awesome price point that's what they're looking for um on the flip side in the web three STS we integrate with a lot of nft marketplaces and stuff like that so um the way that we look at kind of data store storage in general in web 3 is through that lens of where do we put all the things that can't fit on the blockchain and uh let's make sure that we can create a decentralized way in an

ethical way and um a really high security posture and sovere and interoperable way that you can kind of Leverage that from Storage that work locally with just a simple smart so that's how we look at it and from a data availability perspective for like really large files what we focus on is this is on our road map it's coming in a a little bit but um the ability for smart contracts to read and write to offchain data so it's one thing to store something and then just retrieve it when you need it but it's another thing to like update files with a simple smart contract call it's kind of another thing where you're looking like large scale data availability for like really really large files that's kind of where we uh we we slot

in is the reason that the that you're not able to store all these files on chain is it is it just sheer size is that the the limiting factor of the chain yeah yeah usually it it kind of depends on a bunch of different things right so if you look at the validators the validators are kind of like these purpose built machines for compute really um because they do a lot of computational heavy things and then um usually like the parameter for how much storage fits on the Chain um is set by the code right so um I don't know if you've realized like this or all thing everyone's like or lower lower than those right now right um but you've noticed that the ordinals of the inscriptions have kind of like made their rounds and just taken chains down

this is because you're trying to jam like too many transactions and these transactions are like really big because you're like taking a JPEG and just jaming on the Chain right um and this is an issue for them and their chains are going down specifically because the validators can't handle it so the way that Jackal operates and the the reason why we kind of do something like this is we have like two peer toer networks one is a validator so they're like kind of the police of the storage Network and then we have another PP Network which is storage providers so these are like machines purpos built for storage and uh they submit attestations to the validators and validators say okay you're still starting their stop and that's kind of how we make sure and maintain redundancy and all that stuff

so um in short that's the reason why in bles aren't really meant to like store PDFs right or like CCT camera footage or large video files that's why you kind of Leverage an offchain price to store it and then you reference it onchain um using a URL or a pointer something like that right so how does that apply then to IC into Cosmos with your um storage Outpost of how does that kind of concept translate yeah so when you look at kind of IBC and Cosmos the reason why we built on the cosmos Tex St is we want to kind of purpose build a blockchain for storage and then have the ability for all other blockchains that are iidc

connected to use those resources easily right so um if you look at like other infrastructures and older infrastructure when it comes to like storage layers in general it's we're going to build like this monolithic L1 style thing most of all of them are still proof for work chains and then what we're going to do with that is if you want to build an application that requires a lot of storage you have to build it in your wall garded on our model they all want um we think of that great for a lot of different use cases but for the majority of use cases um that's not really going to fly because I don't want to build on top of file coin or ar weave or these other chains because I want to access liquidity or I want to build in a place

that's more vibr so a lot of people end up building on etherum and then they leverage like ipfs pinning through an Amazon server and that's their way around so they can kind of have the best of both worlds our vision is horizontal scalability and kind of why the reason we built on Cosmos is that you can build on Juno or Archway or Neutron or Seer Network and if you want data storage and we build an outpost on that chain that means that with a simple smarton trap call you can get it and it's really really smooth it's really really simple and we use kind of IBC controller contracts basically so as long as we can get a message back to the Jackal chain and then update like file trees or storage accounts or storage payments then you can easily access it from

anywhere which is uh pretty pretty cool stuff at the end of the day that's like basically enabling then um all these different chains to to use jet well that's like a work in progress how's that in coming what's the pl and does that integration look like yeah so right now what we're doing is we've been working closely with the inching foundation so that's kind of like one of the PLM like governing bodies that exist and they're they have uh we've been working with the controller contract and we've been working really closely with them on that now for us we have the contract up and running it's it's working so um the next thing that we're doing is we're trying to find the

perfect Landing space for like an outpost like this so whether that's on Archway or whether that's on Neutron or whether it's on Celestia um we're just doing our research and just try to figure out the best place to deploy it first and then um in the end game for us is that every blockchain is going to have a jackal Outpost where all the developers and users can just access scale of offchain storage quite simply nice and is that uh in it to be Transit um and used somehow in like non IBC environments or kind of what's the what's the uh the limitations of of that yeah so we can't like really put the cart before the horse on some things

because the infrastructure is not there yet so there's some guys that are building like IBC connections to ethereum for example or IBC connections to Solana or Avalanche as soon as we get those IBC connections we can start deploying outs there as well so it's uh it's either we're going to build it ourselves or we have to wait for someone else to build it it's kind of like a matter of like when not if for us but the other thing is in the event that we don't use IBC we can also use a GMP so as long as we can get like a packet back to the jao protocol that's all that really matters so um we're just kind of leveraging whatever is the best technology to do that and whether that's IDC or gmpd um it all works right now we're just focusing on IBC because we're in the

cosmos and there's we have like 80 or 90 connected blockchains that we have to work with so uh we don't want to boil the ocean all at once you know yeah I think Union's coming ready to do some ZK ICM stuff yeah that's only wild um I'm excited to that pretty much uh yeah as soon as we start connecting all the blockchains I think I Cosmos thesis lens too right it's you're just there two types of blockchains there's ones that are cosm of blockchain are just not IBC connected yet so we're we're patiently waiting of course key word is yet yet exactly exactly yeah I think bring everywhere is kind of like a motto and I think it's

some think for the better of of kind of all people all yeah like all Builders all that yeah you can go on like rack I don't know if You' ever been on like rack a it's like that one of the the tough it's uh one of the worst websites but also like the most entertaining websites of all time but uh essentially with that like they rank like power rank pretty much all like the defi hacks and almost all of the her Bridges right um the beautiful thing about IBC has never watched a single dollar and connected it's we do billions and billions of dollars with the volume so it's it's really beautiful that we have this um this wonderful way that we can uh kind of integrate with

all BLS with and reduce like the bridge risk as much as possible so it's just really a matter of time for us we think it's the best bridging technology out there by landl yeah I'm a beli as we kind of um come to a close I'm curious what you guys are um of what we discussed amongst many different kind of angles and verticals what you guys are most focused on for uh T and into Q2 here and um and any opportunities for our community to get involved yeah so for us right now we have a road mapap that we just published a few days ago so that's like the best place to get anything of like what we're up to internally and we also have our every Wednesday at two EAS is when we had our like Town Hall session so if anyone

wants to hop on those that would be wonderful but for the rest of key1 we got Hardware wallet support going so pretty much anyone with a hardware wallet has like this infinite USB drive that you can just rip around with which is awesome um so that's great and then kind of moving into the next thing as we're looking at um our V4 upgrade which kind of is hopefully where fingers crossed is like the last like major upgrade for the J B which enables us to do bunch more stuff which you're really excited about I don't know I can get into that right now because it's we'll be here for another three hours that everything go on a V4 but um outside of that it's uh kind of focusing on um Ledger compatibility done V4 upgrade is the next important thing and then we

have one op this laun chain right now um right now is in like hopefully by the end of this week which is radiant so cool thing about radiant so we have like the Jackal dashboard and Jackal Drive which is like kind of like a private Dropbox type product and now we have a radiant which is this kind of non-private it's like public publishing of information and it's stored for 200 years in perpetuity so you have this kind of the balance of both both extremes I guess is the good way to put it and then we can kind of fill in the Gap so we continue to go very very cool excited for that and we'll uh we'll keep uh we'll keep on the lookout for

radiant yeah cheers man thank you for coming on that was a pleasure and uh to bring ABC everywhere being Outpost everywhere and that that keep those keep that state loow down and that data EAS and just fashion plan that's the plan all right boys you guys say goody yeah