I think what's really powerful about crypto is its ability to like mobilize a massive amount of resources. Right, Like, if you look at just the number of GPUs that were in the proof of workdays of ethereum set up to mine ethereum, it's like fifty times as much compute power as what was used to train chatchipt. Right, So you have like a huge amount of resources. And AI model required huge amount of resources to be trained, right, it can cost like ten million dollars of train a model.
This content is brought to you by Bitco, which is one of the top crypto custodians in the crypto industry. Bitco works with many big companies and brands such as Pantera Capital, Bitstamp, and bitcoin Ira. Nike also selected Bitco to power its wallets for its NFTs, and Bitco has many great services such as hot wallets, custodial wallets, self
managed cold walls, and NFT wallets. Many institutions trust bitgo with its top level security and incredible services such as being able to deploy your capital while it's in custody, which includes lending, borrowing, trading, staking, DeFi access and more. If you'd like to learn more, about bitco. Please visit bitgo dot com link in a description. Welcome into the
Thinking Crypto Podcast. I'm your host, Tony Edward, and with me is Anna Keslawskis, who's the co founder and CEO of VANA and a great to have you on.
Yeah, great to be here.
Anna. You're doing some great things and very interesting things at VANNA with the convergence of AI and blockchain, and I'm really excited to speak with you because I've been noodling and thinking a lot about these things and data and how you can monetize your data and own your data in the web three world and token i should say token iss world. We're headed to. So lots of questions, but let's start with your background. Where are you from and where'd you grow up? Yeah?
Yeah, so I was born in Montreal, lived there for the first five years, moved to Sweden, and then moved to Minnesota, so mostly grew up there. I yeah, half Lithuanian, hence the last name, half Filipino, and then I ended up.
Moving to.
Boston for school at MIT, and then yeah, left school, moved to San Francisco and have pretty much been here since.
And before you started Vana, what was your professional background. I noticed you were at crypto company Celo. If I'm not mistaken, tell us a bit about that. Yeah.
Yeah, So I got into crypto through, like I guess, an interest in traditional currency. So I had worked at the FED and high school. I actually had a picture of Janet Yellen, chair of a FED and in my high school bedroom, which is like totally like who does that, right, And then got to MIT and learned about these decentralized currencies. This was back in twenty fifteen, and then it was like the five person MIT Bitcoin Club. I started mining ethereum from my dorm room and just got really fascinated
in kind of the world of decentralization. And I also sort of came across the world of AI through MIT as well. I was doing some research at c Sale. Basically this was when the attention is all you need paper, which is kind of the foundations of chat. GPT had come out, but people hadn't figured out how to use
it to generate new texts. They had just figured out how to use it to kind of model some text and so I use some of that to automate my job sorting documents at the World Bank, and then ended up leaving school, going through I Combinator, discovering this world of like Silicon Valley, my life's purpose is not to sort the world's documents, And so I actually ended up joining Cello as an early engineer there focused on kind of building a mobile first cryptocurrency and kind of having
this stable store of value that you can use from any mobile phone around the world.
Now, and you're at the forefront of two major emerging tech sectors AI blockchain and crypto. What do you love the best or do you love them both equally?
Yeah, I would say ideologically crypto like, and I think because with crypto, it's really it's a very foundational shift in terms of how power structures in the world could work. And it is a sort of like I would say most crypto outcomes, it's sort of the sort of it's the thing where if it works, it really really works in this massive way like bitcoiner ethereum. And I think that that is just a really interesting problem space to
be in. And it feels like the frontier right of like this is really still the very beginnings of the crypto kind of decentralized industry. But I guess there's some really cool math with a so like maybe, yeah, there's like definitely some fun math of AI, and yeah, I mean today a lot of the AI work is like training AI models, so it's almost like DevOps side stuff.
It depends, so yeah, hard choice. I think that you can apply a lot of the principles of crypto, which are kind of like censorship resistance, true like self sovereignty towards AI, both in kind of an open source AI context and a decentralized AI context. And yeah, then that's a lot of fun. So sort of depends if I had to choose one. Yeah, crypto's ideology and like AI's core technology. But yeah, I mean I kind of like a little bit of both, so for sure.
So what do you think about the I guess what you would call symbiotic relationship with crypto, excuse me, blockchain and AI, because AI we're seeing is being used to enhance certain blockchain attributes and crypto attributes, while blockchain is being used to leave AI a bit with deep fakes and things like that, and there's different technologies being built. So do you see them working together as the future as things progress.
Yeah, yeah, I think that's a good way to break it down, right, you sort of have like AI for crypto, so that means like how do you make models available on chain use them for trading and stuff like that, And then you have crypto for AI where you're using the tools of blockchain and for crypto to really advance AI progress. And that's where Vona sits, and so that's where I spend most of my time kind of thinking about the intersection. I think what's really powerful about crypto
is its ability to like mobilize a massive amount of resources. Right, Like if you look at just the number of GPUs that were in the proof of workdays of ethereum set up to mine ethereum, it's like fifty times as much compute power as what was used to train CHATCHYBT. Right, So you have like a huge amount of resources, and AI models requires huge amount of resources to be trained, right. It can cost like ten million dollars of train a model.
And that's just thinking about the compute like the GPU cost, which I has been talked about a lot, and not even yet thinking about the data costs. So I think that kind of using crypto incentives to allow for very large scale projects to be funded in more kind of community oriented, open source type ways is like what's most
interesting to me. And what I would say too, is that's distinct from just sort of like the naive approach is basically Okay, let's put AI on the blockchain, right, but I think it's important to take a step back and be like, why do we even have a blockchain? Okay, it's because we value like censorship, resistance, like true self sovereignty.
And then how do we create a system that uses some of that distributed consensus that's worked really well in blockchains, but applies it towards this AI native world that we're moving towards over the next five to ten years.
So when did the idea come about to start VANA? And what's your mission?
Yeah? Yeah, So VANA was first started actually back at MIT as like a research project super early on. This is like twenty eighteen of like, basically, what I had seen is the only thing that matters for AI models in the long run is data. Like data is really the new oil. It is just this immensely valuable thing.
And so the question I was asking along with my co founder, who I met he was doing his godduate degree at Harvard when I was doing my undergrad at MIT, and so we had taken a class together at the MIT Media Lab. This question was like, how can you leverage the power of like many people rather than just having like centralized institutions training a single model to all collectively contribute. At the time, we were doing like data labeling, where we would have people label data from their phone
super early on and earn based on that. And then I started working on Vona full time a little over like three and a half years ago. That was still quite early in the crypto AI space. In twenty twenty two, we got our patent for kind of non custodial data, which was kind of a way for you to be able to manage your data in a similar way to how you can manage your funds with a crypto wallet, so you can use your meta mask to manage your data.
That's on Vana, and yeah, our mission is to kind of like shift the power of data and AI towards many people rather than just one centralized institution. The economic framing I have it is to ensure that users own
their data and the value that it creates. But that's a little bit abstract, so it's really just about how do you make it so that things that get created from your personal data that you've created you kind of are attributed to you and you earn from an and also allowing AI progress to be pushed forward through that m.
You know, for years, I've been thinking a lot about this and data going on the blockchain and where I can lend it out to advertisers or I could do whatever I want and monetize it. And it sounds like you guys are building that solution. And I think in the world we're headed towards where with digital identity and updates to different advertising laws, GDPR and privacy and all
these things. Do you see that eventually all of us will have wallets, it'll have our digital identity, our data and so forth on a smart contract, locked and whatever it may be, and then we choose how who we want to give it to or lend it out to and so forth.
Exactly. Yeah, you're just pitching VANA right back. So there you go.
But let's talk a bit about the economics of that. So would it be like let's say, Apple compute, right, I don't know, and I'm spitballing here because I don't fully understand all the concepts of this yet. Would it be I give them my data and they pay me a monthly fee or a one time fee, Like, how do you think that might work?
Yeah, that's a good question. And I think that it's sort of been this shift as people have become aware of the immense value that AI creates. Right, Like, if you think sort of five years back, the value proposition of hey, we're going to give you twenty dollars for your data, it's honestly not that interesting. Like to some people it might make sense, but like to the average person, it's just like not that interesting, of like why would
I want that? Whereas if the value prop is like, hey, contribute your data in exchange for ownership in an AI model, right like own the part of the next CHATCHYBT, and it acts much more like equity, like a stake in an AI model, rather than a one time payment. I think that that is much more compelling for users from what we've seen at Vana, and I think that it also points towards like this economic shift that's happening where a huge amount of economic value is created through AI,
right like chat GBT or open ai. They hit three point four billion dollars in revenue this summer, Right, that's just an immense amount of money that's primarily coming from intelligence that is trained on everyone's knowledge, right, everyone who's created data has created that knowledge which has helped teach that AI. And our view is that if you have the right incentive system, you can build much better AI
and also align incentive. So yeah, kind of moving towards like, Okay, what does it look like tactically today, Like the Reddit data daw which is one of the projects that's built on VONAM. They aggregated like one hundred and forty thousand users data they trained an AI model from that. It's
really good at sounding like a snarky redditor. I've heard some people joke it's sort of like the leading shit posting model, right, So like in terms of it's not yet feeding CHATGBT, but like at some point there will be enough data across all these different data sources that it will and every time that model is used, all of the users actually get paid out, right, And so you have kind of this nice flow where, yeah, in the future, it's like every time someone's interacting with an
AI model like chat GBT or claude, the people who have contributed towards teaching it the knowledge that it has are the ones that are rewarded, and so you have this really strong kind of data flywheel in place.
That's so fascinating because you know, you mentioned about the shit posting and so forth. So it made me think of you have these niche AI models, so to speak, backed by communities. They have their own personality, their own take on things. That's so fascinating to me. So is that the world we're we're headed to where it's different AI? Am I artickling it well? Is in models or bots? I don't know what to call it.
Yeah, yeah, no models or boss however you want to say it. That's both of those are perfectly correct. I think that, yeah, the world we're headed towards. Yeah. I think there's sort of this debate right now of like what is truth right because we're trying to all build a single AI model and be like the AI model knows this and this is true. But that's a really hard question, right, Like, actually we have not solved that
as human beings aside from an AI context. Like that is a massive debate that happened in the social media days and now we're kind of having the same thing and in the AI days and our view is just having this healthy ecosystem. I think it aligns actually a lot with the open source work that Meta is doing, just putting AI models out there and allowing many people to sort of modify them, customize them, align them towards a particular set of preferences or views depending on what
they'd like. Right, so you have this balanced, healthy ecosystem rather than one single massive AI model that is controlled by just one or two companies.
So and give me some rope here. Let's say with the Thinking Crypto podcast, I have a community of people. Let's say it's a total is like over two hundred thousand people that listened to me? Right? Can I build an AI model to teach newbies about crypto and my community sources that info to that that model, and then people can pay to have that model teach him crypto. Would that be the economics or business models?
Totally?
Hmm. That is fascinating. I have to talk to you after this, and I.
Think the way to think about that is like, so, think about what AI models can't do today, and then think about the knowledge that you uniquely have or your community uniquely has beyond what's included in AI models today, and that's really where the sort of like alpha is, right of like where what information is contained in the data that most models don't have access to. So generally the Internet you can publicly scrape. That means just like all the sites you can visit without logging into them.
So that's sort of already in a lot of these AI models. But there's a huge amount of information that, like you know and your community knows that not very many other people know, right, So having the ability for you to train an all powerful AI model based on that unique knowledge and have that truly belong to your community.
And I guess, like you said, aside from what's scraped on the web and what's already out there, the nuance of hey, I'm from this part of the United States or this part of the world, I'm thirty something years old, I'm twenty something years old. Here's what my investment story and journey was here, here's how it impacted my life. That's something the AI can pull from it anywhere except someone sharing that story, right totally.
Yeah, that's fascinating I think too. So often the data that's really valuable is this like contextual data on a person, right, So it's like, hey, like here's maybe how I got into crypto, and then also here's kind of my here are my demographics, here's my perspective on the world. Kind of being able to share like both your tweets, but also your financial history, which could include like your trades, your credit card history, and maybe even your journal entries.
And there's really no single tech company that has access to all of that, right, they might have access to a small slice of it, but if you want to get all of those different data sources across a given user, you really need to work with the user directly. And
so that's kind of the power of using decentralization. We can actually leverage data regulation like GDPR which you had mentioned CCPA in California to ensure that a user can get that data out of where whichever platform they've kind of posted it, and then actually contribute it to these powerful at models.
Now let's go back to the redded data DOWT question. I guess first part is why DOOS did decentralize a toon of autonomous organizations? And is that the only way to do this where you can have the crowd sourcing and to have it in a decentralized way.
Yeah, that's a good question, and I think like doos have a pretty mixed reputation in crypto, right, Like, if we look over the past maybe five years of dows, I think there are a few breakout success stories, but as a whole, like, it's actually very hard to get a decentralized group of people to govern something, and so
it can be a challenge. What we've seen with data dows is that often you have a large number of people come in initially and connect their data, and then a smaller percent of them stay and help manage that data and sort of like decide, hey, this company wants to buy access to it, or this machine learning engineer wants to train an AI model on it. Should we
grant them access? What are fair economics, et cetera. And so by default all that is decided by the TAO, because it's very hard to price data, right like, how should you decide whether? So we basically we see the doos as making that decision ultimately. But in the future there are ways too where you don't need to create a DOO and there can just be kind of stable
coin based payments for a particular data pool. And this is a little bit technical, but it's kind of the distinction between what we call a data liquidity pool, so a DLP on vana. Most of them are data DAOs, but they actually don't have to be a DOO if they don't want a token affiliated with their particular data set.
So yeah, we have this slightly more general framing, which is a data liquidity pool, which is just about kind of building up that data liquidity, getting many people's data all pooled together, because your data is only valuable once you combine it with others. Right, you can't train an AI model on just one person's data. You really want a lot of data. And so that's that's a good question, and there is some nuance there.
So the Reddit data dow, how are they monetizing that data, what are they selling it for? And what's the currency being used?
Yeah, yeah, so they created their own currency which is used as a governance. It's basically the DOW kind of token that's used to govern the Reddit data set, and its primary purpose is governance, basically voting on hey, should we let new people come in and contribute their data and reward them, and also should we let someone access
our data? So one of the folks they granted access to the data is actually like ex Google Brain machine learning engineer who trained an AI model that like ship posting model that I had mentioned, so that it's in the phase of like research preview right now and then they're rolling it out and to interact with that, you'll have to actually sort of like hold a minimum amount
of that Reddit data that governance token. They're also sort of finalizing a data sale in the works right now, and so the way I'm actually not sure exactly how they structured those economics. It could be that they're having the data buyer burn the governance token, or it could be that they're distributing revenue to the governance tokenholders. So it could be either of those two, and it depends on what the two decided.
A couple questions of follow questions on that. Do you have to create your own token or can you use an existing token or a stable coin as you mentioned before, Yeah, because if you get into the creation of a token, you know, could you have securities issues, security laws issues, I should say.
Yeah, yeah, you can use a stable coin, you can use a pre existing token, and so you can kind of collectivize however you'd like. And I'll mention too the Reddit data now actually only intended for their token to be used for kind of like governance and voting, and so that was their intention from the beginning. I think there was some trading activity that came up on it, but it's primarily used for voting, so I think in the long run we'll see sort of some activities like that.
And yeah, depending on kind of your risk profile and how you structure it, right, each daya now has their own preferences in terms of what they're looking for from a governance perspective, from a revenue distribution perspective, and so they can structure their own token. It's similar to like an ERC twenty token, right, so you can kind of modify it however you'd like.
So, are you primarily using eth to build these dows or are you using all the block chains as well?
Yeah, that's a good question. So we actually started as basically we were like, okay, let's start and let's deploy this too like ETH or an L two and sort
of see what happens. What we realize is that like a lot of the optimizations that people have made for primarily financial transactions don't work that well in a vana context, which are data transactions, right, So on VONA, basically all the transactions are like permissioning out data or interacting with a data token, or interacting with an AI model, and
that's a very different set of constraints. You care a lot about low fees, you care a lot about pretty fast block times, and then you also have data regulation constraints. So actually, if we were to run our own kind of like sequencer as an L two, that would be regulated as a data processor. And so we realize like, actually we need a standalone set of VANA nodes that are basically operating this kind of like data transaction network. And so yeah, VONA is it is a layer one
blockchain that's designed for private data. It's fully em compatible, and I see a world in the future where everything is kind of cross chain, right, so you have all these different bridges and it's really easy to kind of like be interoperable across different chains. But yeah, for data regulation reasons, we actually had to land on a standalone
layer one for private data. And the tokenomics look pretty similar to ethereum, where essentially there is kind of like a native gas fee that's paid for the different data transactions, and then the top sixteen data dows actually earn a share of block rewards because they're kind of onboarding data into the network, and so similar to I'm not sure how familiar you are with bit tensors kind of model. They emit rewards to the top thirty two. I think
maybe now they're up to like forty six or something. Subnets. Yeah, the Vona protocol admits rewards to the top sixteen data dows, and that's also kind of governable as more top data dows emerge.
That's fascinating. So let's say I create that now for my podcast? Right? Can I use the Vona token as as as the currency sort of thing.
Yeah, Yeah, you can choose to use, yeah, the Vana native token, which is similarly just sort of like the gas token on main net, which isn't yet launched that will launch later this year. You also could use like a stable coin, or you can use, yeah, an e ARC twenty like token, depending on what you're looking for.
M You and I have a lot to talk about offline, but the steam engine is going here. Question why would someone want and I know this is a question people listening and watching, what would have Why would someone want to buy that AI model from Reddit? And and I understand certain use cases that you know, certain companies may create AI models to do different things, but why would they want to buy that Reddit AI model?
Yeah, so the Reddit one, I believe that the Reddit Data NOWS model is being used. Okay, it's a company that's kind of building like crypto Reddit, And what they're doing is they're allowing you to simulate, Hey, if you post this comment or you post this post, what are how are people going to respond?
Right?
What are people going to say? And so they're actually using the AI model to simulate that, so that before you post something, you're like, hey, this is probably how people are going to react to it. And so yeah, for that particular case, that's how one of the models is used. There's another model that I think they're creating, which is predicting how many up votes the posts would get, right,
and then you can kind of optimize for that. So I think there are many different applications beyond just sort of your standard chatbot. Like it it's not that useful to chat with just a chatbot. That's good as you're posting, right, Like, I think we're on the same page there, and so sometimes it's these like industry specific niches where people are finding a lot of value for the AI model.
I would love to test that out because I want to drive more engagement. I want my tweets, my posts, whatever to drive the I want to get more of vote. It's on Reddit, right, So that makes sense? Okay?
Yeah? Yeah, I mean I think too, Like if you look if you use like chatubt and claude and you're like, hey, like modify this tweet or whatever, and it adds hashtags, and it just sounds so AI generated, right, and you're like, no, like that's not what it should be.
Right.
So you need a trained from kind of higher quality data, right, like actual real data, like the organic good stuff that you can get directly from people, and then you can create better AI models that can serve that purpose.
So are you working with any let's put it this way, real world companies like the Apples or McDonald's or whoever, right, that are trying to build these models.
Yeah. So there's a Fortune five hundred fashion company that we've worked with really closely basically doing kind of like consumer insights type prediction. So their model is focused on how do you predict what's going to be in style in three to six months using like cultural data about fashion and music, And so I think those sorts of
use cases are really powerful. The ones that I'm most excited about that I think are going to grow really exponentially are these new kinds of entrepreneurs Like the guy who started the LinkedIn Data Out. He actually just raised money for the LinkedIn Data Out and he used to sell data at LinkedIn and really deeply understands the value of that data set. He's selling it to sort of
like Web three native LinkedIn type competitor. And so there are all these different use cases of like, hey, what if as a business you could just like access all the data that today is stuck in these web two wallt gardens. But there are also kind of more traditional companies like the Fortune five hundred fashion company I mentioned, who are basically just trying to get data that they can't get access to otherwise and train AI models on that.
What are the laws around that if any or is this just a business model you can create because I know that you know you have the data privacy stuff, but then it's going through the AI. So how do you once I guess sorry collecting my thoughts first, how do you secure the data that's being put into the AI model. And two, are there any laws that govern these things?
Yeah, so there are a lot of laws that cover and data privacy, which generally I think are actually very strong laws. I think a lot of people don't realize just how powerful the laws are and actually how much ownership they have over their data. Right, So when you use a platform, it's sort of like when you park your car in a parking lot, Like the parking lot
doesn't own your car. It's still definitely your car, and they can't just like randomly take parts out of your car like it is yours, you get it back, you can do what you want with it. And so data regulation like GDPR and CCPA and just basically the laws view on data is that you fully own your data. When you use a platform and you check that like terms of service box, you're granting a very permissive license to them to be able to use your data for
different things. They generally cannot train AI models on your d anonymized data. They can train it on your anonymized data, so they can like, yeah, remove personal identifiers, but that ends up being kind of not as good of a data set to train on In Avana context because we work with what we call non custodial data, which is where like a user has full control over their data, they can grant whatever legal rights they would want for someone to do something to their data. You're actually able
to give ai researchers access to de anonymized data. But then you have the question of, like, well, how do you ensure privacy, right, because you want to make sure that if you're contributing data. And it's both from the perspective of like I, as ANNA, don't want my data to leak, and it's also from the perspective of with economics, like okay, if something is public, it's really hard to charge someone money for it, right, So it's both about
protecting one's privacy but also about protecting one's economics. Right. So if I in the future have like ai ana, who can kind of autonomously act on my behalf or revenue, et cetera, like, I want to fully own that. I don't want it to just be like open source and randomly out there, like I want the full rights to it. So that's why that kind of privacy piece is important. We have the distinction between renting data and selling data.
So by default, you are just renting your data where you're allowing an AI model to be trained in a privacy preserving way on your data, but you're never letting the data leave the system, so your data sort of stays in this secure compute environment. And that's actually one of my old professors who's been really helpful and kind of the monitoring invented like privacy preserving machine learning ten years ago at MIT, and so a lot of his kind of work had he was really pioneering a lot
of this. But now you're able to rely on like a secure computing environment to train these models and protect the model ways, so that you're protecting the economics and the privacy of the data.
It's fascinating. I feel like I can talk about this for hours, but I know we got to talk about some other stuff. You guys also got some funding. I believe Paradigm is one of your investors. Tell us a bit about that.
Yeah, yeah, So poly Chain let our initial round and then Paradigm led our most recent rounds. So we've raised
about twenty million dollars a day. I think that standing up like a new network for user own private data is like is quite an undertaking, like we've been building for years, and I think that often, Yeah, you need to basically be willing to make a pretty big investment from a technical perspective in order to deliver like an end user product that's actually usable, right, Like, Okay, one of our guide and goals is how do you get one hundred million people to all contribute their data and
train like a user owned AI foundation model that actually beats something like GPT six So a future version of chat GPT in performance. And to get one hundred million people to contribute their data is like quite an underdate. We're like a little over one percent of the way there, right, But it takes time, right, and it also takes like empowering like builders to create really good experiences that bring
users on board. But yeah, we chose to work with Paradigm because honestly, they're really close to the MIT folks,
like the original MIT Bitcoin Club. I think they've invested in like sixty percent of our companies and so yeah, the just sort of like a lot of MIT overlap, and I think they really see the long term potential of what crypto can be, right, Like they're thinking on a ten twenty year time horizon, which I think is so important when building in this industry that just has these massive cycles that kind of go up and down, Like you really need to keep that sort of long
term view of like, hey, like what should the Internet look like in ten years? Like what should data look like in ten years? And what role will kind of on a play in that in order to build like a long lasting protocol that truly shifts the way that people think about their data.
Now a quick follow a question on that, what do you guys charge for companies and people to build these things? And on the flip side, what are you seeing people selling their AI model data for?
Yeah? Yeah, So I'll give one example of like the Fortune five hundred fashion company that I mentioned, So they're paying users and so you can structure the economics however you like, right, So you could have it be like equity and a stake in an AI model, which I think in the long run is going to be the majority of how people interact. Some people choose to use
more of like a stable coin payment today. So five hundred fashion company, I mean they don't touch tokens like that's you know, like they have a hat a log of dogin for this, so they have, Yeah, they pay people twenty five dollars for their fashion data, their Spotify data, and their demographic data. And that's pretty high, right, Like, if you were just selling your Spotify data, it might
be worth like ten cents. But because you're combining the data from across different sources, they're able to use that for forecasting of like, hey, based on the music you listen to, what's going to be in style in like three to six months, and that's super super powerful. So that's kind of one example from just like a pure data yeah sales view. In terms of Vana's economics, so we charge a small like the protocol charges a small data transaction fee. It's like less than a cent on
test net right now. It's similar to Ethereum, where basically as volume goes up, like fees come up, but we've architected it to try to keep fees really as low
as possible. And then by default, Vana is the currency for data sales, right, So if you are actually buying data on if you're buying a quarter million dollars a million dollars of data on the platform, you are buying the von a native token and then distributing that to the particular data daut that you want to offer data to, whether that's in the form of buying their token burning it.
And the reason why there are all these different tokens in the system is that you really need quite flexible economics to work with data, right Like, if I were to ask you, like, hey, how much is your Amazon
purchase data worth relative to your Spotify data? Like that, that's possibly an unsolvable math question, right like, And so we've basically pushed it to Hey, you have these data set specific tokens which allow for both data DAWs and flexible economics, right of, like, Hey, each data dou can choose how much they want to charge for their day, how much they want to yeah, what they would say yes to, what they would say no to, And so you get a lot of that flexibility.
And I apologize, I'm thinking a lot about the things you're saying and how I can leverage it. So I guess I'm thinking about let's say I set up that model. I can pay my viewers and listeners to contribute to the data set, and they can also be stakeholders. So the revenue that comes in when we sell the or lend the model out, whatever it is they get a cut of the.
Revenue exactly, Yeah.
And how are the payments? Is it done automatically like if you use stable coins.
Yeah, yeah, it's sort of all in smart contracts. So anything you can express in solidity. We've seen some people use viper, which is just like the Python way to write smart contracts, because some of them are more like AI people. And so yeah, anything you can express in solidity, you can kind of used to distribute payments and kind of reward the contributors who kind of put their knowledge
towards making the AI model more capable. And then one thing I'll mention too is one important concept within VANA is proof of contribution, just basically measuring, Hey, for a given piece of knowledge or data that comes in to help an AI model, how much is that actually helping, right, And so you can basically measure, hey, how much new information new knowledge is contained in this data point and then reward them proportionally to that. And that is implemented
at the data set and AI model level. So there's a different one for Reddit, for LinkedIn, for your AI model. But yeah, I mean I think also the best way to understand a system is yet to think about how you would fit into it. So I think these are great questions to be asking.
And what does the interface look like for people to submit that data? Is it like a gouge interface app? What does that look like?
Yeah, so Vona as a network is like as a it's a period of peer protocol similar to ethereum. Right, So like many like you can interface with a network through like a command line terminal that's like the raw way to interface, so that most people don't do that, right,
Most people use like an interface. So you can add it to MetaMask and then you can basically see your kind of like data financial profile in MetaMask, and then there are a couple of different UIs that people have built out, so like the Reddit data down they built their own UI so you can contribute your Reddit data.
There's another project that has this automatic data scraper where they sort of go and get your data automatically for you, and then they're working on kind of adding away so that you can contribute that back to the Vona network. So yeah, you can use like a web It just looks like a website like an app, which is sort of like traditional DApp, or you can kind of like open things up and get into it if you want
to touch the code. But yeah, the core network is a peer to peer network, so there are lots of different UIs on top of it.
Interesting, what's on your twenty twenty four rown map or the remainder of twenty twenty four.
Yeah, so a lot of our focus right now is really supporting the dataaps that are kind of being built on Vana and then getting ready for main net launch and so kind of going through the audit process, going through kind of like load testing and all sorts of other kind of like preparations from that front, and then yeah, so main net will be stood up by a set of independent validators who are all kind of like choosing to run the Vona network from scratch, so stewarding that
launch process. It is. Yeah, it's sort of basically standing up the new blockchain from scratch and supporting these like data now applications on top of it.
Well, you and are going to talk offline about what I want to do, but I want to get your take on with the limited time we have on the aggress of the crypto market. Obviously, you started out in blockchaining crypto working at Celo. What are your thoughts on how the market is matured. We got ETFs, now crypto regulations seemed like they're on the cusps, we're getting past. You got politicians you know, supporting crypto. What are your thoughts on the growth?
Yeah, I mean I think that broadly, we've seen this shift where bitcoin is pretty much like mainstream accepted as like a store of value, right, And if we look back nine or ten years, like that's kind of crazy that people are now just accepting like, yeah, it's bitcoin, Like it's like a reasonable store of value that has like predictable kind of like inflation and other properties people value in terms of like having a store of value.
So I think that that is an oudible shift. It's really easy to forget, right when you go through these kind of like bull and barcycles, you're like, oh, like bitcoin is it? But it's like okay, Like if we look over a long time horizon, like this is doing amazing. I think it's also led to some really cool shifts around how people think about kind of like true ownership and control. I think crypto has yet to deliver on
this promise of like owning your data, right. That is, like I think the promise that actually got a lot of people into crypto super early on, but we're not there yet, Like Vana is working to make sure that we get there over a several year time horizon, but I think we're getting closer and closer to mainstream adoption.
I think that there are a lot of kind of emergent use cases that we're seeing related to yeah, kind of moving beyond just like stable coins or like stable means of payment, which are more like decentralized applications, right, Like I think Farcaster is an example of like they're basically using sovereignty like that that's the core proposition they have, it's not financial, and I think those are quite cool. So yeah, generally super optimistic, like we're still so early
in building so much of this out. And I think another good sort of Yeah, something I like to track is just like where are smart people spending their time? And I think a lot of my like smart friends are spending time in decentralized AI, and that makes me think, like, Okay, there's a lot of growth here. There are a lot of like there's a lot of talent basically going towards solving some of these key problems. And so that's another kind of internal metric at track and there's some cool
stuff happening in like DSi as well. I think more smart people are starting to spend time there figuring out Yeah, some of the decentralized science questions, do you.
Think AI can help defind to be more secure and we avoid some of these exploits and things we've had over the years. And I know one version one point oh of anything is going to have flaws, right, But I haven't really touched a lot of DeFi because of all these hacks and all these exploits. But do you think AI could help solve some of those problems?
Yeah, I think that there's a lot Like if you think about AI is almost giving you knowledge on demand. There's so many different ways you can use that, right, like having knowledge on demand to like do engineering work or do code auditing work. It seems plausible to say, yes, that could help reduce exploits, Like I would say, I haven't spent enough time thinking about it because there are
a couple of different variables, right. You also could have more bad actors on the AI side, so I like would have to kind of like reason through that a little bit more, and I haven't thought about it in depth. But yeah, generally, when you have like knowledge on demand, like it really changes the economics of a system and kind of what's possible where value accruise. We think it accrues in data and how markets play out.
Yeah, yeah, as you're saying now, I'm thinking of AI wars AI models fighting each other, trying to one's trying to police, one trying to attack, and you know, hack things.
Yeah, it's a crazy world out there, you know, Like I think a lot of that's already going on and we don't even we might not even realize it. But yeah, I want like an AI and a agent who kind of like defends my interests and make sure but all.
As well for sure, And I know I kept you a bit over time, so I got some wrap up questions there for you. First, If you could create your own metaverse, what would the theme be?
Oh, I see myself as kind of already living in this like digital world primarily, so it's it's themed around data sovereignty and data ownership and kind of like some research related topics.
Yeah, yeah, that makes sense. I figured it was going to be like something related to what you're already doing. Rapid fire questions. Favorite food.
Probably green tea. I guess that's a drink, but it's just consistent.
Yeah, favorite musician or band.
I actually didn't start listening to music until a few years ago, so I probably like like a major laser workout playlist that I just work out to.
Yeah, favorite movie, let's.
See, you know what. I actually I only watch movies on airplanes, so you know, I actually, okay, I'm actually a huge fan of mister Beast. I think he's a really interesting entrepreneur, So he's in the video category. I would say, like mister Beast and sort of the arc that he's pulled off from an attention perspective.
Yeah, and favorite book.
Favorite book, Actually, I love reading. I think David Sedaris is this amazing writer. He's like a humor comedy typewriter, but he's a really good storyteller and he has this like arc of stories that always like comes back in just the right way. And so yeah, he's like one of my favorite writers.
When you're not working at Vanna, what are you doing for fun?
Probably like biking or skiing, maybe like some I don't get to write that much code anymore in Havana context, and so I still try to just like explore, especially some of the new AI stuff when I have some time on the weekends.
And a pleasure chatting with you. I'm really really fascinated and excited for the work Van is doing, and we're going to talk more. But thank you so much for joining me.
Yeah, thanks so much for the questions.