.webp)
Today at Permissionless III, we are thrilled to welcome Anna Kazlauskas, a visionary entrepreneur and tech innovator with a remarkable track record in data empowerment and blockchain technology. As the co-founder of Vana, Anna is on a mission to help users take control of their data and harness the value it creates, fostering a more equitable digital landscape. Prior to her work at Vana, Anna was an early core engineer at the Celo Foundation, where she played a pivotal role in adapting the Celo blockchain for mobile devices, ensuring that essential financial infrastructure is accessible to the billions of people who need it most.
Anna's journey into technology began at the Massachusetts Institute of Technology, where she earned a degree in Computer Science and Economics while mining Ethereum in her dorm room. Her academic pursuits also included notable research at CSAIL and contributions to various central banks and international organizations, including the World Bank and the Federal Reserve.
With experience as the co-founder and CTO of Iambiq—a YC-backed startup that leveraged machine learning to automate document sorting—Anna is not only a technologist but also a problem-solver dedicated to driving impactful change through innovation. Join us as we delve into Anna's insights on data ownership, the future of blockchain, and her journey as a female leader in the tech space.
Anna: I'm Anna, co-founder of Vana. Vana is building user-owned AI, which is AI that's collectively owned by everyone who has contributed to it by contributing their data. We're exploring how to create a foundation model that outperforms some of the best performing models like GPT, ChatGPT, or GPT-6 in the future, by getting 100 million people who all contribute their data to create a better AI model.
James (CryptoBit Mag): Interesting. Is the platform currently running?
Anna: Yes, it is. We have one data DAO that's live, while all the other ones are in testnet right now. The Reddit data DAO, which launched a few months ago, saw about 140,000 users join. Close to a million people signed up with their wallets, and 140,000 real Reddit users actually contributed their data. What's cool is that this actually led to Reddit changing some of their data policies to be more researcher-friendly.
You can think of a data DAO or these pooled data sets as acting like a union. They can use it for bargaining power, like negotiating with Reddit to change some of their policies. You can also use it to train AI models. The Reddit DAO trained an AI model that, since it's just on Reddit data, is really good at shitposting. It's now the leading model at shitposting, and you can go play with that model.
We have around 200 data DAOs on the testnet. One exciting project is the 23andMe data DAO, which is actually trying to buy 23andMe the company outright. 23andMe's market cap is around $150 million, their whole board quit, and they sit on all this data that people don't necessarily want to have sold.
James (CryptoBit Mag): So this is really on the blockchain, but also utilizing some Web2 infrastructure as well?
Anna: Exactly. Vana itself is a standalone layer-one blockchain designed to work with private data. It makes data non-custodial, meaning your private keys permission your data, just like with cryptocurrency. This is important because if a centralized company goes to Reddit or 23andMe and asks for user data, they'll refuse. But because users legally own all of their data, they can request a copy and contribute it to one of these pools within Vana, these data liquidity pools.
This decentralization is crucial. When the Reddit DAO hit 100,000 users, Reddit tried to shut the project down, claiming it violated their API terms. But because of Vana's non-custodial data approach, where users process their own data, the blockchain can handle it. This challenges the relationship between users and platforms like Reddit, which is enforced by data regulation stating that users actually own this data.
James (CryptoBit Mag): Is there a lot of competition in the marketplace, or do you bring something completely different to the table?
Anna: Right now, we bring something completely different. One historic project that comes to mind is Ocean Protocol. They focused on letting businesses take their data, bring it on-chain, and sell it. But if a business already wants to sell its data, the blockchain isn't enabling that much. We're starting with end users, telling them they own something they probably didn't even realize they owned. I think this is a much bigger opportunity.
A lot of the builders creating these data DAOs are from the Bittensor ecosystem, so they've built out early subnets. There's a fair amount of overlap in terms of the community we've seen.
James (CryptoBit Mag): How long has this project been around?
Anna: About three years. We started with the goal of having users own their data and the value it creates through AI, which was a very abstract concept in 2021. Early on, it was about data labeling and on-chain payments. Then we realized that data collection – having people contribute data that already exists on a platform – was where the market was constrained on the supply side.
It's very hard to convince people to contribute their data to AI models. This is different from data labeling, where you can have anyone label data for $2 an hour. It's much harder to convince an everyday person to sell their message data for $20 unless there are strong privacy guarantees. But because we're offering collective ownership in an AI model rather than just a one-time cash payment, we've seen a shift where more people are willing to contribute their data.
Recent milestones for us include announcing a round led by Coinbase Ventures, and then Paradigm led our Series A. We've brought together really great, trusted backers in the industry to build this fully user-owned AI system through data.
James (CryptoBit Mag): That's incredible, especially to get to Series A in this space. What's your background before working with this?
Anna: I initially got into crypto through an interest in traditional currency. I was obsessed with central banks – I even had a picture of Janet Yellen in my high school bedroom and worked at the Fed in high school. I loved currency because I saw it as central to the whole economy.
At MIT, I learned about decentralized currencies. I was part of a five-person MIT Bitcoin club of super ideological, almost anarchist people. It was around the time Ethereum was launching, and I got into crypto through mining Ethereum with GPUs I'd get from dumpsters that MIT labs would discard. I set them up in milk crates for better ventilation.
I loved the fact that I was running what I saw as a decentralized central bank from my dorm room. I became obsessed with the idea of decentralization and using cryptography to shift control from one entity to a large group of people, achieving true censorship resistance and sovereignty.
At MIT, I also got interested in AI through research at CSAIL, MIT's AI Lab. I was coming at it from an economic perspective, wanting better, more complex econometric models to model data. Then I started seeing some of the deep learning developments, like the "Attention is All You Need" paper in 2017.
What I realized is that data is ultimately the only thing that matters for AI models – an AI model is only as good as its data. So the question became: how do you get the largest data treasury available? It's almost like a central bank for data, but instead of being centralized, it's owned by everyone. That's how we found this intersection of decentralization tools and AI progress acceleration.
My co-founder, Art, is a lawyer by training. He was in grad school at Harvard when I was doing my undergrad. He previously worked in data sales, selling data to big tech companies at Appen, which works with large tech companies for their data collection needs.
James (CryptoBit Mag): How big is your company right now?
Anna: We're a 16-person team. We've kept the team pretty small, but the team of people building data DAOs is probably around 50 developers across all of them. Our view is that if you're building a protocol that you ultimately want to be decentralized, you need to ensure the ecosystem can flourish organically. In the current crypto dynamics, there are incentives to raise more private venture funding and build massive teams, but we don't think that works well for decentralized protocols, at least in Vana's case.
We've taken the approach of keeping our core team small while supporting various ecosystem players. There's a company called DLP Labs that's focused on building data DAOs on Vana, and we support a few other ecosystem companies.
James (CryptoBit Mag): Is there anything else you'd like to share?
Anna: One question that often comes up is: what is decentralized AI? Most people are familiar with centralized AI, like OpenAI and ChatGPT, where one company trains an AI model and keeps it private. This lets them charge money for it, which is necessary because training these models is very expensive.
The alternative is open-source AI, like Stable Diffusion or projects by Stability AI, where you train an AI model and release it openly. The challenge here is that it's hard to charge money for something that's open, making it difficult to sustain a business and keep improving the model.
Decentralized AI offers the best of both worlds. It's like open-source AI because it's controlled by many people, but it has the benefits of centralized AI because you can monetize the models. You get many people contributing to an AI model while still having the ability to monetize it. That's a lot of what's happening in decentralized AI today, and it's a really exciting space to be part of as these new primitives are forming.
James (CryptoBit Mag): That's really interesting. Is there anywhere people can find you? Social media, website, URL?
Anna: On Twitter, it's @withVana, and our website is vana.com. There, you'll find all the documentation and details, and you'll be able to link out to many of the different data DAOs and explore applications in the ecosystem.
James (CryptoBit Mag): Perfect. Thanks very much, it was great meeting you.
Anna: Great meeting you too. It's exciting to see someone interested in AI and making it happen. You know about all the different Fed chairman – Ben Bernanke, Paul Volcker... I think data is becoming the new currency, and so currency itself is becoming a bit less important. We're moving beyond currency in a way.
All interview transcripts have been meticulously transcribed by Stylo.Network, ensuring accuracy and clarity in every detail.
You can follow Vana through their Website, X.com, Github, and Discord.
Exploring the Innovations of Vana: A Deep Dive into Decentralized Data Ownership
Have you ever wondered where your digital footprints lead? In an era where data is the new oil, Vana.org stands out by transforming how we interact with our digital information. Here's a look at what makes Vana an intriguing player in the realm of decentralized technology.
Vana introduces a paradigm where you, the user, retain full ownership and control over your data. This isn't just about privacy; it's about empowering users to gain from the AI models trained on their data. Imagine contributing to a global AI project and actually owning a part of the outcome. That's Vana's vision.
The platform breaks down the traditional "walled gardens" of data. Users can contribute data from various platforms (think your posts on social media, professional networks, etc.) towards training AI models. This approach not only fosters more personalized AI but does so in an ethical manner, ensuring that the AI development process benefits from diverse and user-consented data sets.
Vana leverages blockchain technology to ensure that your data remains non-custodial. By logging in with Web3 wallets, users maintain control over their data's usage, significantly reducing the risks associated with centralized data storage. This privacy-first approach not only protects your digital identity but also allows for monetization of your data.
Here's where Vana gets even more interesting. Through Data DAOs (Decentralized Autonomous Organizations), communities can collectively decide how data is utilized and shared. Every piece of data can become a token, tradable and traceable via smart contracts. This tokenization ensures transparency and fair compensation for data usage.
Vana's technology isn't just for the tech-savvy. Its projects aim to democratize access to AI technology, allowing individuals worldwide to reclaim and profit from their data. This could mean a farmer in a remote village or a blogger in a bustling city; everyone gets a chance to contribute and benefit from the AI revolution.
Designed as an EVM-compatible Layer 1 blockchain, Vana isn't just another platform; it's a foundational layer for data liquidity. It's built to support private, user-owned data, making it a robust base for developers interested in creating applications where data privacy and user ownership are paramount.
Vana.org isn't just changing the game; it's creating a new playground where data privacy, user empowerment, and AI innovation intersect. By leveraging blockchain's inherent features, Vana promises a future where your data works for you, not just for big tech companies. Whether you're deeply involved in the tech world or just a casual observer, Vana's approach to data might just redefine what you expect from your digital interactions.
As we move forward, keep an eye on platforms like Vana. They're not just participants in the digital economy; they're potential architects of its future.
An overview of what Vana.org is, its mission, and its core services around decentralized data ownership and AI.
Details on how Vana uses blockchain technology to maintain user data privacy and control.
Information on how users can potentially earn from their data through the platform's tokenization and Data DAOs.
Explanation of Data DAOs, their function, and user involvement in decision-making processes.
Insight into the types of AI models that can benefit from the data contributed by users on Vana.
Step-by-step guide on how users can contribute their data to the platform.
Information on any fees or costs associated with using Vana's services or contributing data.
An explanation of Ethereum Virtual Machine compatibility and its implications for developers and users.
Details for developers interested in leveraging Vana's infrastructure for building applications where data ownership is key.
Insights into upcoming features, expansions, or strategic directions Vana might be planning.
DISCLOSURE
CryptoBit Magazine is a leading media outlet dedicated to covering the cryptocurrency industry. Our journalists adhere to a strict set of editorial policies to maintain the highest standards of integrity and accuracy. CryptoBit Magazine operates with complete editorial independence, ensuring unbiased and trustworthy reporting on all aspects of the crypto world.
Nothing on this site should be considered financial advice. Always conduct your own research and consult with a professional before making any financial decisions.
