You are swimming in the data. You create new data day -day. If your health app counts your steps? That’s the new data. The ring oura that monitors your bio-metrics? Important data. Your social media posts, even the stupid jokes that have got zero? More data.
This is all the data that AI companies want to harvest. You can’t build great AI without great data, which is why many have viewed data as “new oil ‘in the race for AI. The problem, though, while your data is important In theoryThe fact is that it is difficult to make your own personal data money, as you have no action as an individual. (Open AI doesn’t knock on your door to buy your old tweets.)
Come in Old. “I think data is this basic resource that empowers the next generation of AI, and really the next generation of our digital economy,” said Anna Kazlauskas, co-founder of Vana and CEO of Open Data Labs. “Many people are openly not only realized that they really do have their data.”
But you have your data. And it’s important … If you can join forces with millions of others who also have their data. It will give you the power of bargaining. And that’s the vana mission: to create an ecosystem for user-owned data, which is a user-owned AI-owned AI.
That ecosystem involves a mixture of data daos (a “union of manufacture” for data), decentralized data markets, launched recently VRC-20 tokenand a new one Collaboration with flower lab to form the first world-owned Foundational model. (Show a decentralized AI creeping into the mainstream: Vana/flower cooperation is covered by Wired.)
Kazlauskas will give a keynote to AI Summit in Consensus 2025 This vision is described, and he gives a glimpse of it. And he sees the momentum shift. “We are starting to see this change in which many people realize that, ‘My data is really important to AI’ and ‘I really have it.'” He guess that in a few years, more than 100 million users have been above. In 10 years? “World’s population. More than 10 billion.”
The interview is that -condemned and lightly edited for clarity.
Why is user-owned data so important to you?
Anna Kazlauskas: Most people assume the data is owned by platforms sitting, but that’s not the case. In the same way that when you put your car in a parking lot, the parking lot does not own your car. You can restore it. You have the whole person here.
And there is a huge amount of money made today, mostly by large tech companies, from that data, but users are legal -owned. So I think it’s important that we restore that owner, both from a user’s view and from a developer’s perspective.
Can you connect the dots to how it helps developers?
As a developer, especially in an AI world, having access to the right data is really important. And it’s super hard to do now, because most of the data is locked inside the big tech garden walls. So many of my real smart friends do things -AI’s work is working on Big Labs, because there is data and where the compute is. But that doesn’t have to happen.
How does the DAO data applies to this vision?
So a Datadao is kind of such a manufacture union for data. Where you really have a large group of people with their data together, then can make collective decisions on what’s going on with that data.
The reason that is important is your data, in itself, isn’t it worth it, is it? This is more useful when there is a large pool of it. When there is enough to train an AI model.
What are some of the Daos data that you are excited about?
There are some in the health space that is really interesting. There is an early one who actually makes the full export of medical patient notes, which I think will really help promote a lot of expanse research. There are some associated with biometrics, sleep, and health. There is one with the DLP (Driver Loyalty Program) Labs; They build car data. And within their data-set, Tesla data is really interesting because most people are thinking about Tesla as important because they have a data lead, right? Actually, users can get a lot of data-set.
You pivoting from theory to practice with the new collaboration with flower labs to produce collective-1. What is the goal there?
Collective-1 is the first user of the user’s property. Usually when people think about a foundation model, they usually think of a company that runs a massive training job at a single data center, right? Like Openai. And the reason why this is usually done in a centralized way is because it requires, one, a whole lot of compute power, and two, a whole lot of data.
Flower AI is a type of leader in Federated (Decentralized) training. They have done a really good job in building great resource libraries. They enter the edge of the training and the part of the algorithm. And with Vana, we really focus on that piece of data, right? So we usually have all this data that people can train. You will then give users to the end of the model’s ownership, and users can decide what the model allows? So this is the first foundation model of its type.
And the theory is that eventually, with better data, you can develop AI that is not just Competitive with the middle players but BetterIs that right? So it’s not just about ideology, but also performance.
Exactly, yes that’s 100% right. From a decentralized context, I think people often agree with the principle, “Yes, we must have the AI that people belong to. We must have decentralized AI.” But what is the thing we can really do better in a decentralized context? The data is the answer. For each company, they only have a single cut of a set of data. Apple got their data. Google got their data. But if you go to the user, you can cut the platforms and actually build better data sets than any single company. Data is the secret sauce that makes it all.
Love it. Thanks Anna, see you at AI Summit in Toronto.
Jeff Wilser will Mago -Host at AI Summit in Consensus 2025, and host of THE PEOPLE’S AI: The Decentralized AI Podcast.
