You’re swimming in information. You’re creating new information on daily basis. In case your well being app counts your steps? That’s new information. The Oura ring that’s monitoring your bio-metrics? Priceless information. Your social media posts, even the silly jokes that obtained zero likes? Extra information.
That is all information that AI corporations would love to reap. You possibly can’t construct good AI with out good information, which is why many view information because the “new oil’ within the race for AI. The issue, although, is that whereas your information is efficacious in idea, the truth is that it’s onerous to monetize your personal private information, as you don’t have any leverage as a person. (Open AI isn’t knocking at your door to purchase your previous tweets.)
Enter Vana. “I feel information is that this elementary useful resource powering the following technology of AI, and actually the following technology of our digital financial system,” says Anna Kazlauskas, co-founder of Vana and CEO of Open Information Labs. “Lots of people frankly simply do not understand that they really personal their information.”
However you do personal your information. And it’s worthwhile… in the event you can by some means be part of forces with thousands and thousands of others who additionally personal their information. This is able to offer you bargaining energy. And that’s the mission of Vana: To create an ecosystem for user-owned information, which in flip fuels user-owned AI.
That ecosystem entails a mixture of Information DAOs (a “labor union” for information), decentralized information marketplaces, the not too long ago launched VRC-20 token, and a brand new collaboration with Flower Labs to construct the world’s first user-owned foundational mannequin. (Exhibit A that Decentralized AI is creeping into the mainstream: The Vana/Flower collaboration was coated by WIRED.)
Kazlauskas will give a keynote on the AI Summit at Consensus 2025 outlining this imaginative and prescient, and she or he offers a glimpse right here. And he or she sees the momentum shifting. “We’re already beginning to see this shift the place extra individuals notice that, ‘My information is absolutely vital to AI’ and ‘I’m truly the proprietor of that.’” She predicts that in just a few years, over 100 million customers will likely be onboard. In 10 years? “World inhabitants. Above 10 billion.”
Interview has been condensed and evenly edited for readability.
Why is user-owned information so vital to you?
Anna Kazlauskas: Most individuals assume information is owned by the platforms that it is sitting on, however that is not the case. In the identical method that once you put your automobile in a parking zone, the parking zone does not personal your automobile. You possibly can all the time take it again. You will have full possession over it.
And there is a big sum of money being made in the present day, largely by huge tech corporations, off of that information, however customers are the authorized homeowners. So I feel it is vital that we restore that possession, each from a person perspective and from a developer’s perspective.
Are you able to join the dots of how this helps builders?
As a developer, particularly in an AI world, getting access to the best information is absolutely vital. And it is tremendous onerous to do proper now, as a result of a lot of the information is locked up inside the walled gardens of massive tech. So lots of my actually sensible mates who do stuff in AI go work on the huge labs, as a result of that is the place the information is and that’s the place the compute is. However that does not need to be the case.
How do Information DAOs match into this imaginative and prescient precisely?
So a DataDAO is type of like a labor union for information. The place principally you’ve a big group of people that pool their information collectively, after which could make collective selections over what occurs to that information.
The rationale why that is vital is that your information, by itself, isn’t that helpful, proper? It is far more helpful when there is a huge pool of it. When there’s sufficient of it to coach an AI mannequin.
What are a number of the Information DAOs you’re most excited by?
There are just a few within the well being house which can be actually fascinating. There’s an early one which’s truly doing full exports of affected person medical information, which I feel can actually assist advance lots of analysis within the house. There’s some associated to biometrics, sleep, and well being. There’s one with the DLP [Driver Loyalty Program] Labs; they’re constructing automobile information. And inside their data-set, the Tesla information is absolutely fascinating as a result of most individuals take into consideration Tesla as worthwhile as a result of they’ve a knowledge lead, proper? Really, the customers can get lots of that data-set.
You’re pivoting from idea to apply with the brand new collaboration with Flower Labs to construct COLLECTIVE-1. What’s the aim there?
COLLECTIVE-1 is the primary user-owned basis mannequin. Often when individuals take into consideration a basis mannequin, they sometimes consider one firm operating a really massive coaching job in a single information middle, proper? Like OpenAI. And the explanation why it is sometimes finished in a centralized method is as a result of it requires, one, an entire lot of compute energy, and two, an entire lot of knowledge.
Flower AI is type of the chief in federated [decentralized] coaching. They’ve finished a very nice job of constructing these nice open supply libraries. They’ve are available in from the coaching aspect and the algorithm aspect. And with Vana, we actually deal with that information piece, proper? So we principally have all this information that individuals can practice on. You then give customers end-ownership of the mannequin, and customers can resolve on what the mannequin is allowed to do? So that is the primary basis mannequin of its form.
And the speculation is that ultimately, with higher information, you possibly can construct AI that’s not simply aggressive with the central gamers however higher, is that proper? So it’s not nearly ideology, but in addition efficiency.
Precisely, yeah that’s 100% proper. From a decentralized context, I feel typically individuals agree in precept that, “Sure, we must always have AI that is owned by the individuals. We must always have decentralized AI.” However what’s the factor that we will truly do higher in a decentralized context? Information is the reply. For every firm, they solely have their single slice of a data-set. Apple’s obtained their information. Google’s obtained their information. However in the event you’re going by the person, you possibly can reduce throughout platforms and truly construct higher data-sets than any single firm. Information is the key sauce that makes all of it work.
Find it irresistible. Thanks Anna, see you on the AI Summit in Toronto.
Jeff Wilser will host the AI Summit at Consensus 2025, and is host of The Individuals’s AI: The Decentralized AI Podcast.
