I'm the cofounder and CEO of Vana, building the first network for user-owned data. With Vana, users own their data and the value it creates, including the AI models trained on their data. We believe data will power the AI economic shift over the next decade. Giving users true ownership of their data opens up walled gardens pushes AI progress forward through data abundance.
I studied computer science and economics at MIT, where I got into crypto through mining ethereum in 2015. I left school to start a machine learning company (YC W18) and was an early contributor to Celo. I also did research at CSAIL (MIT's AI lab), the Federal Reserve, the European Central Bank, and the World Bank, helping me see that the future currency will be data.
I got into decentralization through mining ethereum in 2015. I was obsessed with central banks: I worked at the Federal Reserve in high school and had a photo of Janet Yellen in my bedroom. Crypto was fascinating as it allowed me to to run a decentralized central bank from my dorm room, using cryptography to distribute power from large institutions to everyday people.
I saw the world through data ever since I was a kid. At one point, I created a program to turn US inflation data into a song. I was eager to model all sorts of data at World Bank, but noticed other interns had to sort documents all summer, which I decided to automate. This turned into my first company, Iambiq, and I left MIT to go through Y Combinator.
It was the past wave of machine learning: the Attention is All You Need Paper (2017) had just come out, but generative models were still 5+ years out. I was so excited for the day when models could not just classify, but generate.
When you've spent time training machine learning models, you become obsessed with getting good training data. Better data in -> better data out. Very high quality training data was everything. I came to the belief that the future “currency” was going to be data.
What would the central bank for data look like? It worried me that a few big tech companies had the largest data treasuries, making it hard for developers outside of big tech from training leading models. How can we create a global data treasury that's owned by the people who contribute to it? How can we create user-owned AI models created collectively by a community? These questions led me to start Vana.