Anna Kazlauskas | CEO and Co-Founder of Vana : Solving AI's Data Dilemma

Diksha Lyer (0:00)

Let's start with a question. So what is that one thing that you like and dislike at the same time about working in the decentralized AI space, and why?

Anna Kazlauskas (0:13)

Yeah, I think the pace of development is so fast, especially open source AI. Like I'd say, actually, it's maybe more true in open source AI, and then decentralized AI is almost catching up. But everything moves so fast. It's a ton of fun, right? But then it's also like you really need to like stay on top of things because just what's possible from a research perspective changes, like every three months or like every month even. So I think there is, I see mostly the good in that, but I think there's also an aspect of like makes it so you kind of always have to be online and like really tracking what's going on.

SPEAKER_3 (0:51)

Yeah, the space is definitely super cutting edge, but it's also extremely fun because of that. I wanted to dial just back a little bit. We always kind of start by this origin story of, you know, how, I actually wanted to ask you, how did you get into crypto in the first place?

Anna Kazlauskas (1:11)

Yeah, so I first got into traditional currency. So I was really interested in central banks. Like I had a picture of Janet Yellen chair of the Fed in my high school bedroom, which is just a very unusual thing for any high schooler to do. I like worked at the Fed in high school and to me, currency was like central to it all. I was like, whoever controls currency is really controlling the economy. And so very interested in understanding that. And then I got to MIT and I learned about decentralized currencies. There was like the five-person MIT Bitcoin Club, which was very ideological, kind of like anarchist of like, hey, like, let's take back control of currency. And around that time, Ethereum was kind of just getting off the ground. And so I started mining Ethereum with one of my friends who, yeah, we both had like minor rigs set up in our dorm rooms and we would just use the free electricity from the dorms. And then I would actually go get GPUs from the dumpsters that MIT labs would get rid of. That's what I was talking about just before we started the podcast. Like I went back and visited and there were some kids there actually like hunting, scavenging through the hardware, looking for stuff. And they lent me their screwdriver and their wire snippers. So I actually brought a motherboard home, just like for sentimental value. But yeah, that's how I got into crypto. I think mining is a really good way to get involved because you're literally running the network, right? And I think in the proof of work days, it was really cool because you didn't need any capital to get started, right? You could just come in, have hardware and like contribute. And then that got me just fascinated in kind of like decentralization and understanding how you can apply that in Vana's context to really like data and AI. The sort of AI path I took came out of some research that I was doing at CSAIL. So I was really interested in econometrics, which is like just ways of modeling data, mostly about humans. And I wanted to push the methods of like what was possible in terms of like measuring things. And so I started looking at machine learning as a way to do this and ended up taking some classes and doing research with Regina Barzolet, who at the time was really this like pioneer in natural language processing and started kind of trying to understand how can you model data. And this was, yeah, so doing stuff with vector embeddings and sort of figuring out how to represent data with AI.

SPEAKER_3 (3:34)

Did you cross paths with SPF or Gary Gensler?

Anna Kazlauskas (3:41)

Okay, I did. So yeah, I overlapped it. Yeah, so Gary Gensler was teaching a class. I didn't take it because, you know, a lot of the undergrads, I think, kind of like wrote off some of the Sloan MBA grad classes. Like, oh, it's not technical enough. Like, we're going to take like these network design, distributed systems classes. I wish I had. He would be a really interesting character to have shaped some of my like world view. I wonder what he was, I should actually think the course is online, right? I should actually look at it and kind of understand what he was teaching. But I didn't overlap with SBF. Yeah, I was supposed to graduate in 2019 I actually ended up dropping out because I started a company that came out of automating my job as an intern sorting documents and then like went through YC, learned this whole, found like Silicon Valley and was like, this place is amazing. Like, you can just build technology and scale it.