Aug 17, 2023

Venture-capital giant Andreessen Horowitz (a16z) is jumping into crypto tech research with the release of a pair of open-source software projects aimed at speeding up the core tech behind zero-knowledge (ZK) proofs. a16z crypto chief technology officer Eddy Lazzarin joins "The Hash" to discuss the latest projects titled "Jolt" and "Lasso."

Video transcript

We're joined by Eddie Lazin of Andreen Horowitz A 16 Z. We are gonna talk about ZK stuff and for those not in the know that means zero knowledge technology. Hey, Eddie, how are you doing today? Hey, how you doing? Sweet. Um So you know, so a 16 Z obviously huge presence in the space known more for investing in companies than developing tech of its own. Talk to me about sort of the departure or what I understand to be a departure from that existing model and going a bit heavier into this deep tech research specifically around ZK. Yeah, it's funny, I, I don't actually think of it as a departure. Uh When I joined the firm, there was a really deep focus on research. The way I think about research is getting really deep into the market from a purely technical angle. You know, when I, our, our crypto team is and has been from the beginning, engineers, former engineers, people with some kind of technical background, everybody's technically literate that's been in the DNA from the beginning. And I, I remember uh when I joined the firm like a little over four years ago and one of the first things Chris told me is like he loved the research spirit of the team, right? And we, we had used to have like this old uh you know, A I repo back like eight years ago, there was a bunch of public open source stuff. So it's, it's been in the DNA of the firm, I think what's changed is that crypto is really complicated, obviously. And our portfolio companies uh have a great demand and need for research and engineering thinking and partners and support and, and all those types of things. So as the, as the crypto team has grown, it's been something that we've invested more and more and more in. And we're getting to the point where uh a lot of the research we've been investing in for a couple of years now is able to bear fruit. And one example is uh which is uh you know, a new kind of cryptographic construction, both great names. Uh So congratulations on that, Joel and La. So can you sort of explain what each of them do, what, where, where they are similar and where they are different? Yeah. So it's uh it's really hard to go into it without getting deep into the weeds. You know how this is ZK tech and cryptography in general. It is really difficult to talk about in a high level way. But the really quick summary is that Lasso is a new type of look up argument which is a critical component of snarks like ZK snarks and jolt is the application of Lasso to instruction set architectures. I know that sounds kind of complicated. But all that means is uh you get to apply this Lasso technique to uh programming languages like two general programming languages like the EVM like Solidity or Rust or something else. And what that means is general programs like regular conventional software programs can be proven in zero knowledge using Lasso. And that has a lo jolt and that has a lot of consequences for uh for crypto all the Zkevm hype we've heard about and they talk about ZK, roll up EKVM and all that stuff. Everybody's trying to do that. Like that's, that's the goal. Everyone wants to prove high level programming languages and zero knowledge with minimal overhead. And this helps us get there. And I can say a little bit more about why that matters. But that's, that's the, that's the gist. I gotta follow up actually kind of on what you're saying with Zach saying just as a research focused institution, but it's also VC, it seems like it's more difficult than ever to apply capital to the space because things are increasingly becoming math based, increasingly becoming technologically driven. It's like, how, how do you throw capital into this and make sure that it's going to like work out on the other side. So we're looking at the ZK space specifically, how are you guys approaching some of these investments. What is your guys' process to do uh your due diligence? Yeah, I, I'd say, uh you know, there's always a ebb and flow in, in technology investing. There's sometimes there's times where there's tons of people raising exactly in the areas you want to be investing in and sometimes that slows down a little bit. Uh We tend to look at things at a longer horizon than that. So it's uh you know, I'm, I'm not exactly sure that that's slowed us down or change our thinking as far as our diligence process. Um Yeah, this, I mean, I got to admit this plays right into it, right? When, when people are raising for a very complex interesting new crypto project, having people on your team who can think really deeply about exactly that underlying tech uh helps, helps a ton. Uh So, II I don't think we've had challenges deploying capital like when we've wanted it. In, in this case, the goal is honestly to accelerate the space. You know, we are aligned with crypto in a really, really interesting way because we're very long term investors, we don't buy and sell assets like week to week or month to month or even year to year. Honestly, we hold things for very long periods. We invest in the space and all the interesting areas in the space and if the space develops faster, uh and we can accelerate it by a year or two years or something like that through our applied efforts in open source and specifically to our portfolio companies, we're gonna do better. So we win if crypto wins. And that's the thinking honestly, there's aren't that many spaces where you can apply so much leverage with really high quality researchers thinking really honestly about how to advance the entire space at once. So lots of very interesting aspects to the space every bill cycle, we see capital kind of flow to one corner of it, right? And there's been a lot of hype around A I if you were to look into the crystal ball and look towards the next BILL cycle, where do you think we're going to see funds flow into? Is it going to be like Z key and ZK and like infrastructure layer things is going to be A I like, what is it going to be? Do you think that's so hard to predict? I I can't possibly, I can't possibly pretend that I know. Yeah. Yeah, it it's super, it's super difficult. I mean, I'd say like if you look at, if you look at A I, you know, A I has had a similar history people, you know, it's super hot right now. It's amazing. I've been a huge fan of A I for forever, but uh we just kind of got out of a long A I winter people kind of forget this just 23 years ago. Uh People weren't even saying A I, they were using euphemisms like machine learning and uh other types of things to talk about it. Uh You know, there were people saying, oh, it's just matrix multiplication. It's just linear algebra. Like why are we calling it A I? Right. And now A I is back. That's great. I love it. I love to see it, but that's kind of the way technology develops. It's in these summer, winter, hot, cold periods. What exactly the next summer is gonna be for the next area of tech? It's super, super hard to say. My hope is that we finally develop the infrastructure over the next year to the point that all of the key blockers preventing cool applications from being developed are resolved. I'm very hopeful about that. I could talk about it for a long time. But I think that uh if it's up to me, it would be really, really cool app layer stuff enabled by the latest changes in infrastructure which are super exciting. We won't ask for any more price predictions. That's, that's Jen's thing. Sorry to put on the spot there. Let's go overseas for a second. There's been a lot of talk with what's happened with Gary and others like, hey, I'm a tech person or I'm a crypto person. I need to go overseas and a 16 Z did just open up a new office in London. So somewhat kind of following on that whole narrative even though it's probably planned, you know, months to years beforehand. How do you see, like from your basis being the head of research at this company? Are you seeing a lot of people wanting to be outside the United States or is that kind of an overblown narrative? And people are just as happy in San Francisco or elsewhere in the United States as they were a few years ago? Yeah. Yeah. So, well, well, first let me say I'm not the head of research, I'm the CTO and we have an incredible head of research named Tim Ruff Garten who leads the whole research team. So I don't wanna, don't wanna get any confusion about that. But as far as moving overseas, you know, we have seen, I don't know if you saw in our state of crypto report, uh there's a lot of developer activity that seems to be shifting overseas. Thankfully, crypto developers are pretty sticky once they get into the space and they start, you know, developing real tech, especially open source tech that we can see, you know, we can see their activity on github. They tend to stick around. Uh and when you look at where they're committing their code from, it does seem to generally be shifting proportionately overseas. Exactly why that is, is kind of hard to say, but I don't think anybody could doubt that the regulatory climate has been a little bit difficult in the United States. I'm not a lawyer. So I'm not gonna go super, super deep into it. But, uh you know, it's, it's uh an inevitable thing. Crypto is gonna be built where it's gonna be built is probably a question for people who set up the regulation. So I'd love to, uh, and, you know, I, I, I'd love to see regulatory clarity and really thoughtful, clear thinking about what we want to see built where, I don't know. That's just, I, I'd love to see it when we're, we're, we're not gonna be um you know, we're not moving the whole firm or something to London. We're investing in London and thinking really, really deeply about how our portfolio companies should position themselves and where they should position themselves and where to focus on investing. Like that's kind of, that's kind of our angle. Um I don't think it would affect our research efforts, right? I'm taking this one. So we had the, we had the, the prediction, we had the, the, the attempted Gary Gensler bashing. I want to get back to the tech. All right, I want to get back to the, all right. Bear with me guys, bear with me guys. So there's been a lot of advancement I think in 00 knowledge research over the last five years, like, you know, catalyzed in part by the applications in the crypto world, talk to me about the importance of ZK and just help me zoom out for an audience that may not be super familiar with the importance of zero no technology. So I think a lot of people think that's like a privacy preserving technology where I think we're seeing a lot around scaling relating to these uh various CKK tech. So I just want to, I want to understand from you what you think the mix is because ZK can be uh overarching any number of different things. What is the real sense of what this accomplishes? Well, first, I really appreciate that framing because you're exactly right. Right. The generally the two properties are privacy and succinctness, which just means making things really short, like proving knowledge or proving facts in a really concise manner. And obviously, if it's short, it's easier to read. If it's easier to read, then you can read it faster. And if you can read it faster, then the computer can do it faster. Right. That's so it's, that's why secs leads to scaling and that's why ZK roll ups are typically a scaling technology. There's really only one uh major privacy preserving Ziki roll up project. That's Aztech, which we can talk a little bit about at some point. But the big picture is that crypto needs really interesting game theory and cryptography to work. And the more that we can put into the cryptography, the more that we can guarantee uh the more that we can design these systems so that computers can guarantee what happens. Right. That's one of the cool, really cool things about cryptography. You can use math to enforce the properties of the system. ZK tech is one of the coolest corners of cryptography because it lets you really, really play with proving facts, proving facts, concisely, scaling computations that needs to occur across many, many, many systems. Of course, this is true in ZK roll ups. But here's like another cool area. What if you wanted to prove what A A I model made as an inference, right? Of course, you can go get a rent a machine on Amazon, you can run a, an LLM yourself, you can train, train one, you can do all that stuff. But if you want to do it in a way that minimizes the trust relationship that you have with the person running that computer, you might need ZK tech in order to verify the computation in order to prove that the computation was run exactly as you wanted. Uh There's a bunch of companies thinking about this type of thing. One of our portfolio companies Jensen is thinking really deeply about how to train in a trust list decentralized network. There's a bunch of cool projects on this. But uh the the fact of the matter is is that if you want computers to do really complicated trustworthy things and you want computers that can make commitments and you want it to be fast, you probably need CK tech uh at a, at, at a minimum and Uh I, I haven't even waded into the privacy area because it's a bit complicated to talk about. But uh that's, that's the big picture. It's a cornerstone of crypto and will be the way that crypto scales in the future. There are no doubts about that. Thanks for doing your best to keep us succinct on, brand on theme. We appreciate it. That is CTO of a 16 Z Eddie Lazin. Thanks so much for joining us, Eddie. Have a great day. You too. Thank you.

Learn more about Consensus 2024, CoinDesk’s longest-running and most influential event that brings together all sides of crypto, blockchain and Web3. Head to to register and buy your pass now.