Antonio Juliano, founder of the decentralized exchange dYdX, and Eli Ben-Sasson, cofounder of StarkWare, discuss the recent launch of dYdX on Ethereum layer 2 using StarkWare’s scaling technology. In this episode, they cover:

  • what cross-margin perpetual trading is and why dYdX chose StarkWare to be its scaling partner (1:13)
  • ZK rollups and how they contrast with optimistic rollups (8:49)
  • how many transactions per second StarkWare can handle through its layer 2 solution (16:57)
  • lessons learned and critical takeaways from their six months spent working together (20:15)
  • the similarities between Ethereum and StarkNet (26:07)
  • why Antonio believes that scalability is more important than composability (27:30)
  • the benefits of decentralized exchanges and how dYdX fits into the dex marketplace (31:54)
  • why dYdX chose Chainlink as its oracle provider, and how it got oracle prices onto layer 2, which previously had none (38:11)
  • why Cairo, StarkWare’s Turing-complete programming language and platform, can be the “one verifier to rule them all” (42:10)
  • what it was like for StarkWare to compete in the Great Reddit Scaling Bake-Off (45:41)
  • whether or not DeFi could handle a GameStop/Robinhood situation (48:56)
  • fair launch tokens versus VC tokens (54:12)
  • the likelihood of dYdX and StarkWare each creating a token (58:19)
  • upcoming developments and projects for both companies (1:02:14)

Thank you to our sponsors! 

Download the Crypto.com app and get $25 with the code “Laura”: https://crypto.onelink.me/J9Lg/unchainedcardearnfeb2021

Square: https://square.com/go/unchained

Episode Links:

Antonio Juliano

Eli Ben-Sasson

dYdX

StarkWare

dYdX and StarkWare Relationship

Other

https://www.reddit.com/r/ethereum/comments/hbjx25/the_great_reddit_scaling_bakeoff/

Transcript:

Laura Shin:

Hi everyone, welcome to Unchained, your no hype resource for all things crypto. I’m your host, Laura Shin, a journalist with over two decades of experience. I started covering crypto five years ago and, as a senior editor at Forbes, was the first mainstream media reporter to cover cryptocurrency full-time. Subscribe to Unchained on YouTube, where you can watch videos of me and my guests. Go to youtube.com/c/unchained podcasts and subscribe today.

Crypto.com:

The Crypto.com App lets you buy, earn and spend crypto, all in one place! Earn up to 8.5% interest on your Bitcoin, and 14% interest on your stablecoins – paid weekly! Download the Crypto.com App and get $25 with the code “LAURA” – link is in the description.

Square:

Building a web application to sell products or services? Integrate with Square’s robust and secure APIs to easily take payments. Learn more at square.com/go/unchained.

Laura Shin:

Today’s guest is Antonio Juliano, founder of dYdX, and Eli Ben-Sasson, co-founder, president, and chairman of the board at StarkWare. Welcome, Antonio and Eli.

Antonio Juliano:

Thanks for having us excited to be here.

Laura Shin

Antonio, you and Eli partnered to launch cross-margined perpetuals, which went live at the end of February. Explain to us what cross margin perpetuals are, why you chose that as your project to scale and why you chose StarkWare.

Antonio Juliano:

Yes, absolutely. So, perpetuals are a type of synthetic product that’s really popular in the crypto space. Not a lot of people know this, but perpetuals by volume are the most popular type of trading instruments in all of crypto. Basically, the volume for perpetuals is more than all of the volume for all of the other products including just base spot trading on crypto combined. So it’s a really big market, but I think we’re one of the first in DeFi to get into this market. So we think it’s going to be big. The reason why we’re so excited about it long-term is that normally in traditional financial markets, you see the derivatives or synthetics volume being about 5x-10x higher than kind of the base spot volume.

One interesting thing that happened in crypto just in the past year is for the first time, the derivatives volume on crypto surpassed, like I said, all of the other volume in crypto combined, and it’s well on its way to kind of reaching we think that 5x-10x, in the volume of the entire rest of the market combined. So it’s a huge market. We’re really excited about it. And that’s the main reason that we’re excited about launching this product. More specifically what perpetuals are, is like I said, they’re a type of synthetic product. All synthetic means is that when you’re trading these types of products, the actual currencies that you’re trading don’t actually exist in the contract. So for example, if you’re trading a Bitcoin perpetual, there’s no actual Bitcoin being custodied and there’s no actual Bitcoin under the hood. We just kind of come up with this synthetic contract and it tracks the price of Bitcoin.

And the reason these are so popular is that they can be traded with leverage. Leverage basically just means that you can multiply your gains and your losses and trade as if you had a lot more capital. So it’s much more capital efficient to trade on these types of products. And that’s why they’re so popular. All cross-margin means is, it’s a bit of a technical term, but it basically means that you can have one account or one pool of collateral that collateralizes any different type of market that you might want to trade on. So, as compared to isolated margin, which is what had on our layer 1, or kind of our original product, that means that for every different type of product that you want to trade, so say you want to trade both an Ethereum perpetual and a Bitcoin perpetual, you have to put down collateral into two separate accounts. So it’s not nearly as capital efficient. Whereas, because of the scalability that StarkWare is now able to provide, we’re really able to offer cross-margining and that’s a really important product feature that we’re excited to launch to our users.

Laura Shin:

Yeah, that seems like a no-brainer if that technology enables you to do the cross margin because I could imagine as a user it would be very frustrating if you wanted to trade multiple assets and had to collateralize each one. And so why did you choose StarkWare?

Antonio Juliano:

Absolutely. So we did a lot of research on this about six months ago, right before we started working with StarkWare. Kind of the main things we were considering was obviously StarkWare, and kind of, by extension, zero-knowledge roll-ups, and Eli can talk a lot more about that. We also considered Optimistic rollups and specifically Optimism. And then the third kind of thing we also considered was other layer 1 blockchains. So things like Solana or NEAR. The reason we chose StarkWare was for a couple of different reasons. So first of all, the level of scalability that they’re able to provide is really, really high. So I won’t touch on this too much because, obviously, Eli can dive a lot more into it than I can. But, basically, the way it works is you take just an arbitrarily long list of, in our case, trades, and through zero-knowledge proofs, or STARK proofs, you can basically transform them down into a constant size data object.

And then, so it basically doesn’t matter how many trades you have, you just transform it into this constant size thing, put that on the blockchain and that’s where all the scalability comes from. So the main, or requisite thing, that we were obviously looking for is a really high level of scalability. And this isn’t exactly rocket science, but right now, for anybody that’s used DeFi, just the fees and the gas fees on Ethereum specifically, are obscenely high. Currently, it costs about $100 to sometimes up to literally a $1000 per trade in gas fees to execute a trade on layer 1 on Ethereum. And this was just obviously something that we had to solve as soon as possible. And kind of the second main point was exactly that: the as soon as possible. StarkWare already is kind of one of the very few scaling technologies that already had a really good history of building quality products in production. We’re not the first exchange or decentralized exchange to launch with StarkWare. They’re also launched with another decentralized exchange called DeversiFi and they’d been running successfully in production at that point.

A lot of these other optimistic roll-ups and other L1’s are very much unproven stuff. It was a really big selling point that it was already live and it was already working in production. We solved all the architecture problems and we knew how to do it. Kind of the last point, which is a bit more subtle, is something called withdrawal. So for anybody that’s kind of familiar with optimistic roll-ups, one of the drawbacks is that you actually can’t withdraw from optimistic roll-ups for, you know, it depends on the configuration, but oftentimes up to a week. For StarkWare with its STARK blocks, we have basically instant finality once the block is mined on layer 1.

And that’s really important because, you know, for us, like who’s going to use an exchange where you can’t actually withdraw for a week. There are certainly ways to get around this. You can have liquidity providers front fast withdrawals, which we’re actually also doing on StarkWare. So that means you can have instant withdrawals, you don’t even need to wait for the STARK blocks to be confirmed to the blockchain to get withdrawals. But if you’re thinking about becoming a liquidity provider, I would be happy to talk more about this in detail. The thing that you need to think about is what the withdrawal period is. Suppose you’re a fast withdrawal provider for Optimism, or by extension to any optimistic roll-up protocol, you basically have to have enough capital to cover an entire week’s worth of withdrawals for the system, which even for us, a relatively smaller exchange compared to centralized exchanges is like millions and millions and millions of dollars to the point of where it was just totally financially unfeasible.

Whereas for us, because the StarkWare blocks are mined roughly once every hour or so we only need to have a smaller amount of capital to front withdrawals on the system. And then the last thing I’ll say is compared with just kind of other layer 1 chains, StarkWare, and the system we’re building on top of StarkWare settles back on Ethereum. So it has basically exactly the same security properties and exactly the same decentralization properties as Ethereum does, which is really important to us. And even more important, it works with all the Ethereum wallets. And this is just absolutely critical from a user experience perspective. If you’re a decentralized app and you want to get users, you have to support the wallets that people are using. And because we’re just settling back on Ethereum, we can use exactly the same wallets that all of our users are used to using. And this continues to provide what we think is a really top-class product experience on DeFi to them.

Laura Shin:

And Eli, I know Antonio described briefly how the StarkWare solution works, but why don’t you dive a little deeper into the ZK rollups and also maybe contrast it with optimistic rollups, which he mentioned, and that is another popular layer 2 solution that a lot of DeFi projects are trying out.

Eli Ben-Sasson:

So the basic problem that all rollups try to solve is the problem of this limited resource, which is the block space on the blockchain. In the case of Ethereum, you only have 12.5 million in gas to allocate among all the competing transactions in every block, every 15 seconds or so. And because there’s a lot of demand these days for this resource, prices are shooting up.

The way that roll-ups work or all scalability solutions, L2 scalability solutions work, is that they sort of take a lot of this computation and data and everything that costs so much and takes it off-chain. It’s a little bit like an iceberg where you put the tip on L1, which is Ethereum, and you use Ethereum to sort of verify or do the minimal kind of stuff that needs to be done to make sure everything’s safe. And then all the rest is taken off-chain. So like an iceberg where you only see on-chain the tip, okay. Now, this is the idea of all L2 scalability solutions that we’re aware of. Now rollups are a particular version of scalability solutions that have added security. It’s sort of a trade-off.

What they say is we need all of the information for each and every user to know exactly the state of her account at any given point in time. We need that information, that raw data to appear constantly on L1. We don’t want any of this data to sit with some data availability solution, which is another option. So, this is a trade-off. You get the security of L1 for your data. The downside is that there’s a linear amount of where basically more transactions means more data on L1. But, it’s a trade-off. We at StarkWare support both options of this data availability situation. So dYdX chose to go with the rollup model, which means that all of the data sits on L1. DeversiFi, for instance, chose to work with offloading the data to a data availability committee that is also very secure and operates in a different way.

And so, these are two types of trade-offs. There’s even a third type where you can let the users decide whether they want to pay extra for having their data on-chain, or whether they’re willing to. If they want to save costs, but rely on this L2 data solution, they can. We call this volition. And we’re also going to support that with some of our customers. In any case, that’s what roll-ups allow you. Now within roll-ups, there are two main ways to deal with security. One branch is called validity proofs, which StarkWare and STARKs are one of these ways, which basically means whenever you move the state, whenever you change this tip of the iceberg to something new because you’re updating the state of your system, it must be accompanied by a cryptographic proof.

In our case, those are STARKS, which are very lean and short and cheap to verify, exponentially cheaper than the amount of computation that, you know, leads to this change on-chain. You cannot move the state without validity proof. This ensures that you will never have any movement to a state that is not valid and does not have integrity. So this is one way to deal with it. And these validity proofs in the cryptographic assurances they give allow you to have the instant finality that Antonio was talking about and that the dYdX customers are going to benefit from. The other kind of solutions are game-theoretic, fraud-proof based solutions, Optimistic rollups, which say something like this: anyone can move the state of the system to any new state, including to a fraudulent or invalid state.

And then there’s going to be a period during which parties can sort of raise a flag. There’s going to be some process by which you sort of sort out whether the state of the system is invalid or valid. There’s a certain game going on that has a bunch of interactions between two parties. One is saying, no, this was fraudulent. The other party is saying, no, this was good. And you have to wait a little bit and allow this thing to sort of peter out. And this is what causes this capital inefficiency that comes from funds needing to be locked for a long enough time in order to make sure that no one raises a flag. And if the flag is raised, you know, to sort out what happened. So those are the Optimistic rollups.

So just to summarize again, scalability solutions take a lot of the load of computation off-chain. Rollups mean that all the data, the raw data, not the computation and not the storage, still appears on L1 for added security, but there’s a trade-off, you pay a little bit more. And then within the two families of rollups, you have validity proofs, like STARKs that allow you instant finality, and you have Optimistic rollups that come with this, you know, finality period of roughly a week or so till you can remove your funds.

Laura Shin:

And as far as I understand, just from the description, it seems kind of obvious that it’s better to have the instant finality, and yet there are teams that are choosing Optimistic rollups. As far as I understand, is that simply because the technology is maybe easier to implement now, like it’s further along, is that what the case is?

Eli Ben-Sasson:

So I don’t think that’s the case. I think that our technology is further along. We have more projects that are launched or ready to be launched with our technology. One thing that is offered by some Optimistic rollups project is to say that…to describe it this way, you basically take all of your existing Solidity code, you press a button, and it automatically works just as previously. Now I’m not entirely sure that actually is the case, but this is a very tempting proposition that causes a lot of projects to say, okay, we want to press that button and have it automatically work. I think a lot of projects are definitely saying they want to wait to see that happen. Now it hasn’t happened yet for any meaningful project.

You know, we have to wait and see. So, I think they’re saying we are willing to bite the bullet of the capital inefficiency and the extra sort of security issues with having these fraud proofs, you know, the ability to attack them and delay the system and whatnot. Because they are tempted by this proposition that they won’t have to modify its code significantly. You know, I’m biased, of course, but I think Antonio and dYdX made the right choice of going with a better solution in terms of capital efficiency. We started working with them roughly six months ago, and it’s already on mainnet. So I think this practically shows that this is a better route, and we hope other teams are gonna go down this route.

Laura Shin:

Yeah, definitely. When I realized the timeline on that, I did think that that was pretty quick how you guys rolled this out. And before I forget, because I’ve now forgotten multiple times, I wanted to mention that DeversiFi, who we’ve mentioned a few times, is a previous sponsor of my shows. And so then, in terms of scalability, how many transactions per second are you now able to process versus, you know, what it was before?

Eli Ben-Sasson:

I prefer to talk about things that we actually demonstrated live on mainnet rather than various [tests]… I mean, there are a lot of numbers floating out there for things that theoretically could be done. On mainnnet, for the Reddit Bake-Off, we demonstrated a single proof that settled 300,000 transactions. And it was in rollup mode, meaning that all the data was on-chain. The effective TPS for, and this was on mainnet Ethereum, and this was a few months ago before you were working with dYdX, it was a TPS of 3,000. With DeversiFi we have demonstrated the ability to reach trades of 9,000 TPS, and for payments, we can go to 18,000 TPS with dYdX.

The system is just now being ramped up, and we’ll be very proud to announce the effective rates that we get once we actually demonstrated them. We saw this tweet that there are more than 80,000 people waiting to join the dYdX system once it launches. So of course, this causes… we’re up for the task… but it sort of causes you some sort of angst, hoping that everything will function properly. We will report, in a small number of weeks, the actual numbers that you’ll see on mainnet live.

Laura Shin:

I can just see it now; I’m going to have to run an update and be like, by the way, they touted all these numbers, then it… No, I’m just kidding. All right. So for the perpetuals products, some of the features are that traders can now trade with zero gas costs. And so how are you making that possible? Is dYdX just paying for them because it’s such a small amount compared to what it was before?

Antonio Juliano:

Pretty much. Yeah. Obviously, because it’s a roll-up that settles on L1 Ethereum, there are still gas costs, and we’re basically paying for every time a STARK batch is committed to mainnet, we’re paying for that. But it’s really not that much. And that’s where all the scalability features that Eli was talking about come into play where it’s just costing us however many X less than it was before. It’s gotten down to the point where we’re really able to charge fees that are much more in line with what traders are used to paying on centralized exchanges. So even fees on centralized exchanges in crypto are fairly high, right? Like even 20 BPS is a lot higher [than centralized exchanges]. And that’s what you’ll experience on like FTX or Binance than what you’ll see in traditional finance. So the revenue that you make in fees is more than enough to cover, in this new case, what we would have to pay on gas. So we’re, like I said, just aiming to charge fees that are really in line with what other exchanges charge, and no longer will the fees have to be deterministic, or kind of dynamic, based on the current gas price environments on Ethereum.

Laura Shin:

Very interesting. And so, in working together, as I said, it is impressive that you managed to roll this out in six months, but I’m still wondering, everyone, in DeFi in particular, is looking at different scaling solutions… that is really deemed to be probably the main issue right now. So I was wondering what lessons you feel like you learned in working together, especially for DeFi in general or other DeFi protocols that are interested in scaling?

Antonio Juliano:

Yeah. Maybe I can start. I think the first thing is what Eli touched on before, which is that in my opinion, for a good technical team, it actually doesn’t really matter what language your smart contracts are written in. Like we’re happy to port our smart contracts to Rust or Solidity or, you know, soon to come out Cairo, which is StarkWare’s new programming language that they’re coming out with for their STARK rollups, which Eli can talk a lot more about. I think the first lesson was that it’s really not that big of a deal if you’re a good technical team to move to a new coding language. I think people just put way too much weight on that in terms of making things backward compatible. I’d say maybe one-second learning that we had is that basically for the current contracts that we have, StarkWare wrote them. They wrote them in their own Cairo language.

And one of the things that they’re working on, which I’m really excited about, is what I just mentioned before, is their Cairo programming language, which will allow third-party developers to comment and write their own contracts. I think that’ll be a really big step up from the build process that we had, because of course, whenever two different companies are working together, it’s just a little bit harder to have coordination across the two teams. For example, StarkWare was writing all the contracts and we gave them pretty detailed specs, but of course, there’s some back and forth on various edge cases for those contracts and just time making sure we were able to have all the kinks ironed out and have it exactly the way we want. If later we’re able to come out and our developers are able to just code in Cairo directly and have that run directly through StarkWare, I promise you, we could get the build times down to three or four months or even less. So, that’s something I’m excited about looking forward into the future.

Eli Ben-Sasson:

Yeah. So on, on our part, the first thing I just want to say is it was, and still is, such a pleasure working with such a professional team as the counterparty on this. We learned a lot about, first of all, the business and the business logic that is needed in these cases. This of course helped us and will help future customers and players that want to build on top of this and extend the capabilities of STARKs and Cairo. The second thing that was really important for us, Antonio mentioned this: Cairo is a programming language, which allows you to express in a way that programmers express themselves, any computation in any business logic that you want to scale.

And I want to emphasize this because probably not all of your listeners are aware that when you come to ZKP technologies, one of the hardest parts is taking the computation, the high-level computation that you have in mind, and then converting it to these very intricate mathematical objects that are really, really hard to deal with (various things like polynomials or finite fields and so on). What Cairo achieves is a very clean abstraction of all of that. It sort of allows you to just write your business logic in a very clean way. And this partnership was for us to get another test case for the agility and the feature velocity of Cairo. And of course, along the way, we were constantly improving it. And the whole framework that it builds on.

All of this is really important for us. Because as we already announced, and this also ties to what Antonio said, our plan is to allow teams like dYdX, and of course, anyone who wants to permissionlessly write their own contracts and then deploy them on an L2, to reach an arbitrary scale. We call this StarkNet and it will be launched later this year. And basically, it will be an L2 on which you can deploy your logic. So each further iteration where we work with the teams, especially such a professional team as dYdX, we sort of hone and make the tooling much better so that when we release it to the public, it’s going to be much more ripe. I just want to mention that when I say releasing it to the public, it’s not something that is in the future.

It’s actually something that happened in the past. Everyone’s invited to go to cairo-lang.org, and basically download Cairo, start programming, and send proofs right now to a service that is on Ropsten, it’s called Sharp, basically shared proofing. So you can have several different applications sending their proofs all to be combined into one single proof. It goes on-chain and is verified for all of them. And then you can amortize gas costs a little bit better. So Cairo and the same programming tools that we use in order to build the system, our part of the system with dYdX, these tools are already available for everyone to use, with a lot of tutorials. And so we’re very excited about that, and much more is going to come from that.

Laura Shin:

And so, just so I’m understanding this idea of a StarkNet, it’s almost like, I mean, it just sounds like a new kind of Ethereum in the sense that it has its own programming language, it’s Turing complete. And so now people can use that programming language to create DeFi programs that use ZK roll-ups and are inherently kind of like scaled from the get-go. And maybe even then, also have that feature of composability that layer 1 DeFi protocols have on Ethereum, is that correct?

Eli Ben-Sasson:

Yes, Laura, it’s exactly that. We think of it as just like Ethereum and being Turing complete. It does not have the gas limit of Ethereum. That’s the whole point, it’s an L2. In terms of security, it’s security is as safe as that of L1 Ethereum and is going to rely in a lot of aspects on L1 for various aspects of sequencing, allowing users to send commands to the cell to extract their funds if they need. You will have composability, you will have a permissionless deployment of smart contracts that will be transactions. Again, permissionless, and users can submit it. You’ll have decentralized operators, approvers, and, yes, that’s our vision.

Laura Shin:

That sounds really fascinating. So my next set of questions is about dYdX and perpetuals and kind of diving more into that, but why don’t we just jump to one of them right now, because I was gonna ask Antonio about the fact that obviously these DeFi protocols do benefit from composability on layer 1. I think it’s slightly less of an issue for dYdX, but Antonio, can you just tell me how you have been thinking about that and how that factored into your decision to go with StarkWare and ZK rollups?

Antonio Juliano:

So it definitely is a factor, first of all. But I think you touched on it, and I think for us, specifically for derivatives or synthetics, it’s actually, in my opinion, much less of an issue than it is for different spot trading protocols. For example, let’s suppose Uniswap wanted to migrate to any given layer 2. One of the things they would have to really think about this is, like, well, okay, are there actual tokens that people want to trade on this layer 2 project? For us, again, we don’t need that because we’re just using synthetic contracts and all we need is one, literally one token, as basically just a collateral token to exist on this layer 2, and then synthetically, we can create just any given asset that users might want to trade. I’d say I also have kind of a counterculture view on this, and especially as it comes to DeFi, but really our thing at dYdX is to take a very full-stack approach to building our products.

We like to think that we do everything — everything down to building the smart contracts, to building like the backend servers that run the order books and the matching engine to building the front end, to eventually like building the apps that we’re going to come out with. Honestly, we just think that we can build it better than other third-party teams would be able to, for the most part. Whereas a lot of other DeFi has this ethos of composability and is like, this product is cool. It’s built on top of like 0X API, which is like building on top of Uniswap and Curve and like all this stuff. And that’s awesome. But that’s just like, for us specifically, like not exactly the target that we’re going for. We really just want, we have like a singular mission, which is just making the best product for advanced traders on cryptocurrency.

And we really think that we can offer that in a really high-quality way without too, too much composability just through perpetuals on layer 2, even right now. And that’s not to say that, like, there will never be composability for dYdX, especially with a lot of this stuff, like Eli was talking about, with kind of moving just other contracts to the same, STARK rollup chain. I think that’s really awesome and will certainly unlock some things for us, but just the, when you think about the trade-off of like, okay, like, you know, a 1000X or whatever, scalability versus like, you know, some composability it’s like, you know, I will take a thousand X plus scalability right now over that composability.

Laura Shin (30:38):

All right, in a moment, we’re going to talk more about the perpetuals products, but first, a quick word for the sponsors who make this show possible.

Crypto.com

With over 10 Million users, Crypto.com is the easiest place to buy and sell over 90 cryptocurrencies. Download the Crypto.com App now and get $25 with the code: “LAURA”. If you’re a Hodler, Crypto.com Earn pays industry leading interest rates on over 30 coins, including Bitcoin, at up to 8.5% interest and up to 14% interest on your stablecoins. When it’s time to spend your crypto, nothing beats the Crypto.com Visa Card, which pays you up to 8% back instantly, and gives you 100% rebate for your Netflix, Spotify and Amazon Prime subscriptions. There’s no annual or monthly fees to worry about! Download the Crypto.com App and get $25 when using the code “LAURA” – link is in the description.

Square:

There’s a lot to think about when building a commerce web application. Integrating with Square’s APIs to take payments, however, is a no-brainer. Trusted by millions of sellers worldwide, Square’s APIs are now available to developers, making enterprise-grade payments accessible to everyone. They also have your back. Use their APIs and you’ll get dedicated developer support, dispute management and fraud detection. Start building your online payment form today. Visit square.com/go/unchained. Again, that’s square.com/go/unchained.

Laura Shin:

Back to my conversation with Antonio and Eli. As you’ve noted, perpetuals are some of the most popular trading products in crypto, at least on centralized platforms. And I wondered what benefit do users get from trading perpetuals on a decentralized exchange?

Antonio Juliano:

Yeah, absolutely. So I think it’s pretty similar to the benefits users would get from using any decentralized product. So the first is just security. It’s a lot more secure to hold your own keys. I’m certainly not the first one to say this. But trading directly on a smart contract through your hardware wallet is just a lot more secure than trusting your keys to a centralized exchange. And, actually, I have a lot of experience with this. I used to work at Coinbase as an engineer and I was an engineer where one of the things I worked on there was their hot wallet service. So I’m really familiar with kind of the security level of centralized exchanges. One of the cool things is that even if you do a good job on security, as a centralized exchange, there’s just so much overhead that goes into securing that level of funds because there’s like a $90 billion incentive or whatever to hack Coinbase. And you have to as Coinbase, spend resources proportional to that, to secure your platform, even if you’re doing it the best way. So like I say that to mean, like for us as a decentralized exchange, we can move a lot faster, can offer like a lot more products after we just like focused on securing the smart contracts. I’d say the second is transparency. So especially for more advanced financial products like synthetics and derivatives, you’re not only trusting your coins to the exchange, you’re also trusting the entire execution of the contracts that you’re entering to said exchange and are counting on that exchange to honor the contract exactly as it was written up.

So for things like, you know, BitMEX and FTX and not to pick on anyone, I mean, this is just like the way like technology works before blockchain. It’s just like really a black box, how a lot of this stuff works on the centralized exchanges. Like, you know, when were people getting liquidated? What were the index prices that caused these liquidations to happen? How much money is on the insurance fund? Like all these different things. When does the insurance fund used? Those all kind of factor into it and, on dYdX of course, because all of the contracts are open source and transparent, anybody can just go and audit or look at our audit reports and understand exactly the contract that they’re entering into. And then I’d say the last thing, and this is something we’re going to be focusing on, just more going into the future, is access. So we’re not quite there yet in terms of dYdX, right now, is what’s known as a hybrid exchange.

So that basically means we have some decentralized components and then we also have some centralized components. So the central components are kind of our order book and our matching engine which basically we have running on our servers to just offer a much more performance level of trading to our traders. But one of the things we’re going to be focusing on going into the future is just decentralizing more parts of the stack. And that’ll allow us to just offer dYdX to more people and more places in the world. Because again, once we get closer to this vision that all of DeFi has, where it’s just open source code running on a permissionless blockchain, anybody, basically anywhere in the world, will have access to that code.

Laura Shin:

Yeah. And something that interested me was when I was looking kind at the descriptions of dYdX, it sort of almost seemed to me like dYdX was aiming to be almost like the Uniswap of perpetuals, maybe offering more like a long tail of a variety of perpetuals. But then as I read more, I realized, oh, well, since you’re kind of using this closed order book based dev, which then competes against these other derivatives exchanges that operate more like automated market makers, that are like Uniswap, I was like, hmmm, maybe I’m wrong because maybe those are the ones that are gonna fill that spot. So kind of what vision do you have of dYdX?

Antonio Juliano:

Yeah, it’s a great question. I think the order book versus automated market-maker question is certainly not just a consideration for dYdX. I think anybody operating a DEX is thinking about that right now. My opinion on this, and this isn’t a super original thought, is kind of somewhere in the middle. I think that order book-based models are vastly superior for trading kind of the top — call it like hundred or 200 markets by volume. We’ve seen various centralized exchanges be extremely successful working directly with market makers and, you know, FTX is a prime example of this. It probably is the exchange that has the most different pairs. And you’re able to just offer a really high amount of liquidity on those top markets. I think automated market-makers certainly do serve a purpose. But in my opinion, I think they’re better for just trading like the long tail of assets.

And in my opinion, that was the product-market-fit that Uniswap found, which I think is awesome. Just people coming and trading all these various tokens before they were listed on centralized exchanges, or that weren’t even listed ever on centralized exchanges, but they were listed on Uniswap, like the moment that they launched. That’s a really powerful thing. Again, for us specifically at dYdX, our goal is to become one of the biggest crypto exchanges period. But on a 3-5 year time horizon. And we think, kind of the combination of, you know, what I talked about at the beginning, with derivatives already being the biggest market. And then just, we think it will become even more of a percent of the market. And then obviously decentralized exchanges as well, which are currently like, I don’t know, 5% or so of the market but are super new and we think we’ll grow over the next three to five years.

So if we’re able to win that market of derivatives plus DeFi, we think that’ll be really big. I’ll conclude by just saying, I think specifically for derivatives, where we see the real volume is not in the like 200 plus markets by volume, it’s really in those top hundred markets, and honestly, really just in the top five or so, like the Bitcoin perpetuals to a lesser extent, like the ETH perpetuals that are traded on centralized exchanges. Like literally just Bitcoin perpetuals, I think almost do more volume than literally everything else in crypto combined. It’s like, it’s a real power-law thing. And if you want to be the biggest exchange, you have to serve like the biggest market that’s 10x bigger than all the other markets in the highest possible, quality way. And we think that’s by order books, just because order books are able to provide just a much higher level of liquidity, given a certain amount of capital for the markets that they operate on.

Laura Shin:

And also, because you’re using perpetuals, which, as you mentioned, are a type of synthetic, the oracle that you’re using becomes ultra-important because that then becomes a main attack vector. So how did you decide upon Chainlink as your oracle provider?

Antonio Juliano:

Yeah, absolutely. So I think Chainlink, almost the name sort of speaks for itself in DeFi at this point. Most of the leading products are building on Chainlink or at least things that use technology that’s similar to Chainlink. Like dYdX also actually operates on MakerDAO oracles which are, you know, not exactly the same, but kind of use similar technology as Chainlink. But we’re really excited about partnering with Chainlink. One of the really important things that we had to design with this layer 2, which was an open question, and we designed this alongside StarkWare was, how do we solve this oracle problem? There are no oracles right now on layer 2, basically because nobody’s on layer 2. Right. So like how do we get these oracle prices on layer 2?

And we thought about a lot of different things, like potentially building our own oracles that are specifically specific for L2, and we potentially could, but we would have had to have sacrificed on security there. So the thing that we ended up on, was just working directly with kind of the three of us, like us, StarkWare, and Chainlink themselves actually did a good amount of work for this, was basically just putting the exact same oracle prices that they’re already reporting on layer 1, but putting them on layer 2. And this is a really powerful thing because now we get the exact same level of security that everybody else in DeFi is used to. It’s kind of the leading Chainlink oracles, but we get it on layer 2 instead. And it’s actually even better than that, because the whole point of layer 2 again, is that it’s just much more scalable and much faster.

So one of the really big problems we ran into on layer 1 with any oracle provider, and this is not at all limited to just Chainlink, but the problem we faced was there was a lot of lag or like a lot of latency in the oracle prices, especially when prices were moving fast, which is the worst time to have a lag in your price oracles. Like every time, this is a bit of a generalization, but more or less like every time the price would drop by like 10%, it would take, you know, 5 to 10 to sometimes 15 minutes for that to be reflected on the layer 1 oracles. And this is a real problem if you’re operating a leveraged exchange because your contracts don’t know that they should liquidate people. And if the price keeps dropping, the entire protocol could become under collateralized.

And this was the main thing that was holding us back from offering higher leverage. So on our layer 1 product for our perpetuals, we were only able to offer a 10x maximum leverage, and it was exactly due to this problem of the oracles lagging. But now on layer 2, it takes the same security, as I said, but we can reduce the latency, can reduce the lag by a factor of like 10x-20x roughly. So now instead of having to wait minutes for these oracle prices to be reflected on our system, we can do it in seconds. And that’s kind of what centralized exchanges can do as well. Because again, they’re not even using a chain, but for us, we can offer like a similar level of leverage starting out at 25x, potentially could raise that higher, just, you know, starting out a little bit more conservative.

But even just this fact of now the oracle prices are much more performance [level] allows us to offer a new feature on our product. That’s really important in higher leverage. Higher leverage is actually super important for operating a high leverage crypto exchange, as you might expect. Binance recently released statistics on their user’s usage of leverage and what kind of multiple leverages all of their users use. I believe the most popular leverage was around 20x. I can’t remember the exact statistics, but it was something like 30% or so used 20x. It was like some small percent like 5% to 10% use that like crazy 100x leverage. But if you’re able to offer this kind of like 25x like maximum leverage, that’s just a huge step up in terms of the quality of our product.

Laura Shin:

Yeah. I could see that appealing to people in the crypto space. And so let’s also dive a little bit more into some of the StarkWare things, like Cairo in particular, I wanted to ask a little bit more about, because Eli, I saw in some of the blog posts it was being described as “one verifier to rule them all.” Can you explain what that means?

Eli Ben-Sasson:

Yeah. So that’s a terrific question. Typically what happens is, and this is something I spoke about earlier, you have a computation in mind and you want to use the ZKP to scale it. So this computation, you have to somehow express it and convert it to some algebraic format that is very hard to deal with securely and efficiently (like a bunch of polynomials, or I could throw a whole bunch of math terms here at you that basically say the same thing: some complicated algebraic problem that may have some sort of number theory in it… It’s very complicated). So whatever means you did for that, you would have to work very hard, and really, there is a limited amount of complexity that you could deal with. And then you ended up with some system that is very specific for the problem at hand.

So for instance, in Zcash. I’m a founding scientist of Zcash. There’s a very specific circuit that deals with shielded payments, and it’s crafted towards that. At StarkWare, in our initial version one of DeversiFi, there was a specific set of, you know, the analog of a circuit, the specific set of polynomial constraints that had to deal with that. So it’s a little bit like building an ASIC, or some chip that is, you know, one per computation. Now with Cairo, basically we have a set of constraints or, you know, this math object that is the ZKP analog of a CPU. So the CPU is one chip that can deal with any computation, right? You send the computation as a program, as data, and it is executed.

That’s precisely what happens with Cairo. In fact, the name Cairo, the “C” comes from CPU and the next three letters are air, which comes from algebra. So it’s like a CPU algebraic object, and you put it once out there, and you can basically express all computations that you want and have one verifier that you don’t have to modify and change that deals with all of them. And again, the analogy you should have in mind here is that of a CPU. So you bought a laptop, there’s one CPU there till you replace it in two years time with a faster CPU, which likely will also happen at some point with Cairo. You’re going to run all of your computations, all of your programs, using that one CPU without replacing it. So that’s the same idea with Cairo. This wouldn’t work well if the way you had to express computations to the CPU was very hard, but it’s not. So there’s a programming language called Cairo that is very developer-friendly. It looks like a programming language, and now you can write all your computations in this language and they get automatically compiled and proofs for them are generated for the correct execution and put on-chain through this one verifier to rule them all.

Laura Shin:

Yeah, it’s really impressive. And you also earlier mentioned the Great Reddit Scaling Bake-Off. That was when Reddit and the Ethereum Foundation invited any Ethereum scaling projects to show how their project could be used to scale Reddit’s community points, which now I think has used by just hundreds of thousands of users, but later they want to roll it out to all 430 million Reddit users. So explain a little bit more about your participation in that and how it worked. I mean, you talked a little bit about what you were able to achieve, but just describe the experience.

Eli Ben-Sasson:

Yeah. So they gave a challenge in which they said, well, we have this point system, so think of these as some tokens that currently are administered over the Reddit servers, but they would really like to put them on the mainnet. And actually the way I think, it started, at least from our point of view is that we took one of the more popular subreddits and we basically took all of their transactions, I think it was like something like 12 million accounts if I’m not mistaken or something like that, or 1.2 million accounts, I forgot, and we just put one proof for, you know, for a whole bunch of stuff. I think this impressed Reddit. I’m speculating that maybe some other L2 solutions said, “Oh, you know, it’s not a fair fight, why don’t you do a challenge or something like that?”

Anyways, they said, okay, we have this grand challenge. You’re supposed to show how you can deal with 300,000 transactions, which is a huge amount, and show us how your scaling solution is gonna deal with them, had they been given to you as part of the system. And you want to sort of have these transactions be spread out over a period of five days because 300 transactions, you know, that’s a big amount of transactions. The time for this bake-off was given five weeks. So we did that in four weeks. And instead of spreading the transactions on five days, we put it inside 10 blocks and these 10 blocks, basically nine of them, the proof was one block, it could fit in one block, but there was all this data. They didn’t ask for it to be in rollup mode, but we said, you know, what the hell, let’s do it also in rollup mode. So we put all of the data on-chain, and then it took us 10 blocks. So instead of five days, we put it on the mainnet in one proof, you know, 10 blocks. Most of it is just data, not the proof itself. I think the Twitterati agreed that we sort of hit the ball way out of the park. I would love to have Reddit pick up the phone and say, “Hey guys, you know, you nailed it let’s work,” but I don’t know. I don’t know what they’re doing there at Reddit with respect to this thing… they haven’t announced anything since then, so, okay. We got some pretty nice PR and we moved on.

Laura Shin:

I guess we’ll have to see where that goes and what happens with that. So as we all know, there has been a lot of talk in not only crypto, but also in the traditional financial world about GameStop and Robinhood and scaling of these centralized solutions, I mean, or not solutions, but exchanges. And even if we just think about Coinbase and how they’ve had so many outages and when you look at how markets act the way they did in the GameStop period, which I think we all know is pretty much how crypto acts pretty much all the time, And we see that that caused problems for Robinhood, we know it does cause problems for Coinbase. I was wondering what you, what your thoughts were on how DeFi is positioned to address these issues.

Eli Ben-Sasson:

I’ll try to answer this. I mean, there are two aspects to it. There’s one, you know, the sort of a technical scale, and can you deal with it? And, you know, in the conventional markets, it’s not the scale of the trading systems that are the problem, right? They can deal with that. In the crypto world, this used to be the problem. I think with technologies like ours, this is no longer the problem. I mean, we can reach the scale of, you know, Nasdaq or Visa or, if need be even Alipay, WeChat. This doesn’t have to be the problem.

The other issue is that of the, you know, the sort of T+2 that everyone’s talking about. That the way the markets work is that finality, you know, you’re given some time for finality, and, in that respect, crypto is way better. Be it on L1 or L2, with instant finality, like dYdX perpetuals, it is basically as fast as the market or as fast as the blockchain progresses or the L2 progresses, that’s basically the state of affairs. Then there’s a third aspect of like centralization and censorship. You know, there were various rumors that I have no idea about, about, you know, various parties, not being happy with certain transactions and sort of stepping in. Of course, the decentralization of blockchains is what prevents this. So blockchains have instant finality. They are censorship-resistant. Had such a thing as GameStop been played on crypto, the main problem would have been scaling: can you deal with such an amount of transactions? But with our technology, it is kind of solved. So, you know, Nasdaq, if you’re listening to this episode, please pick up the phone, we’ll be happy to serve you.

Laura Shin:

So I don’t, I don’t know what the number of transactions per second was when the GameStop mania was happening. But are you saying that it was an amount that something like StarkWare and ZK rollups could handle?

Eli Ben-Sasson:

We can handle it… I mean, you know, we’ll have to put a lot of engine… Okay, the technology can handle any reasonable TPS that is currently, you know, in any financial system in the world. Okay.

Laura Shin:

You said it was 3000 transactions per second. Was that the number…?

Eli Ben-Sasson:

We displayed on mainnet something that is a 3000 TPS. This is something we displayed. But if we, you know, if we had to service Alipay with its 250,000 transactions per second, there’s nothing in the technology that couldn’t put that on L1. Of course, we would need to have a lot of engineers move to bigger machines and whatnot, but there’s nothing that prevents it. If we needed to deal, let’s say with Visa payload, which is in the 2,000-25,000 TPS, we could deal with that pretty much today.

Laura Shin:

And Antonio, did you want to add any thoughts on the GameStop issue and Robinhood?

Antonio Juliano:

Yeah. I mean, I think Eli gave a really good summary on that. The only other thing I would add is what I touched on before in terms of transparency. And I think that was a really big factor that went into just a lot of the chaos around GameStop as well. I don’t know exactly what the facts were, but I think one of the issues was that there was more than a 100% short interest on GameStop. Once those hedge funds had to actually cover because the price was going up so much, they literally couldn’t buy the shares that they needed to cover because it was more than the entire float of GameStop. Whereas something like that in crypto would have been, you know, people would have seen that coming. And people would have been able to see like, oh, okay, well we know like all of them, the short positions on DeFi, like total this much. We know the float is this much. So it would’ve just been able to have a lot more transparency around the market, which I think is important as well.

Laura Shin:

Yeah. Or the systems even could have been designed where it would not be possible to short more than existed. Something else I wanted to ask about is a trend in DeFi that differentiates between these fair launch projects and then VC projects. And both of you have gotten investment from big-name VC firms, like dYdX has investment from Andreessen Horowitz, Polychain Capital, One Confirmation, Three Arrows, Defiance Capital. StarkWare has investments from Paradigm, Sequoia, Coinbase Ventures, Scalar, MultiCoin Capital. And so I know neither of your projects has launched tokens yet, but I just wanted to hear your take on this debate between fair launch tokens versus VC tokens.

Antonio Juliano:

Yeah. It’s an interesting question. I definitely don’t think that there is one that’s objectively better in terms of the path for any given system. I really think it has to do a lot with the complexity of the systems that you’re building. For example, like let’s take us or, I mean, actually, StarkWare is a great example of this too. Like the things that we’re building are just very technologically complex. Sure you could come out of the gates and be like, “Hey, I’m dYdX and have this dream to like someday hopefully build a protocol for perpetuals. And this was like, theoretically, how I’m going to do it, like give me your money for a token.” But I don’t really think that that’s like an optimal way to go about fundraising or an optimal way to go about company building more generally.

People always ask us constantly, “why have you not done a token yet? Like, it’s pretty obvious that you should do one.” And I agree, but trust me, all the things going on on in the market are not lost on me, but the thing that I always tell people, and the thing that I believe is we’re not doing a token because we’re like, you know, fundamentally opposed to them or something like that. We’re not doing a token right now because we believe that it’s more important for us to build a really great product first. And then, once we have a really great product and we have strong product-market fit, we have a lot of great users, when you can throw just fuel on the fire on top of that with the token, then you have an opportunity to hit our goal of like potentially becoming like one of the biggest exchanges in crypto, period.

So for us, like that’s always been, the goal has been really big to become one of the biggest exchanges. And we really feel like this is the best, like expected value, like way for us to do it long-term to go through. We’ve been around for like three and a half years now, which in crypto is literally forever. We just tried to go through and take these like three and a half, four years, like building an amazing product. I think we had kind of started to hit our stride on layer 1, but then we just ran into all of this nonsense with having to pay hundreds of dollars in transaction fees, and there a bunch of like other product limitations that we had on layer 1, like finality was pretty long, like block time, and things like that that are now solved on our new product.

And we’re really excited about our new products. I mean, you know, Eli touched on this before, but we’ve literally had 50,000 people sign up for our waitlist on our alpha. And it’s getting to be more every day. So I think a lot of other people are excited about that as well. And again, that’s like with no token or anything like that, so it’s just organic interest in our actual product. And then like once we have strong product-market fit then if we, you know, later and no guarantees or anything like that, but later if we go out and launch a token that’s how we think we become really, really big. And then we feel like the amount of the company and therefore the like hypothetical tokens or whatever that we like sold to investors would be well worth it. Because now we have just like four years of development experience behind us and a lot of expertise to continue to build great products.

Laura Shin:

Yeah. And I meant to mention earlier that I did an interview with Antonio, actually, I don’t know what year that was. Do you know what year it was?

Antonio Juliano:

A long time ago? It was either like 2017. I think it was 2018. So it was right after the ICO hype. So probably talking about the same thing.

Laura Shin:

Yeah. So you know, clearly the project has evolved since then. It was quite different. And so people should listen to that just to hear more about Antonio’s history. So it sounds like you probably will launch a token. And so should we expect it would be a governance token, and maybe you would, you know, add a little liquidity mining scheme on top of it.

Antonio Juliano:

Yeah. I’m not going to talk too much about it in particular, but it’s not rocket science, probably. We’re taking a look at all the things that have been successful, like in the DeFi space. Those things that have been successful I would say, are governance and liquidity mining. So we want to fulfill our goals of being able to grow in the maximum possible way. You can use your own intuition about what we might do with such a token. But again, it’s not a hundred percent that we’re going to be doing this. Like even in this bull run or whatever, no matter how long that lasts, we’re not going to launch one until we’re really happy with the current state of the product and we really think and honestly believe that once we throw fuel on that fire, that we’ll be able to grow. And we just appreciate everyone’s patience until then. And of course, it is like in our long-term plans, like I was talking about earlier in the episode to decentralize more of dYdX and to not just have, say, for example, like our contracts be controlled by like some multi-signature contract forever. And clearly, that goes hand in hand with governance too.

Laura Shin:

And Eli, so you also talked about how StarkWare will be decentralizing. Do you plan to do that with a token?

Eli Ben-Sasson:

Yeah. So, I’ll get to that in a minute. I won’t comment much [on the token]. I just want to add to what Antonio said about the fair launch versus a VC. We incorporated three years ago at the height of the ICO craze and everyone was saying, “Oh, you know, you must do an ICO.” That was the previous version of it. And we didn’t, even though, you know, a Filecoin was raising a quarter billion and Tezos and so on. Those were the happy days. What we said to ourselves and also to other projects we were talking to who did do tokens, was that you’re basically going public from day one. There’s this price ticker, right, that everyone sees, you know, your employees, the outside world, and you haven’t really done anything yet. So it could be a very demoralizing experience.

There are these market fluctuations. So that’s another aspect of just, you know, managing expectations of your employees and everyone, and then the public as you’re sort of building something. We wanted to build something very quietly and get top engineers and mathematicians to work very patiently on this. So, that was the obvious route for us, but others have gone the other way. Now regarding, you know, decentralized StarkNet tokens, I won’t comment on tokenomics. I will say this: it’s clear to us that an L2 that is decentralized will need to have proper crypto-economic incentives for the various players out there. We’re going to have decentralized proving services that are operated by numerous players that you need to have coordination among them. There are going to be sequencers of transactions. You need to take care of payments to the L1, amortize the gas cost, and so on and so forth. All of this begs for game-theoretic mechanisms, design, and crypto-economic mechanisms. But those are yet to be determined. And just like we waited patiently with any sort of tokenomics until now, we really need to understand how StarkNet will be governed and administered before we can make decisions on the right mechanisms.

Laura Shin:

All right. So neither of you want to comment too much on any tokens, but are there any other developments coming up this year that either of you wants to mention?

Antonio Juliano:

Yeah, maybe I can start. I think the main thing that we’ve just…. I guess it’s even hard for me to like, even for the next year… We’ve just been so heads down and focused on building out this new L2 product with StarkWare. The other thing I think I didn’t mention but would like to talk about too, is it’s not just that we took our layer 1 product and put it on layer 2. We literally built an entirely new set of products with new servers, a new UI, a new everything. So we’ve been working really hard alongside the protocol work that StarkWare has been doing then. And I’ve kind of been in the thick of that as well in terms of coding and stuff. So yeah, certainly we’re about to have the public launch in about a few weeks from now. That’ll be really exciting. For further on down the line, I think the main thing we will be focusing on for the rest of this year is just further decentralization of the protocol. So starting with kind of decentralization of like the governance of our contracts and things like that. And then starting to think like, how can we decentralize, like the order book and the matching engine components that we still run in a centralized way right now. So no timelines or anything like that for any of that, but that’s what we’d be focused on for the rest of the year. And I’m excited about that. Like, it’s really hard problems, but that’s how we’re really going to realize the long-term vision of what we’re building.

Eli Ben-Sasson:

So on our side, first of all, I’m very excited and anxious about the public launch of dYdX, which is going to be the first major public event, the closest public event, on our horizon at StarkWare. We’re really, really excited about this. Shortly after that, Immutable X is going to be launching. This is for NFT and massive off-chain minting of NFTs, which is a big, big problem that no other solution can deliver right now and Immutable X is going to do that. That’s going to come very soon over the next few weeks. Looking a few months ahead, we’re going to be launching with Badger DAO an AMM on L2 and that’s another exciting thing.

I mean, those are the projects that we can sort of announce right now. There are other things, of course, being worked out, and we hope to announce them really soon. Another exciting thing is basically the launch of phases one and two of StarkNet called the Planets and Constellations. That’s going to be in a very small number of months. These will allow developers at large to just write smart contracts and deploy them and benefit from the same scale that dYdX is going to have when it goes public. And I want to mention, in this respect, and this also ties to this question: we’re asked a lot you know, how do we, how can we participate in this thing that is you know, StarkWare’s journey… it started with the things like that, that people ask about, you know, tokennomics and whatnot. And there’s only one very simple answer that I, and we at StarkWare, always give them, which is: go learn Cairo. This will be like if you asked about, you know, the early days of Ethereum and what was the best thing you could invest your time in, it would probably be to learn solidity or whatever precursor was to there. So right now, this, this language, on which all of StarkNet is going to be built with all its scale, is Cairo. It’s already accessible. Go to cairo-lang.org basically start learning. There is a bunch of tutorials there; there are tutorials for running your own AMM. So you can build today your own AMM on layer 2 with much lower gas fees, there are voting and signature aggregation tutorials that you can do there, and again, you can start building your own products and have them integrated into StarkWare. I think this potential of having a developer ecosystem that learns Cairo, writes in it, and then you have this massive amount of devs on StarkNet. This is very exhilarating to us.

Laura Shin:

Yeah, it definitely sounds interesting to me because yeah, we see all these Ethereum killers trying to come and take developers to these other blockchains and attract them with you know, DeFi capabilities on their platforms. And yeah, I could see the appeal to a developer who wants to stick with Ethereum. All right. So we will have to keep an eye on that because that does sound interesting. So thank you both so much for coming on Unchained. Where can people learn more about each of you and your work?

Antonio Juliano:

Yeah, for us the best place to start is just going to our website dydx.exchange. There, you can sign up for our private alpha, which will live really soon. You can join our discord. If you want to ask me anything, literally tag me in our discord and I’ll probably answer. So just head over to dYdX.exchange and check everything out there.

Eli Ben-Sasson:

On our side, first of all, follow us on Twitter @StarkWareLTD. You can check out our website is starkware.co. Finally, if you’re a developer and want to start building in Cairo, you go to Cairo-Lang.org.

Laura Shin:

Great. And I meant to mention earlier that Immutable X has been a previous sponsor of my show. But that also, you know, when you mentioned that, I hadn’t been aware of the difficulties of minting NFTs, like in terms of gas costs. I only that’s something I only found out about recently, so I will also be watching that. I think that sounds very interesting. All right. Well, thank you both so much for coming on Unchained. This has been super fun.

Thanks so much for joining us today. To learn more about Antonio, Eli, dYdX, and StarkWare check out the show notes for this episode. Don’t forget: you can now watch video recordings of the shows on the Unchained YouTube channel. Go to youtube.com/c/unchainedpodcast and subscribe today. Unchained is produced by me, Laura Shin, with help from Anthony Yoon, Daniel Nuss, Mark Murdock, Dan Edlebeck, and the team at CLK transcription. Thanks for listening.