Said Loi Luu, head of Kyber Network, the second-largest decentralized exchange
|Jul 1||Public post|
Good morning defiers, hope you had a great weekend! For this week’s interview I talked with Loi Luu, co-founder and CEO of Kyber Network, the second-biggest decentralized exchange by volume.
Luu said inefficiencies in decentralized exchanges are creating profitable opportunities for traders, but that the space is still too immature for professional investors and institutions. He talks about why he thinks Kyber is uniquely positioned to connect all other decentralized exchanges, about how governments can limit access even to exchanges that are fully on-chain, about the difficulty of preventing front running and how Libra can be a threat to crypto.
The interview has been edited for brevity and I bolded what I think were the best quotes.
Camila Russo: What led you to create Kyber?
Loi Luu: I started doing my Phd in the National University of Singapore back in early 2014 on applied cryptography and cryptocurrencies and since early 2015, I started doing full time research on Ethereum. Together with my team, we developed a couple of open source and nonprofit tools. One tool allowed people to check whether their smart contract had any vulnerabilities before deploying it and we also developed a decentralized mining pool. These tools were intended for a relatively small, specific audience. By the end of 2016, we wanted to work on something that could create an impact in the entire cryptocurrency community.
In 2017 we started reviewing decentralized exchanges and we couldn't find any solution that was easy to use for the mainstream user, and also no solution that supported the smart contract ecosystem. Most of the existing solutions were hybrids as they still involved some centralized server.
We thought that was a super interesting problem to solve and after a few months and several iteration we thought of the idea of Kyber and have been working full time on it since then.
CR: So the main goal behind Kyber was to connect smart contracts to a decentralized exchange without having to rely on centralized solutions?
LL: That was the initial initial goal back then when we wanted to run something that’s completely trustless. That means we don't have to trust anyone in the middle. There wasn't an easy way for people to develop Defi applications on a smart contract and integrate with a decentralized exchange. So the goal was to develop something that is first of all, natively available on a smart contract so that people could integrate with it easily. And secondly, make it completely trustless so that the end users when they need to interact with the smart contract, they don't even need to trust Kyber. We launched on main net in April, 2018.
CR: How has the initial concept developed and changed since then?
LL: Part of the design is still the same except that we opened it up to allow more people to join, to provide liquidity and we allow different kinds of strategy that the liquidity providers can use. Secondly, we make it super easy for people to integrate with a smart contract. So they only need to call one smart contract function so they can talk to all the available liquidity providers.
We integrated with other Dexes like Unisawap, Oasis Dex, Gnosis DutchX, and many Defi applications like Nuo, bZx and even Melon Port, and also cryptocurrency wallets to allow the end users to trade within the wallet so they don't have to connect to any Kyber server. They only need to query the smart contract.
CR: When you connect with other exchanges like Uniswap, does it mean that you're both sharing each other's liquidity? If you don’t have volume for a specific token, but Uniswap does, will Kyber users be able to access it from Uniswap?
LL: Yeah, so we aggregate all the on-chain liquidity sources via Uniswap, Oasis Dex, Eth2Dai and other guys and provide one single interface for anyone to talk to all these liquidity sources so we can guarantee that the users can just talk to one single function and they can get the best rate available on the blockchain.
CR: Do you see all Dexes integrating in the future? Will it be like one big connected exchange?
LL: I think Kyber is in a unique position to connect all the other liquidity sources because we don't dictate on any mechanism for people to provide liquidity. Currently if you look at other existing platforms, Uniswap uses an automated market-making model, and if you look at something like Eth2Dai, they maintain the on-chain order book based model, and Gnosis DutchX works in its own away. But at Kyber, we don't dictate any model at all. I think that that flexibility offers a lot of room for other market makers to join.
CR: Wanted to ask you about geo-fencing. You talked about how you wanted to make sure to build a decentralized network, but other Dexes, like Bancor, have still needed to exclude users.
LL: In the case of Bancor, they exclude the users from the website. They banned all the IP addresses from the U.S., but I think technically they cannot, on a smart contract level, they cannot do anything at all because there's no way for the smart contract to know where the user is.
This is one of the main reason why we focus on developing Kyber as a fully on-chain platform, that means everything is on the smart contract and users can find a lot of different interfaces to get access. They can go to any wallet among the 20 or 25 different wallets that are already integrated with Kyber to trade. They don't necessarily have to visit a website.
CR: But that's a risk for any exchange, even if it's completely on chain and decentralized that a country can still block access to it through the website and just ban the wallets.
LL: Any country can do that if they want to but there is always the option that you just run your own full node and then you can interact with any contract from your own full node, so you don't have to rely on any software, on any wallet or any website.
CR: So the difference there with a centralized exchange is that you still have the option of somehow getting access to the exchange, but with a centralized exchange, if a country just blocks the website, then there's like no way of of using it.
LL: Right. For example, if someone blocks IDEX’s website, then there's no other way for you to integrate with the exchange.
CR: So I also wanted to ask you about front running and bots on decentralized exchanges. The fact that it's slower to trade on a decentralized network because you have to like wait for confirmations to go through, opens up more space for trading bots to front run less technical traders. What's your view on thaT and how can that be resolved?
LL: This is always a topic that gets brought up when, when it comes to decentralized exchanges. In a public decentralized exchange it’s always possible and technically very hard to prevent it. There are solutions, but you usually have to come compromise on transparency or on user experience.
But one property of front running on decentralized exchanges is that anyone can do it because it's public so it opens competition a little and potentially reduces the incentive for people to front run because they might lose out if there are more people doing it
CR: That's, that's true. But if that's still a possibility that there will be front runners and it’s hard to prevent or regulate, will it be a barrier for professional traders on institutions to ever use Dexes?
LL: I think that's a very interesting question because if you look at it, front running, it's always possible on centralized exchanges as well. And it's worse because if someone actually does it, there's no way you can verify it. If the centralized exchange themselves front run you as a client, there's no way for you to verify it.
Another point is that front running is going to be harder when the block count is shorter and there’s less time for people to act. Traders can also just try to set the maximum gas price to avoid anyone else getting in before them –though that doesn’t help is the miners themselves are doing the front running.
The main concern for decentralized exchanges is not really front running, but rather making them more usable and trying to integrate with more application as well.
CR: And with the advantages that, as you you said, Dexes are more transparent, maybe that will actually attract professional traders and institutions in the longterm.
LL: Professionals and people who know enough to actually front run might be interested in doing other activities on decentralized platforms. For example, arbitrage, because there’s still a lot of inefficiencies in the market and a lot of rooms for them to earn money from these platforms. But currently volumes and the market size on decentralized exchanges is quite small for anyone to be serious in doing fun running and even arbitrage at a large scale.
CR: About the general decentralized finance space, what's looking the most interesting to you right now?
LL: What I’ve found quite interesting so far is that many like decentralized finance platforms, are trying to connect to each other. For example, Kyber, connected with many, many different platforms over different use cases that we have, Nuo and bZx which offer lending and trading, we also have like Melon Port that offers a decentralized fund management platform. A good example also are the guys from InstaDapp. What they do is they just look at different smart contracts and they try to combine them together to offer new use cases to the, to end users.
CR: And going forward, what aspect of DeFi do you see gaining the most traction?
LL: I think that the lending market is currently very hot because for the first time people can actually lock their coins that they're not using and get some interest from it. Then on the other hand, people can, like, people can borrow coins from other people in the Internet without knowing who they are and just use the coins for whatever purpose that they want. I think a lot of the focus will be on that.
CR: Switching gears, I know you have some strong opinions on Libra. Would love to hear more about that.
LL: I’ll raise three concerns about Libra. The first one is privacy. Facebook is well known for mishandling user privacy and, and now with the launch of Libra and Calibra it will allow Facebook to collect even more data from users. They already have the social data now they’ll have financial data. It might lead to a monopoly situation with the data.
The other aspect is, we have Bitcoin, Ethereum, which are public cryptocurrencies. They’re the people money, the people coin or the internet money. We have central bank coin and now we will have a new kind of currency; the corporate coin. They’re getting all the big guys together and trying to create a platform that everyone will have to join. If you don't work with them, then there's no chance for you to compete against them. People believe in cryptocurrency because they don't want to give that much power to governments and central banks. And now we have a new big guy on the street who is going to have a lot more power.
Also it’s not clear how governance is going to be done. Whether people can actually cast any vote and make any, make an impact. They said Libra is going to be permission-less in the future, but there are no details about how it’s going to be done.
CR: Is it a threat to crypto?
LL: I mean, I'm not sure whether it's going to be a threat at the moment but I think in the long run, yes. I think at the moment it gets a lot of publicity to blockchain and cryptocurrency and public validation from one of the larger companies. But when they launch if they can address all the regulatory concerns,I think it's going to be a threat for public cryptocurrencies for sure.
CR: Is there a possibility that it can be an onramp for bitcoin and other crypto? Like if people are switching from Libra to Bitcoin at some point?
LL: That's a good idea actually, because if people already have Libra then they already have a wallet and private key infrastructure and it's going to be much easier for the users to switch from one cryptocurrency to a different one. People will have to develop some interoperability solution to connect between Libra and other public blockchains.