Respondent: Shawn, developer of the Hyperliquid ecological LSD project

The recent Jelly short squeeze incident and a series of public opinion influences have become hot topics in the industry. In response to this matter, while accusing Hyperliquid, many people have also advocated the superiority of "centralized exchanges" for the sake of benefits.

In this regard, the "Xianrang" team had an in-depth conversation with Shawn, the developer of the LSD project in the Hyperliquid ecosystem, and discussed topics such as Hyperliquid's technical architecture and current status, the shortcomings and optimization methods exposed by the Jelly incident, whether the KOL site-centralized exchange attacks Hyperliquid, and the future development trends of other tracks.

The following is a text version of this interview, and everyone is welcome to read it.

1. Wuyue: Let me briefly introduce myself first?

shawn: I entered the industry in 2017 and have been working on technology, including wallet, exchange backend support, and public chain-related development. I have done some blockchain network maintenance and on-chain protocol development at Mantle, and now I am working on projects within the Hyperliquid ecosystem. What I am doing is similar to LST within the Hyperliquid ecosystem. Users can pledge HYPE tokens and receive the LST tokens we issue.

Users can store HYPE in the protocol to automatically obtain income. These assets will be used for Validator pledge of the Hyperliquid chain, and then predict the opportunity to send the income data back to the contract on the chain. The entire asset flow is self-custodial, which basically achieves decentralization. The asset security on the user side and the protocol side is guaranteed to the greatest extent.

2. Wuyue: So you should know more about Hyperliquid than most people? So, can you briefly explain the technical architecture of the project?

shawn: Based on the current public information, one of the biggest features of Hyperliquid is that it not only creates a blockchain network dedicated to DEX, but also builds on this chainLayer2. The Layer2 I am talking about here refers to HyperEVM, which is the EVM compatible chain. Hyperliquid converts user-side behaviors, such as pending orders, into an on-chain event, and then puts the event on the chain, so it is a decentralized exchange that goes on the chain throughout the process. The consensus it currently adopts is DPOS, which has made some modifications to the Hotstuff consensus algorithm. Simply put, it adopts a modular approach and abstracts the consensus layer and DEX-based application layer classification.

One ​​time ago, a North Korean hacker tried to attack Hyperliquid's validator network, but failed. A problem was exposed at that time. At that time, Hyperliquid had only 4 Validator nodes, all of which were run by the official. After that incident, Hyperliquid introduced more Validators, currently there are 16 nodes, and their future goal is to reach about 100 nodes.

At the staking level, we can regard Hyperliquid as a regular POS chain. Currently, the official website supports users to pledge tokens to run Hyperliquid verifier nodes, and the system will distribute incentives based on the performance of your node.

We have finished talking about the POS chain of Hyperliquid, and then briefly talk about the Layer2 of HyperEVM. There are some precompiled contracts in it, which is not like the precompiled instructions we see in Ethereum, and is closer to the system contracts provided by Layer2, a traditional sense of optimism. We know that due to the characteristics of the EVM itself, many functions or functions cannot be implemented internally in the EVM, and some external dependencies are required, and system contracts are such external dependencies.

The feature of HyperEVM is that it provides multiple system contracts that connect its DEX application chain and EVM-Layer2 to better adapt to its unique business needs. Then, since it is an EVM-compatible environment, it facilitates the migration of applications in the Ethereum ecosystem. There are actually various scenarios such as Swap. Users can initiate operations within HyperEVM's contract through system contracts to exert influence on the DEX of the first layer. All interactions you generate with HyperEVM will be submitted to the first layer of Hyperliquid POS chain.

Current HyperEVM testMany precompiled contracts have been implemented online, and the functions are more complete, but the main network has not yet fully opened up such precompiled contracts. The main reason is that many components of HyperEVM are not mature and require a period of iterative upgrade. We are currently building all the features of Restaking based on the testnet.

There is another interesting thing. I wonder if you have noticed that when Coinbase was working on the Base chain before, they proposed something called fast block, which is fast transaction confirmation. Generally speaking, it takes several seconds or even ten seconds for a normal public chain to produce a block. A regular Layer2 may be a block in 2 seconds, while Base may produce a block in milliseconds before.

But before Base proposed the fast block concept, Hyperliquid made a plan, dividing the blocks in HyperEVM into two categories, one is large block, and the other is small block, one is large block, and one is small block in one minute, and one is small block in two seconds, which can ensure that users can quickly confirm simple operations such as transfers.目前看，这是一个分别保证快速执行和EVM内复杂逻辑执行的比较有效的方案

这样做虽然有合理之处，但也会带来用户侧的一些体验割裂，比如说，一般用户在以太坊上可以随便发一笔交易，但在Hyperliquid上面，需要预先在账户里做一些设置，才能让交易提交到大区块里。 The confirmation time of this large block may be one minute, and from the perspective of experience, there will be some things that make you confused.

The second is that when we do arbitrage on the chain, the experience will be different from that in Ethereum, because Hyperliquid has two block types. But its basic account system is shared. In this case, we need to optimize the arbitrage strategy and have more personalized strategy configurations to achieve the best results. These are the differences between it and ordinary EVM chains.

3. Wuyue: How is the development status of HyperLiquid and HyperEVM projects you are seeing? For example, the situation in the ecosystem

shawn: If you look at it from the perspective of developers, they currently provide developers with not much. If you look at it from the node level, they have not opened source code yet. What we can get now is a mirror file, that is, a version obtained after compiling the original code, and run the node based on it, but cannot see the original code. So, although we can run the Hyperliquid node,But its code is closed source for us.

Precisely because its code is closed source, we cannot determine whether it contains the logic related to MEV operations that the official added itself. This has caused some difficulties, that is, we cannot directly reuse some experiences in the EVM ecosystem, such as not seeing the transactions waiting to be listed in the transaction pool, which is not friendly to MEV players.

然后再看生态层面的工具，他们其实大量借用了以太坊生态里已有的各种工具，所以很大程度上与以太坊生态给人的体验感对齐。 For example, you can directly use Metamask to connect to Hyperliquid.目前Hyperliquid应用链对应的浏览器提供的信息比较少，像最基础的账户交互历史、资产余额这些东西都不是特别详细，但它会有一些 API 给你去查询。 So things like browsers still have room for optimization.

目前来看的话，Hyperliquid已经能够最大程度的满足dex交易，不会显得非常割裂，然后各种defi项目在HyperEVM上也都有，总体上生态内应用还比较早期，很多项目都是最近两三个月才开始启动的。

4. Wuyue: In this short squeeze incident about Jelly, in which aspects do you think Hyperliquid has exposed its shortcomings? How should such problems be solved in the future?

shawn：首先Jelly是一个非常早期、流通盘很小的一个项目，这种小项目被上到Hyperliquid的期货列表里，显然他们的风控策略没有做好，对这种高危高风险项目的处理不到位。

Hyperliquid's coin listing process does not require permission, as long as you obtain the coin listing permission through the Dutch auction.它引来了Jelly这样一个低市值的项目，有人在Hyperliquid所内开了一个头寸非常大的空单，甚至比Jelly的总市值还高，然后在场外不断的拉升Jelly的币价，所以我们能够想到，一方面持有大头寸的空单，然后在场外不断拉升现货价格，那么我在所内的这个仓位肯定会爆掉。

同时，Hyperliquid作为DEX，以资金池和金库作为开合约者的对手盘，比方说我们在那边开一个多单，他就会开空单，作为一个对冲，或者在我们去平仓时，能够有一A liquidity provides an opponent's plate. The HLP treasury provides the function of taking over the closing position. When the user issues a short order, the HLP treasury will issue a corresponding order with him. When the short order is exposed, he will close the position. So during the closing process, you will continue to buy and close the position, and then there will be a spiral inside.

So his problem is that when the off-site price rises, the short position will inevitably be liquidated, and then the short party continues to withdraw margin to intentionally increase the position risk. When the holder's own margin loss is minimal, the risk of additional position penetration will be borne by the HLP.

This is also a very controversial place at present, lies in the mechanism design of passive market maker Chizi. When trying to use a fixed strategy to entrust these funds to close positions, it will definitely have risks. When this mechanism is abused, for example, more and more small-cap projects short-shrink HLP treasury in this way, then all users in HLP will bear the losses of being short-shrinked.

The problem exposed here is that when we manage some small-cap risk products, users cannot be allowed to open positions of any size. We must limit the position size and leverage ratio that users can open based on its market value, degree of chip dispersion, and depth of over-the-counter liquidity.

Many centralized exchanges have a detailed risk control rating of small-cap projects, especially futures projects, for example, Binance and OKX futures have pilot zones, and there is a complete risk control system for the project's own market value, degree of chip concentration, depth of the on-chain liquidity pool, how many coin prices are sourced, whether it will be manipulated by oracle prices, and whether the price indicators are stable. A certain exchange seems to have clear regulations that small-cap projects can only open contract positions of up to US$250,000. Hyperliquid's experience in this area is obviously insufficient, which leads to HLP being forced to bear the risks of contract liquidation and opening.

Considering that Hyperliquid started with a market maker, market makers may pay more attention to the liquidity of a token itself, rather than considering the problem from the perspective of the platform. After a whale used 50 times of leverage to make trouble with Hyperliquid, Hyperliquid lowered the highest leverage ratio of BTC and ETH, but this decision did not seem to have been fully evaluated to prevent similar risk scenarios.

In addition, Hyperliquid's risk management in HLP vaultsThe dynamic relationship adjustment mechanism between OI Caps and Market Caps needs to be added. For high-risk assets, for example, we have now held a 1M contract long order. When the market value of this asset continues to decline, on the one hand, the system must gradually reduce its positions and dynamically adjust all contract positions holding the asset. Then, after the market value falls, the maximum position may only be less than 1M. These means are used to reduce the risk of vault taking over user positions closing.

So this time Hyperliquid encountered Jelly's short squeeze, I understand that when he was beaten, I must stand upright, and I hope to pay attention to this lesson, but this does not mean that Hyperliquid has done anything that is ashamed of the user, and this is not a serious problem.

Considering that Hyperliquid's listing is in the form of auction without permission, this decentralized listing method is actually beneficial to the development of the industry, but it will bring hidden dangers to the trading platform itself. Perhaps in the future, Hyperliquid can introduce AI Agent+ on-chain voting to screen currency listing objects, and introduce and remove currencies through AI evaluation systems and open source and open risk control and governance systems, thereby reducing the shortcomings caused by the permissionless currency listing method.

5. Wuyue: What do you think about the fact that many KOLs obviously take centralized exchanges such as Binance and OKX, and attack Hyperliquid for "not decentralization"?

Shawn: This is the question I want to answer the most. Binance has clear plans for its own. They will vigorously develop BSC, attract users and assets to feed back to their exchanges, and walk on two legs. So in fact, judging from the bull market last year, many KOLs have made money and achieved results on the Solana chain. Then after seeing this, CZ was preparing to move what happened on the Solana chain to BSC and repeat it. They took the initiative to push up the memecoin on the BSC chain, but we won’t say much about who made more in this process.

As for Hyperliquid, it released the main network in November last year. Based on its architecture system, it can also make more expansions on the application chain side, such as the Solana VM and MOVE VM can also be moved into its ecosystem, and it will be made into a Layer2 just like HyperEVM. It is obvious that Hyperliquid can expand horizontally and is more flexible than centralized exchanges, so Hyperliquid is actually here to dig the grave of centralized exchanges, and there will inevitably be commercial competition.

But the question is, what exactly is the attitude of the founders towards competitors? Is it more to improve themselves or to pull others into the water? This is something that is very interesting and thoughtful. We don’t talk about pulling or how, in this cycle, OK and Binance are definitely the biggest impact on the on-chain exchange. Logically speaking, they will spare no effort to reduce the impact on themselves. These are all open conspiracies and are normal business competition.

But I think there is something more puzzling. Why are there so many KOLs who suddenly advocate centralized exchanges? Then, HLP has penetrated its positions, but Hyperliquid is not good. This critical method obviously raises the local phenomenon to the whole. Can you tell that you are not good? Many KOLs criticize Hyperliquid for the impression that people give people.

For basic industry competition, I don’t think it is necessary to beautify or ugliness anyone, and there is no need to advocate how centralization is better than decentralization. The key to the problem is not here.

In addition, competition is not shameful. Everyone hopes that they can promote the progress of the industry and do good things, but centralized exchanges have original sins and need to maintain a conscientious attitude. As a role who takes commissions based on transaction volume, you should do your job well, instead of doing something that is both pretentious and stable. Splashing dirty water on others, or attacking your competition in the form of a KOL matrix is ​​definitely not a high-end thing.

6. Wuyue: What is your judgment on the future development trend? What do you think of the exchange track?

shawn: If you talk about this, we can talk about it slightly broadly. This bull market started from November last year to the beginning of this year, including AI The craze of Agent concept coins all happens on the chain, and the emergence of these phenomena is actually an inevitable trend. Various tools on the chain platforms such as gmgn are becoming more and more mature, and users are becoming more and more good at discovering early projects, and being bored with VC coins, etc. are all natural things.

Now many projects will not even do VC rounds, but directly issue coins in the form of on-chain PVP, and those P players who are good at mining vote with their hands to vote for some high-quality projects. This is actually very terrible and will make VCs have no food to eat. So at present, not only the exchanges are anxious, but VCs are also anxious. They will feel that they need to go down to dig money on the chain like ordinary users.

Including HyperliquThere are many projects in the id ecosystem that do not receive grants, and they all rely on HYPE airdrop to maintain project expenses. Including this project I am working on myself. Moving more things to the chain will help develop an open ecological system rather than becoming increasingly closed.

Hyperliquid is not becoming more and more closed like Ethereum. First, it provides users with generous airdrops, and then the platform has tens of billions of dollars in transaction volume every day, and all the fee income is used to repurchase HYPE tokens, which leads to some better results. In the future, HYPE is a deflation model, which is generally relatively stable. To put it bluntly, more people who are good at discovering value need to pay more attention to what kind of people are doing in the Hyperliquid ecosystem.

So I think, as an exchange itself, what attitude should it be to promote the development of the industry to support the industry, or do all kinds of sarcastic evil. Including that after Bybit was stolen by North Korean hackers, CZ also tried to back-prick Bybit, everyone can see this.

Centralized exchanges cannot lead the blockchain industry. They are just pumping services. Although they have a large scale and good revenue, they will never be a leader. In the future, projects and platforms are destined to be decentralized. The gradual decline of custodial centralized exchanges will be a foregone conclusion. Users' demand for self-custody, anti-censorship, and anonymity is the general trend.