Original Author: Hahaho@DAOrayaki.org

Podcast address: https://www.ximalaya.com/sound/ 589658039

In the past two days, the topic of FTX and Binance has continued to ferment in the circle. The latest development is that although FTX has declared bankruptcy protection, its exchange is continuously being attacked by hackers. To describe it as a mess, a wave of ups and downs. In my personal thinking, many problems have emerged from this still-fermenting event, such as the direction of market supervision, issues involving monopoly law, issues of centralization and decentralization, liquidity and redemption issues of exchanges, etc. No matter which one, it will have a very profound impact on the future development of the industry.

Shaun

Hello, everyone, welcome to this issue of YakiTalki, I am the host Shaun. In the past two days, the topic of FTX and Binance has continued to ferment in the circle. The latest development is that although FTX has declared bankruptcy protection, its exchange is continuously being attacked by hackers. To describe it as a mess, a wave of ups and downs. In my personal thinking, many problems have emerged from this still-fermenting event, such as the direction of market supervision, issues involving monopoly law, issues of centralization and decentralization, liquidity and redemption issues of exchanges, etc. No matter which one, it will have a very profound impact on the future development of the industry. As a member of BUIDler in the industry, DAOrayaki pays more attention to issues and events that have an impact on the BUIDL circle. This is the Merkle Tree we are going to talk about today. After this exchange war, various exchanges began to "compete for me" to officially announce a technology called Merkle tree, trying to prove the transparency of asset reserves and transaction transparency of their exchanges through this method . What exactly is a Merkle tree, what kind of problems it is used to solve, and what impact it has on the industry. With these thoughts in mind, today we specially invited the core developer of Starcoin and the initiator of MoveFuns DAO, Jolestar Teacher, let's talk about the Merkle tree and some exclusive extensible thinking brought about by this incident.

Welcome to an old friend of the community, Mr. Jolestar.

Jolestar

Hi, everyone.

Shaun

Before I officially start talking about today's topic, I'm curious about your interpretation of this "fairy fight" incident?

Jolestar

It must have been beyond most people's expectations. When I paid attention, I didn't expect that the "fight" would go in this way. Some people thought that the boss would fight with the second child, and the third child would suffer. It was really unexpected that things would move so fast, and FTX was almost powerless to resist. In addition, according to my observation, at the beginning of the incident, everyone talked about withdrawing coins. Users who have been affected by the LUNA incident will be more vigilant and actively withdraw coins. Some people took a wait-and-see attitude at the beginning of the incident, and eventually lost their assets on FTX.

So everyone has also come to a conclusion, once there is trouble, regardless of whether it is true or not, the coins should be withdrawn first. But this behavior pattern will have a great impact on centralized exchanges. From the perspective of the game, this is a multi-player game. Everyone will form experience, choose this way every time, and the reaction will become faster and faster.

From a negative point of view, whether it is the market price or the entire industry, everyone is prone to pessimism and panic; but from a positive point of view, everyone has long realized that centralized exchanges are a very large single-point risk. Decentralization is to eliminate this single-point risk. Before DeFi, many people’s assets were stored in exchanges, and there were very few assets on the chain. DeFi prompted users to migrate assets from exchanges to chains. This incident also made more People realize that the way they keep their assets is the main priority.

This is a good thing for the industry in the long run, and everyone will pay more attention to decentralization itself. In the past, for the sake of early and rapid development, decentralization may be sacrificed, seeking to build up the overall market first and attract users. But after this incident, everyone gradually realized that if we want to move forward more smoothly, we still have to return to the decentralized route.

Shaun

I opened Twitter today and found two particularly notable features. One is that some people think that from now on, no matter whether the market information is true or not, as long as the information comes out, the first thing to do is to act first, and remember not to think twice before acting. For reference, everyone seems to be in a state of panic and caution about the entire market now.

The second is that KOLs have begun to widely propose to store tokens in cold wallets, and some cold wallets have begun to be more actively promoted on Twitter. I think this may be why everyone wants to make decentralization more personal and more personal. feature.

Jolestar

In fact, there has been a lot of debate on the Internet about how assets are managed. Some people think that the asset management ability of ordinary users may not be enough, self-storage may face hacker attacks, unofficial fake apps, private key mnemonic loss and other problems lead to asset loss, this situation does exist.

But in the long run, the overall bankruptcy or loss of centralized exchanges will have a greater impact on the industry than the loss of individual users. The losses of individual users are scattered, not centralized. Some people think that the management methods and technology of the exchange are safer, but because it is a single point, the risk is concentrated. Once an outbreak occurs, the impact on the industry will be far greater than that of accidents. Great impact. And from the perspective of the market, the risk loss of individual users will have less impact on the fluctuation of the overall market, so it is recommended that you manage your own assets.

Shaun

Today I saw a person talking about how he manages digital assets. He said that he would divide his digital assets into two parts, 50% in the cold wallet and 50% in the top exchange, even if he faces hacker attacks or exchanges Closed, still have half of the assets.

This incident forced everyone to think of various solutions, so many centralized exchanges began to talk about a technology, that is, the Merkle tree. They wanted to use the Merkle tree to prove that their asset reserves are transparent and transactions are transparent. So today I invite Mr. Jolestar, let's talk about what is a Merkle tree, and can we explain this technology in a relatively concise way?

Jolestar

Merkle Tree is actually a very basic technology. The problem it wants to solve is that if there are two data sets and you want to compare whether the data in the two data sets are consistent, the cost will be very high if you compare them one by one. So we can make the data into a tree structure, find two nodes, calculate the hash value of the node data (the hash value can be understood as a data fingerprint), and combine the obtained two hash values ​​to generate a hash value , build a tree layer by layer.

After the tree is built, there is a root. If the order of the leaf nodes can be guaranteed, by comparing the root nodes of the two data sets, you can know whether the entire data is consistent. This technology will be used in many scenarios, and it will be used to make a state tree in the blockchain. I understand it as a state proof or data proof.

image description

Merkle Tree icon

Shaun

So the way of Merkle tree can keep the information between the exchange and users in a more transparent state, can it be understood in this way?

Jolestar

It is a bit more transparent than a completely black box state. But for the Merkle tree to be effective, it needs to rely on several conditions:

One is that the exchange publishes the root of the data, and each user checks whether the balance on the root matches the balance on the account. If a mismatch is found, the user will announce it to the public and denounce it on social media. This is essentially a kind of Binding based on social consensus, but not strong binding.

The second is that the leaf nodes of the Merkle tree will not be fully disclosed. It is necessary to rely on a third-party audit agency to verify the total value of all nodes to confirm whether the total value is consistent with the value of all individual users. Otherwise, the value of individual nodes will be generated at will. Even negative, total values ​​can match. Therefore, each node must be confirmed by an audit agency to ensure that the value is credible.

In addition, the data of the Merkle tree is time-sensitive, and it relies on manual auditing, but manual auditing cannot be audited in real time, nor frequently, or even daily. So it can only be audited in stages, and this may lead to the fact that the data is correct during the audit, and the money is misappropriated when the audit is not performed, and it is withdrawn when the audit is performed again. This kind of thing is unavoidable.

Shaun

That is to say, this matter is not only about the technology itself. The Merkle tree is publicized only from the technical level, but it also requires the self-inspection and self-discipline of industry players, as well as the assistance of third-party regulatory agencies. If one or two parties are relatively evil, it is still difficult for the technology itself to meet the requirements of transparency.

So after this incident came out, major exchanges announced Merkle trees one after another, and some began to post snapshots and take some actions. To what extent do you think these actions are effective?

I understand that it actually promotes the industry to a certain extent, but it is not that useful?

Jolestar

Not as ideal, but still works. There are currently many centralized exchanges in the blockchain. If it really intends to run away, users basically have nothing to do. The reason why it did not run away is because it follows the constraints of a certain degree of "contractual spirit" between merchants and users, or between people.

Constraints come from the punishment for breaking the promise. After the other party breaks the promise, I can choose not to continue trading with you. This is the most direct punishment. If the exchange hopes for a longer-term game relationship, this most direct punishment will have a big impact. Therefore, the exchange will tend to continue to trade with you for a long time, because there will be a bigger cake in the future. If you can earn more money in the future, the exchange has no reason to run away. If the industry continues to develop upwards and everyone has high expectations for the future, the user's exit deterrence will be effective.

At the same time, although the centralized exchange is opaque, the transactions on the blockchain are transparent, and the inflow and outflow of funds is transparent. This is a very clear binding force on the exchange. Whether the exchange wallet can withdraw coins normally and smoothly, and the records of inflow and outflow are all public. I think this is also the reason why blockchain exchanges have been growing in a positive feedback loop for so many years. This is not like traditional financial institutions, such as the P2P boom, you will find that the entire industry has almost no positive growth process, because the transfer of funds from fundraising to use is completely opaque.

Therefore, although the exchange is centralized, it still has a transparent component, and the Merkle tree can make it more transparent. In addition, although there are several shortcomings mentioned above, there is a cost to collude with the audit institution, and there is also a cost to temporarily lending funds to fill the hole, and it is easy to be discovered. Recently there is also news that some exchanges are borrowing funds.

Shaun

Yes, the latest news should be three hours ago. Twitter circulated that an exchange embezzled funds to take snapshots. If you are interested, you can follow up on Twitter. The information is updated quite quickly.

As mentioned earlier, the cost of doing evil in this way will be higher now, so what do you think of the announcement of the Merkle tree by the exchange, which is actually good, but the efficiency is limited?

Jolestar

Yeah, so I think a little is better than nothing, something is better than nothing. Although it does not solve the fundamental problem, nor is it the ultimate solution, but at this stage, it has aroused everyone's attention. For example, some exchanges have published the Merkle tree before. If an exchange with 100% margin sincerely wants to prove its innocence, but users don’t care, it will also make the exchange very distressed, because users don’t pay attention. There is no competitive advantage over other exchanges.

Therefore, the widespread attention this time can be regarded as forcing exchanges to self-discipline. If an honest exchange has a competitive advantage, everyone will be more honest. If you can't admit your innocence, it is a process from bad money to good money.

Shaun

You also mentioned that the Merkle tree is not the ultimate solution. I believe that the entire industry is still in a stage of continuous exploration and development. From your point of view, besides the Merkle tree, are there any other solutions worth exploring at present?

Jolestar

The system process of the exchange can be understood as, based on the user's request, use the program to process the user's funds and data to produce the final result. In short, I perform calculations based on parameters, I run the program, I get the results, and I tell you the results. But how can you tell if my results are real and not deceiving you? The easiest way is that you also perform the calculation through the program. If a third party wants to verify the authenticity of the result, it still needs to perform the calculation again by itself.

If some users are unable to run the program verification due to hardware and other reasons, and want them to believe the authenticity of the value, it is necessary for each operating node to discuss a unified rule, how to judge the result is correct, this is the consensus of the blockchain, so The mode of operation is also the mechanism of the chain.

If the data is not only calculated based on parameters, but a state will be left after each calculation, and it will be accumulated for rereading in the next calculation. This involves a problem, how to ensure that the final calculation-based state on each node is consistent. This is also the usage of the Merkle tree. The hash value in the block forms a state tree. The block chain executes the program based on a certain Root (the root of the state tree), gets the result, and broadcasts the result to other nodes. Everyone verifies and passes The consensus mechanism decides a result that everyone agrees with.

Every transaction will lead to a change in the root of the state tree, and the audit is a phased audit, so how to ensure that each step in the process is credible? At the same time, if the user wants to test the results, he must perform the calculation himself, which also means that the program and data of the exchange must be made public. However, in fact, it is difficult to directly prove that a certain transaction was initiated by the user with the transaction data of the exchange. A user’s identity sent a transaction, the user’s balance changed, and the user appealed, but this record does exist in the transaction record. If the information is released to the public, it will be difficult for outsiders to judge whether the behavior was sent by the user or the exchange sent it on behalf of the user.

At this time, the blockchain signature mechanism is needed, that is, there must be a private key that the user knows but the exchange does not know. During the transaction, the user signs the transaction and sends the transaction. At this time, the exchange actually becomes a chain. When operating in the blockchain mode, the user must use the private key to initiate the transaction. All transactions and programs must be made public, and third parties can also verify Verify the data authenticity of the exchange. At this time, it is more like a private chain, but the private chain with public data, because there is no other node to make a consensus, there is no constraint, even if the exchange miscalculates the data, the user finds the error after proof, and uses the social consensus that tweets to condemn the transaction to constrain the transaction So it won't work very well.

If you want to use procedural constraints, the exchange needs to mortgage a part of the funds to the chain. If the user finds an execution error, the user will go to the chain to submit for arbitration and deduct the funds mortgaged by the exchange. If the exchange is determined to do so, users can be trusted.

At the same time, this is also the goal of Layer 2 at this stage, but the difficult problem is how to prove the on-chain arbitration link of exchange transactions. If the calculation of the matching program run by the exchange is wrong, how can the chain recognize the matching error? If the exchange can repeatedly execute the program on the chain, it can be punished, but this solution is still not feasible, because the complexity of most off-chain programs is greater than that on the chain, and the chain cannot support such complex repeated calculations.

Other methods include ZK (Zero-Knowledge Proof), that is, to generate a proof from the calculation results off the chain, which is only verified on the chain and does not need to be re-executed. However, it still requires a certain degree of complexity to make a matching program for a general exchange.

Another solution is the interactive proof of Layer 2, which is similar to the mechanism of repeated proof. The user runs the program of the exchange and finds a calculation error, but this transaction cannot be executed repeatedly on the chain, so it is necessary to find out which long instruction has a difference through the intersection of off-chain, and generate the instruction and the memory state involved. The state tree is submitted to the chain, and the instructions are re-executed on the chain to judge whether it is right or wrong. Through this idea, any complex off-chain program can be arbitrated on the chain.

So in the long run, centralized exchanges will either evolve into decentralized exchanges themselves, or wait to be surpassed by other decentralized exchanges. At this time, only OTC is left in the application scenario of the centralized exchange, which is the fiat currency deposit and withdrawal channel and conversion channel.

Shaun

That is to say, from your point of view, if the industry is developing positively, the model of future exchanges should be decentralized. Does Binance belong to a purely centralized exchange, and how to define it?

Jolestar

Yes. In fact, there are still many problems here. For example, exchanges cannot directly disclose user transaction data, because they have a large amount of user privacy information. I think it is very difficult for a centralized exchange to evolve into a decentralized exchange now, because I have not found any successful precedents of a large company that can create a new trend internally and revolutionize itself. So it must be a new decentralized exchange model, gradually attracting users from centralized exchanges and gradually replacing them.

Shaun

Is it possible for the future development to be centralized exchanges plus strong supervision?

Jolestar

Impossible, even if the so-called strong supervision of traditional finance is not what we are after, there will still be a certain degree of "delay". Under normal circumstances, regulators need to obtain records and books before they can audit. The so-called Layer 2 now needs to ensure the availability of data, that is, to ensure that anyone can get it and audit it. In the traditional situation, it is impossible for everyone to get the transaction records. Only audit institutions can get it. Its data availability is controlled by It is limited to a certain range, and it is not real-time. After all, there have been news of scandals in audit institutions themselves. Therefore, we believe that if there are technical solutions, we should not use human methods.

In addition, the asset types in the blockchain field are evolving very fast, while the traditional regulatory system has evolved for a very long time. For example, the regulatory system of banks has evolved for hundreds of years before a relatively complete regulatory system emerged. The same is true for stock exchanges, that is, the entire link from asset issuance to circulation is within a controllable range. However, the crypto field is only subject to strong supervision at one end. For example, in the current Coinbase model, the circulation of assets is not internal, but external, which will bring a certain degree of challenge.

Shaun

I saw that some community partners said in the comment area that it is best to build a public chain that only trades currency pairs. In fact, the solutions we just talked about are basically based on Layer 2 (two layers)?

Jolestar

There are several problems involved in this. It is not cost-effective to use the public chain as an exchange itself. The public chain itself solves a problem that is the arbitrator problem. Everyone operates based on the consensus rules of the nodes. There is no need to return to the arbitrator mode of the centralized exchange, and there is no need to create another chain and use other mechanisms to solve this problem. In addition, the more nodes in the public chain, the higher the cost, and the TPS will also be affected.

Using Layer 2 to solve is a more reliable solution, and the advantages of centralized exchanges can be satisfied by Layer 2. However, Layer 2 has some difficulties at this stage, that is, most Layer 2s are single-chain at this stage, and the chains supported by centralized exchanges are sufficient. It is obvious that centralized exchanges are faster when doing some complex exchanges.

Therefore, the industry is also thinking about new solutions recently, that is, whether Layer 2 can connect to multiple chains, and expand the connection to multiple chains as the settlement layer, and finally achieve the experience of a decentralized exchange. The type of asset support can completely match that of It is comparable to centralized exchanges and has higher security.

Shaun

Talking about this, the topic is coming to an end. Given that this event is still fermenting, what substantive impact do you think it will have on the future of the entire industry?

Jolestar

In the past, there was an ambiguity between decentralization and centralization. Although "I" myself was centralized, in the decentralized world, nothing could control "I". But now it is accelerating this state, that is, either it is truly decentralized, or it returns to the regulatory framework and removes the vague area in the middle.

I mentioned in a long tweet about Web 3 that I wrote recently that what Web 3 needs to solve is the problem that Web 2 applications are torn apart by regulators, and centralized exchanges will directly face this problem. If you return to the regulatory framework, you can only operate in one country, which will reject users from other countries; or in different countries, split into different exchanges, abide by the regulatory rules of different countries, and separate users. In the short term, the trading volume of a single country may be able to support an exchange in the short term, but it is not a future trend in the long run.

If you don't want to be torn apart by regulators, you can only take the decentralized route, and the current solution seems to me to be only Layer 2. Of course, there may be a route along the path, such as building several exchanges into alliance chains, running each other's nodes, connecting accounts, and supervising each other. But I don't think so far, they will reach such a consensus.

Shaun

That's right, it's still in a stage of killing each other but not loving each other hahaha.

Thanks to Mr. Jolestar, we reserve a few minutes. Does anyone have any questions about it?

Sophie.Fu

Hello, first of all, I am very grateful to the two teachers for their sharing. I would like to ask a relatively novice question, that is, are there no purely decentralized exchanges in the current exchanges on the market? Like Binance or the exchanges involved in this incident, they are actually relatively centralized exchanges, can this be understood?

Jolestar

Now all Swap on the chain can be understood as a kind of decentralized exchange, but the transaction models will be different. Centralized exchanges generally refer to pending orders, and Swap is an AMM (Automatic Market Maker) method of direct exchange. It is different from pending order matching, but essentially the function is the same, that is, exchanging one Token for another.

For the judgment of whether it is decentralized, if you do not understand technology, there are several simplest judgment criteria. First of all, whether the program is logged in with a wallet or an account, although there are also decentralized wallets that do not need a private key to log in, but you can feel who is under the control of the assets, which can be distinguished.

The second is whether the transaction can be closed directly. This involves the point that was missed earlier, which is the so-called unilateral withdrawal ability. Among all the current Layer 2 solutions, the most complicated design is to allow users to unilaterally withdraw money from the chain, that is, if the application is down, can the user withdraw the money as usual through some chain method? The key difference between centralization and decentralization. Because even if the centralized exchange proves that the assets are healthy through the Merkle tree, if it closes the website or application and ignores the users, the users will have nothing to do, and the assets will not be withdrawn.

Another thing is to look at the team. Even if the boss runs away or disappears, the application will continue to run, and no one can stop it. This is truly decentralized.

Shaun

It seems that everyone should pay attention to whether to log in with a wallet or an exchange account and password in the future. This is a good point of distinction.

Jolestar

Let me add a few words about the implications of this event for the future. If the centralized exchange further "rolls" on the route of self-certification and innocence, it will actually promote the evolution of the overall Layer 2 technology, and the route of technical proof such as blockchain state proof and calculation proof will receive more attention , Layer 2 technology will become prominent in the future, and everyone will really pay attention to how to prove it.

Shaun

As you said, the previous exchanges seemed less tense, and the current trend is that everyone is busy trying to prove their innocence. But the current solution still relies more on the "old tune" Merkle tree, which can solve a certain degree of problems, but it may only solve less than 1/3 of the problems, because it also includes the self-discipline of industry players , and third-party audit supervision.

As you mentioned in your sharing, in the future, the entire Layer 2 may have more new technical empowerments and wider application scenarios.

I think there is no winner in this "fairy fight". I don't use much in the exchange, but I think the original FTX is excellent in terms of liquidity and experience. But as mentioned at the beginning, FTX collapsed in just two or three days. This is related to the overall panic in the industry and the run-off of users, but the speed is beyond people's expectations.

SBF himself also talked about the mistakes he made and the lack of dealing with the liquidity run. CZ also thinks that this matter is actually not good for anyone. You also mentioned that it is almost more difficult for an exchange to obtain a license for globalization in the future, and you can only deal with policies on a region-by-region basis.

The most important thing is the capital market. After this incident, everyone will become more cautious, which is not necessarily a good thing. Big capital like Sequoia and Softbank also lost a lot, but there is no way. The real builders of the industry still need funding to support them. I think this may not be so optimistic.

I personally feel that this incident has a profound impact on the future development of the entire circle. On the bright side, it is a good phenomenon that the event forced the industry, especially the centralized exchanges, to make more decentralized improvements. It can bring back many old technologies that are in the right place at the moment, such as Merkel The tree can also force developers to speed up the possibility of exploring related new technologies. But from a negative point of view, the accelerated coming, possibly more stringent industry regulation, the status quo of a dominant centralized exchange, and the panic in the capital market caused by the Lehman effect may affect the future development of the entire industry. The road becomes more difficult and uncertain.

But no matter how bloody the capital industry is, the real BUIDLers are always on the road.

Thanks again for Mr. Jolestar’s passionate sharing, welcome to come and sit down from time to time, and thanks to the friends in the DAOrayaki community.

The audio of today's related discussions will be uploaded later on DAOrayaki's official website, DAOrayaki.org's dedicated podcast channel. We will also release it on the small universe podcast platform in China, Himalaya FM, and major streaming media platforms around the world, such as Apple Podcasts, Google Podcasts, YouTube and Spotify. You are welcome to listen and subscribe. See you next time.