Revisiting Ethereum: What are the bullish reasons?

CN
2 days ago

Is this a turning point for Ethereum to revive its price?

Host: Alex, Research Partner at Mint Ventures

Guests: Zhou Qi, Founder of EthStorage; Lawrence, Researcher at Mint Ventures

Hello everyone, welcome to WEB3 Mint To Be initiated by Mint Ventures. Here, we clarify facts, explore realities, and seek consensus in the WEB3 world through continuous questioning and deep thinking. We aim to elucidate the logic behind hot topics, provide insights that penetrate the events themselves, and introduce diverse perspectives.

Alex: In this episode, we have invited Dr. Zhou from EthStorage and our researcher Lawrence. We are here to discuss a blue-chip asset that crypto investors are very concerned about—Ethereum. We know that Ethereum's overall performance in this cycle has not been good; its exchange rate has consistently lagged behind BTC and has often been outperformed by its competitor Solana. However, there have been many noteworthy changes in Ethereum recently, such as Vitalik becoming very determined about Layer 1 scaling, restructuring, and layoffs, which overall reflect a more pragmatic approach. Is this a turning point for Ethereum to revive its price? Before we officially start our discussion today, let's have our two guests introduce themselves to our audience. Dr. Zhou, please go ahead.

Zhou Qi: Hello everyone, I am Zhou Qi, the founder of EthStorage. I am very happy to share some insights about Ethereum with you today, including some of our recent work on Ethereum. We have been deeply researching Ethereum's technology since 2017 and 2018, including its scaling roadmap, from the previous Layer 2 to the current Layer 1. We have also participated in many research projects related to Ethereum and received substantial support from Ethereum, including research on DA and some related studies with OP Stack, as well as many grants. So today, we are very pleased to share some of our relatively unique insights in this area.

Lawrence: Hello everyone, I am Lawrence from Mint Ventures. I am glad to be here with Dr. Zhou to discuss this issue.

Reasons for Ethereum's Underperformance Compared to BTC and SOL

Alex: Let's get into today's main topic. Before discussing the bullish reasons to look forward to Ethereum, let's first outline the issues Ethereum is currently facing. In your opinion, what are the main reasons for Ethereum's significant underperformance compared to BTC and Solana in this cycle? Let's have Dr. Zhou share his thoughts first.

Zhou Qi: I think there are several reasons. The first reason is the entire Ethereum roadmap, especially the Layer 2-centric roadmap, which many have found to be somewhat inconsistent with Ethereum's core value. This was actually a topic I discussed with Vitalik when I met him last month in East Asia, and it is also his perspective. For example, last year, before EIP 4844 went live, Ethereum was still in a relatively deflationary state. However, after the 4844 upgrade, the fees for Layer 2 submitting data to Ethereum dropped significantly, which also led to the value of many Layer 2 projects not being reflected in Ethereum itself. For instance, many Layer 2 projects like Base and Arbitrum received a lot of user fees, but these fees did not benefit Ethereum's value itself. This has created a significant incentive misalignment issue. The second reason is that Ethereum's performance in the previous cycle was very good, which in some sense affected its judgment, causing it to become relatively slow in certain areas, such as engineering progress. At that time, Ethereum did not have any real challengers, whether it was Bitcoin or Solana. In this wave, many people have complained that Ethereum is changing too slowly, that its roadmap has been planned for a long time without actual engineering results. Each upgrade takes one to two years to implement the corresponding features. Compared to Solana's aggressive engineering advancement, Ethereum has always leaned towards research, with engineering not being a top priority. This approach has caused significant delays in Ethereum's overall development and upgrade roadmap. We have many personal experiences in this regard; for example, in previous years, we contributed to many EIPs for Ethereum, including EthStorage itself, which we can elaborate on later. In summary, these are the two main aspects.

Lawrence: The two points that Dr. Zhou just mentioned are also what I wanted to highlight. Another important reason, I believe, is that in this cycle, there have been relatively few new business models or innovations on-chain. The richness and activity of on-chain businesses have not improved significantly compared to 2021; in fact, if we exclude meme trading, the activity and richness of on-chain businesses may have slightly declined. In contrast, the fundamentals of BTC have improved significantly. Therefore, in this context, the overall performance of all public chains has been poor. Even Solana, which has performed relatively well, is still at least 50% below its exchange rate compared to BTC's peak in 2021. This is a common issue faced by all Layer 1s in this cycle. Additionally, the two points that Dr. Zhou mentioned are also relevant. One can say that Ethereum's strategic issue in the last two to three years, namely the Layer 2 strategy, can be frankly judged as a failure. Another point I summarize is that Ethereum has a long-standing structural issue, which is not just a short- to medium-term strategic problem. I have seen many criticisms recently, especially from someone named Max Resnick, which I find very typical. Max was previously a researcher at the Ethereum Foundation and switched to Solana at the end of 2024, joining Anza, the team that was spun off from Solana Labs for R&D. He has been critical of Ethereum's Rollup strategy and has supported scaling Layer 1 instead. He has made some sharp criticisms of Ethereum, such as believing that those who set Ethereum's roadmap, specifically Vitalik, have expertise in blockchain and cryptography research but lack sufficient research in computer science. This has led to many fundamental and factual deviations in Ethereum's judgment and research on how to improve blockchain performance over a long period, including directional deviations. For example, the Ethereum Foundation has long believed that the bottleneck for improving Ethereum's performance lies in the execution layer. However, according to him, the bottleneck for performance improvement is clearly in the consensus layer. He also mentioned that Vitalik or those who set Ethereum's roadmap have pursued long-term goals that seem relatively abstract, rather than focusing on current users. For instance, Vitalik has consistently discussed privacy applications and social applications on his blog but has spoken less about DeFi, even though DeFi has always been the most used application on the Ethereum mainnet. This issue sparked some debate around July and August of last year, where Vitalik and several leading DeFi projects on Ethereum discussed this matter. The low efficiency of Ethereum's R&D, which Dr. Zhou just mentioned, is also a result of this. On one hand, there is a significant disconnect between the research team and the development team, with a large gap between what the research team studies and what the development team develops. On the other hand, the R&D progress is indeed very slow, with basically only one major upgrade per year. For example, the transition to POS was something Vitalik started discussing in 2015 or 2016, but it wasn't until the Shanghai upgrade in 2023 that it was officially closed. Since 2021, there have not been many upgrades that people can name: the Merge in 2022, the Shanghai upgrade in 2023, last year's Cancun upgrade, and the recently completed Pectra upgrade. Overall, the development progress is very slow, leading to a very high cost of error correction, especially in terms of time. For example, the Layer 2 strategic issue in Ethereum's case is that in 2020, Vitalik proposed a Rollup-centric approach, but it wasn't until 2022 that related elements were implemented, and now in 2025, everyone realizes it is not working and needs to change. There has been a significant waste of time in between. Especially compared to Solana and some new Layer 1s like Sui, the gap in their R&D progress is approaching an order of magnitude in efficiency. This means that Ethereum may take ten times longer than other public chains to make a decision and push something online. Of course, I think there are reasons for this, as Ethereum is the first influential public chain after Bitcoin, facing many issues, including previous regulatory considerations, and it has always maintained a strong commitment to decentralization. However, in terms of results, I believe the issues mentioned earlier are long-term structural problems belonging to the Ethereum Foundation or the core team that sets Ethereum's roadmap. I feel that these are the main three points.

Alex: OK, I would like to add one more point. Lawrence just mentioned that the top-level concept of the Ethereum Foundation's commitment to decentralization is very persistent. Until the last cycle, everyone still believed that this was one of the elements of blockchain orthodoxy. However, there have been many changes in this cycle. One of the most important changes is that there has been a significant shift in the U.S. government. The current U.S. government is very friendly towards crypto, and regulation is quite loose. This has resulted in a reduced urgency for the crackdown and confrontation against crypto projects during this governing period. Ethereum's strong commitment to decentralization has become less necessary in this government cycle. In contrast, projects like Solana and Sui, which may not be highly decentralized but have excellent efficiency and performance, have become an advantage. Moreover, in the long term, I feel that even if the next government changes to a Democratic administration, they will recognize that the crypto investor base in the U.S. is very important. Under this premise, I believe they will not conduct such a brutal crackdown on crypto projects as during the previous administration under Gary Gensler. Therefore, the necessity of decentralization is gradually decreasing with the changes in the industry. We see that many projects that have emerged in this cycle, such as Ethena, and many narratives that investors now consider important, like RWA, are actually products of the combination of CeFi and DeFi, which is an inevitable trend. The decreasing importance of this narrative has also somewhat weakened Ethereum's consensus, which is one of the reasons why it is not performing as well as SOL in terms of narrative this cycle.

Consensus and Non-Consensus on Ethereum Issues

Alex: Let's move on to the next question. We have discussed many issues regarding Ethereum, including its engineering capabilities, understanding of development direction, and slow error correction speed, among others. Given all these issues, what are the points of consensus among Ethereum's leaders, community, and developers? What are the points of non-consensus? To put it simply, which issues are commonly recognized as problems from the core management of Ethereum to the community and developers? Which issues have differing opinions? For example, we might see something as a problem or an obstacle, but Ethereum's current stance might be that it is not an issue, which is a characteristic we are very concerned about. What are your thoughts on this? Let's have Dr. Zhou share first.

Zhou Qi: I think a very important point in this wave is that there has been a significant change in Ethereum's definition of decentralization. I can say that a few years ago, Ethereum had a very idealistic, almost religious fervor in its extreme pursuit of decentralization. I remember chatting with some people from Ethereum back then, and they expressed a desire for Ethereum L1 to become a minimal trust layer, allowing devices like smartphones or even very simple embedded devices to run an Ethereum validator. However, it is clear that after challenges from Solana and others, especially in the upgrade roadmap for L1 scaling, including continuously increasing the gas limit and introducing block-level access lists to accelerate transaction speeds at the execution layer, they are actually seeking a more pragmatic balance between decentralization and execution efficiency. This also means we might need more powerful computers. The direct issue everyone faces is: if you design an Ethereum consensus that allows a smartphone or a device worth 100 yuan to run a validator, you still need at least 32 ETH. At the current price, that’s nearly 100,000 yuan. This is actually mismatched. Your device is not the bottleneck for the validator; the main issue is that you need to hold too much ETH. So under this premise, why not relax this assumption? For example, we could allow computers worth $1,000, $2,000, or $3,000 to run nodes while achieving 2x, 3x, or even 10x throughput scaling for Ethereum L1. This is part of Ethereum's upcoming plan. This represents a pragmatic adjustment in Ethereum's search for a balance between decentralization and execution efficiency. For instance, we submitted a proposal to Ethereum two years ago, the ESP grant. We wanted to research block-level access lists. The idea is that when packaging a block, I can inform other validators which data I will access while executing these transactions. This way, they can use pre-fetching techniques to read data like account balances in high concurrency ahead of time, greatly improving execution efficiency. Two years ago, we only applied for 10,000 yuan in funding to study this issue, believing it would be valuable and helpful for Ethereum, but it was rejected without explanation. We suspect it was because Ethereum thought this might impact decentralization and was not a top priority. However, at the beginning of this year, they suddenly announced they would conduct research in this area and welcomed our participation. This indicates that Ethereum's paradigm has shifted from an idealistic view of decentralization two years ago to a more practical perspective now. Especially with the improvement in computer performance and the decrease in costs, it roughly aligns with Moore's Law. If Ethereum continues to follow the original design framework without dynamically adjusting the gas limit or adding block-level access lists, then execution efficiency cannot be unleashed, making it difficult to achieve, for example, tenfold scaling. I was quite surprised; two or three years ago, I wondered why they didn’t consider this issue. I can only understand it as their adherence to idealism.

As for the non-consensus part, I believe Ethereum still carries a lot of "debt"—including technical debt, cognitive debt, and even brand debt. This means that when it comes to making significant reforms, it often cannot directly negate its past. Historically, we have seen similar situations, such as during China's reform and opening up and certain leadership transitions in the Soviet Union, where slow but steady reform proved more effective than completely overturning the past. For example, Ethereum now adopts a multi-client roadmap. There are four or five different clients written in various languages at both the execution and consensus layers. In contrast, we have seen that particularly successful infrastructure software, such as Solana itself, the Linux operating system, and HDFS, are generally implemented using a single language and a single testing framework. They pursue engineering efficiency. Ethereum, having experienced the DAO attack incident, anticipates that using a single language might introduce mechanisms due to vulnerabilities in that language, leading to downtime. Since Ethereum is still changing very rapidly, unlike Bitcoin, where many functions have become relatively fixed, they would rather spend ten times the engineering effort to prevent such risks. However, by doing so, Ethereum requires five to ten times the engineering power to keep pace with Solana. This is also why Ethereum's upgrades are so slow. We have personally participated in Ethereum's upgrade process, including EIP formulation, DevNet, and TestNet, and have found that coordinating implementations between different clients requires a significant amount of effort. For example, during the recent Pectra upgrade, we discovered that Geth was inconsistent with other clients' configurations, leading to synchronization issues that needed immediate fixing, ultimately finding the problem lay with Geth. This is a trade-off issue that requires substantial engineering effort. Here, I ponder a question: does rapidly changing software like Ethereum necessarily have to avoid downtime? Solana has experienced multiple downtimes, and other projects have as well. Perhaps we can allow for downtime but ensure a quick recovery. Maintaining a single client could enable faster iteration. This is actually a software engineering issue. I have seen that Ethereum is recently recruiting positions like Chief Performance Developer, focusing more on solving progress issues efficiently from an engineering perspective. Having worked in large companies, I know that downtime is actually the norm. Even with money and engineers, problems frequently arise. During every training session, we have to discuss how many times our systems have gone down; for instance, systems at Meta or Google have been completely inaccessible for an hour, which is not uncommon. Therefore, how to overcome these issues is a very worthwhile topic for in-depth discussion and is currently still an area of non-consensus.

Lawrence: I understand that the current consensus is that the previous focus on Layer 2 strategies needs to shift. I have seen that the Ethereum Foundation has indeed taken a series of actions recently, and the community has long been skeptical about this matter. Of course, there are still some uncertain aspects, such as the wording used by the Ethereum Foundation regarding this issue; they refer to it as "reprioritization," which sounds more moderate. However, most people feel this should be termed a "pivot," indicating a need to acknowledge past mistakes. Nevertheless, based on the current actions, I believe that focusing on Layer 1 and abandoning the previous Rollup Centric strategy is the most widely agreed-upon point.

As for the points of non-consensus, as Dr. Zhou just mentioned, I believe it still lies in the Ethereum Foundation's commitment to decentralization and where the limits of that commitment are. Currently, it feels like there is still no consensus on this. In fact, many of Ethereum's efficiency issues can be traced back to their adherence to decentralization principles. For example, whether at the execution layer or the consensus layer, they require decentralization and support for multiple clients. At the same time, they have long insisted on supporting solo stakers, making the participation rate of individual stakers a significant goal. This also affects efficiency because it requires coordinating a large number of clients. I have been following Lido for some time, and I feel that the Ethereum Foundation has not shown a clear stance on this issue, such as no longer prioritizing solo stakers and feeling that some centralization is acceptable. I sense that Ethereum, at its inception, could only be compared to Bitcoin, which in its early years existed as a symbol of resistance against the global financial system. Therefore, Ethereum has emphasized anti-regulation and anti-censorship from the very beginning. Unlike BTC, which has no subject, Ethereum has a subject and is very cautious about being regulated. In recent years, Solana has experienced some of the issues that Ethereum initially worried about. For instance, Solana was explicitly defined as a security by the SEC in 2023. The Ethereum Foundation was very concerned about this in the past, but now it seems that even if it is defined as a security, it does not pose a significant problem, and Solana continues to develop well. To what extent can there be a retreat or concession on decentralization? I have not seen much discussion on this. I believe this is a key issue that will continue to affect Ethereum's efficiency and other aspects in the medium to long term, and consensus has yet to be reached.

Timeline of Ethereum Reform Events

Alex: Ok. We have discussed many issues and touched on some points of consensus regarding these issues from the community to the top of the Ethereum Foundation. What strategies has Ethereum's leadership planned in response to these consensus issues? What is the general timeline for these strategies? For example, there are two recent events that have garnered a lot of attention: the first is that Vitalik mentioned he hopes to achieve a tenfold expansion of Ethereum by the end of this year; the second is the recent restructuring and layoffs at the Ethereum Foundation. What other key events do you think are similar to these? Is there a common expectation for their occurrence? Let's have Dr. Zhou address this question first.

Qi Zhou: First, the first point is about Ethereum's L1 scaling. We can see that Ethereum has launched a very clear roadmap on how to increase the current gas limit from about 30 million to, say, 60 million, and gradually raise it. They also have corresponding EIPs to make related improvements. Secondly, we also see that Ethereum's client, such as Geth, is actively optimizing the current code. For example, recently we observed a very interesting phenomenon where all Ethereum clients have implemented some form of performance caching, specifically data caching. We found that Geth's caching implementation has not been particularly effective over the past four to five years, which surprised us, but we did not analyze it deeply. This has led to the current transaction execution speed being about 100 million gas per second. If the block's gas limit is adjusted to 100 million or even 200 million, for example, ten times would be 300 million, then the time required to process the execution of transaction blocks could take about 3 seconds, which might lead to block timeouts. Because Ethereum has strict time limits for block production, voting, execution, and other aspects within a 12-second window, it is very likely to timeout. Interestingly, Nethermind implemented some caching optimization techniques during their engineering process, miraculously increasing performance by about 3 to 4 times, reaching levels of 400 to 500 million gas per second, which is data we have already measured and can reproduce. We saw that Geth released its latest improvements at the beginning of this month or the end of last month, and we replicated it on our machines, achieving the same performance standards without modifying any specs of the current consensus and execution layers. It can be seen that Geth may not have spent much time optimizing in this area over the past four to five years, but suddenly under pressure, seeing others perform better, they began to optimize significantly, resulting in a performance increase of 4 to 5 times. Initially, when we saw Nethermind's 4 to 5 times performance improvement, it was for all Ethereum clients, and we thought they might have used some black magic or that the data was inaccurate. Later, we realized that everyone had not been pushing themselves hard enough and were still in their comfort zones. Now we know that performance can actually run faster, which is a kind of progress that occurs under pressure. This also means that we have validated through data that Ethereum can potentially increase performance by 3 to 4 times, or even 10 times. Originally, everyone thought that increasing from 100 million gas per second by 10 times would lead to timeouts, but now it seems that this is no longer a major issue. This is a very detailed perspective and indicates that this kind of pressure is actually a very good progress for Ethereum. Although Solana did not perform well against Ethereum this time, the ultimate result may be that Ethereum continuously improves due to competition, which is a good advancement for the entire industry.

Another story regarding layoffs can also be shared. We previously collaborated with Ethereum's official project, Portal Network, which mainly addresses data storage issues after Ethereum's scaling. The historical transaction data of Ethereum is about 300G to 400G, and the state data is also 300G to 400G, totaling about 1T. If we scale by 10 times, the data will quickly exceed 2T or 3T. The ultimate goal is to store the data without affecting decentralization while reducing the overhead for validating nodes and full nodes, thus achieving a better decentralized scaling method. We worked closely with them for about a year and a half to two years, and then one Monday, they suddenly announced that the project was canceled, and all full-time members were laid off. So we can see that there must have been a very strong determination within Ethereum. Any solutions that do not directly address scaling are not their top priority at the moment. Ethereum's determination in this regard is immense.

Lawrence: The timeline currently only shows very rough information, such as the previously mentioned goal of achieving a 10-fold L1 scaling within a year, and a target of 100-fold scaling in 2 or 4 years. Dr. Zhou's comments also indirectly highlight past issues with Ethereum: for example, there are optimizations available that exceed 5 times at the execution layer, but they have not been pursued in recent years. The countermeasures include a strategic adjustment to bring L1 back to the core strategy; on the other hand, there are organizational changes, including the addition of two new executive directors, one being Wang Xiaowei and the other being Tomasz Kajetan Stanczak, the founder of Nethermind. There have also been intensive layoffs in recent days. This morning, the Ethereum Foundation released a general letter regarding the budget, and there was also a reorganization and personnel allocation within the foundation recently. A more detailed timeline may come after more specific actions. But for now, the terms of these two new executive directors are two years, starting this year, and I believe they will provide good assistance for Ethereum's scaling goals, especially in improving L1 performance. Compared to other researchers, they have more of a research and development background. Wang Xiaowei has been researching Ethereum scaling for a long time, focusing on sharding; while Tomasz Stańczak's work on the Nethermind client has recently shocked other Ethereum clients. I believe their addition will be beneficial for improving Ethereum's performance and may lead to more intuitive enhancements.

Is Ethereum Still Worth Watching?

Alex: So, combining everything we've discussed, Ethereum has some issues but has also made many improvements, and of course, it has its advantages. Looking to the future at this stage, do you both still have confidence in Ethereum? I mean in terms of asset investment. What are your reasons for being optimistic about it? Among the reasons you are optimistic, which ones do you think the market is currently overlooking significantly? Let's have Dr. Zhou answer first.

Qi Zhou: My view on Ethereum is cautiously optimistic. From an optimistic perspective, I think Ethereum is a rare decentralized ecosystem with many developer communities participating and contributing. In a certain sense, it is also a very commendable achievement in our entire web3 industry. Including my conversations with many studios that are not very familiar with crypto, everyone tends to compare Ethereum with high-tech companies like Nvidia, Apple, and Tesla. In this regard, I believe Ethereum has a very good foundation and network effect. From a personal perspective, I think if our industry only had Bitcoin and none of these new interesting things, it would be too boring.

Cautiously, I feel that due to some historical reasons, Ethereum is driven by some very idealistic individuals. Initially, they received very good market feedback, but as they gradually dealt with real-world issues, some overly idealistic ideas were exposed. If the foundation cannot make significant reforms, or if those reforms are not thorough enough, I think there may be some backlash. I believe that many times it is human nature; after all, they have reached such a height and still need to maintain their original intention and continue to work hard, which is a significant challenge for them in terms of human nature. So in this regard, I still maintain a cautious attitude towards Ethereum.

Lawrence: I also remain relatively optimistic about Ethereum, though perhaps not as much as in the previous cycle. I think the reasons for optimism are, first, there are still many excellent developers continuously building within the Ethereum ecosystem, like Dr. Zhou. I believe this is a very key competitive advantage for Ethereum. Of course, some new applications may choose not to go with Ethereum and may go directly to Solana or Sui, but there is still a considerable portion of early, outstanding developers who participated in the crypto market and are still developing within the Ethereum or EVM ecosystem. I think this is an important factor for long-term optimism about Ethereum. The second point is that although many new application types have emerged first on Solana in this round, financial applications, including RWA, and new applications like Ethena and HyperLiquid, can still be seen making some new progress in the DeFi space within the Ethereum ecosystem. After further regulatory relaxation is implemented, there may be more financial applications that will still choose Ethereum when selecting chains. Another point is that I think we may soon see improvements in Ethereum's performance or reductions in costs in the short to medium term. I feel there will be a strong short-term rebound effect, where the market's short-term perception may shift.

But like Dr. Zhou, I also remain cautiously optimistic. Although Ethereum may change in the short to medium term, it is still difficult to say whether the long-term trajectory has truly changed. For example, what will happen after the two new executive directors' two-year terms end? Will Vitalik's personal characteristics change in the future? He may become more focused on existing users and short-term metrics, rather than solely on some vague long-term indicators. I think if these characteristics do not change, or if Vitalik's influence on Ethereum remains unchanged and his attitude does not shift, then in the long term, over the next five to ten years, Ethereum may still lag behind competitive chains like Solana and Sui, remaining in a generally slow and overly conservative state. In this case, the competitive advantages that Ethereum gained from being early in the cycle, whether in terms of developers or the early establishment of financial applications, will diminish with each passing cycle. I believe that if Ethereum continues to be slow and conservative in the next cycle, it may not be a very worthwhile investment target.

Under What Circumstances Would You Increase Your Holdings in Ethereum?

Alex: OK, finally, let's return to a question directly related to investment. Ethereum's overall valuation level relative to Bitcoin has been continuously declining. Now there are some positive signs of change. For both of you personally, what facts, data, or information would give you enough confidence to choose to buy or increase your holdings, or to raise Ethereum's position in your overall crypto asset allocation? What might those facts, data, or information be?

Lawrence: Alright. I think making a decision to buy ETH now, if we look at the data, the data might lag too much, and it is very likely that price increases will occur before the data improves significantly. I am looking forward to seeing Vitalik suddenly come out one day and say that we made mistakes in the past, and now we have decided to do things differently. I have a vague feeling that everyone is telling him that he is doing it wrong, and he is saying that we will try doing it your way. Vitalik's overall influence on Ethereum is still unmatched by anyone else. His influence may weaken with the addition of the two new executive directors, but for now, he remains the person with the greatest impact on Ethereum. What I hope to see is that Vitalik himself becomes more aggressive or that relatively aggressive individuals like Tomasz gain more influence within Ethereum.

Another point is that I hope Ethereum can make a mistake. I think they have been too focused on avoiding errors in the past, too intent on creating something that lasts for ten or a hundred years. They can be slow, but they must not make mistakes; their pursuit of this has been a bit excessive. Dr. Zhou just mentioned that making mistakes is actually very normal. Solana has experienced so many outages in the past, and new public chains often go down as well, but Ethereum remains overly cautious. I hope to see Ethereum become more aggressive rather than being as cautious as in the past. If I can see such signals from the decision-making level, I might consider increasing my holdings. Of course, this is mainly because I have been holding a significant position in Ethereum, so I tend to think more about the timing for potential increases.

Qi Zhou: I think from a valuation perspective, our industry's valuation is a completely new issue. So when deciding to buy Ethereum, I might consider two main factors. One is the overall market; Ethereum has already achieved such a large network effect, and as long as they are moving in a practical direction, there should be very good investment opportunities as long as one does not buy during hype. After several cycles, there are still many opportunities for Ethereum. The second factor is whether Ethereum can maintain a more practical attitude in making various decisions going forward. There are some very interesting questions here, not just about Ethereum, but about the various issues that a large organization or community will encounter. For example, many people are FUDing Ethereum this time because they feel that there are many opportunists around Vitalik who are exploiting him or overselling. However, some new members of the community feel that Vitalik is somewhat out of touch with the masses, leading to information asymmetry, with too much noise around him and no good way to identify good or outstanding individuals. I think this is an objective fact; any organization that reaches a certain scale will inevitably encounter such situations. It depends on how Vitalik can build a better team, listen to the community's voice, and face the mistakes that may have already been made. However, I still hold the view that this process needs to be smoother. If a situation like Khrushchev's were to occur, it could easily lead to a split in the community. So in this regard, Vitalik may need more introspection and reflection. An important point is how to identify and discern the voices within the community, allowing more people and organizations to more easily engage with Ethereum. Previously, I also felt that when Vitalik and his team designed Ethereum, they viewed it as a super-sovereign project, holding a distant attitude towards some ideological aspects that did not align with him. For example, this time, the U.S. under Trump has done many relatively friendly things for crypto, including banquets and various hearings, but I feel that Vitalik still maintains a certain distance from such activities. Although I see he has been very active recently, running around, when answering some sensitive questions, such as whether to set up an office in Hong Kong or attend Trump's banquet or go to U.S. hearings, I feel he still has a somewhat hesitant attitude. I'm not saying to be confrontational, but there may be some hesitation. In fact, throughout this process, the evolution of the entire Ethereum ecosystem cannot lack support from many local communities, including the general public, the upper class, and regulatory bodies. I read his blog before, and he may have had many interactions with Russia and Putin because he is originally Russian. But after the Russia-Ukraine war, he was extremely opposed to such matters. So he said he would no longer engage in sovereign exchanges. However, I think if one takes a hardline approach, that is not a particularly practical attitude. In China, we have a saying: make more friends and fewer enemies. In this way, I believe Ethereum still has a lot of room for performance in the future.

Alex: Alright. Thank you to both guests for your generous sharing on this topic today, and I hope we can invite you both to discuss more topics in the future. That concludes our program for today. Thank you.

免责声明:本文章仅代表作者个人观点,不代表本平台的立场和观点。本文章仅供信息分享,不构成对任何人的任何投资建议。用户与作者之间的任何争议,与本平台无关。如网页中刊载的文章或图片涉及侵权,请提供相关的权利证明和身份证明发送邮件到support@aicoin.com,本平台相关工作人员将会进行核查。

Bitget:注册返10%, 送$100
Ad
Share To
APP

X

Telegram

Facebook

Reddit

CopyLink