a16z Conversation with Ethereum Foundation Researcher: Where Will Ethereum Go After the Merge?

Original authors: Danny Ryan, Jeff Benson | a16z crypto

Source: Jinse Finance

The Merge is a milestone for Ethereum. What developments can we expect afterward? Will the GAS issue be resolved?

Ethereum is about to undergo its largest upgrade in history—"The Merge," transitioning from a Proof of Work (PoW) to a Proof of Stake (PoS) consensus mechanism. "The Merge" is considered the first phase of Ethereum's upgrade, aiming to enhance security and sustainability. It is important to note that this upgrade does not include the long-awaited ultimate scaling solution, sharding.

In this context, a16z conducted an exclusive interview with Ethereum Foundation researcher Danny Ryan, who is involved in this upgrade. The first part of the interview mainly discusses how "The Merge" will bring security and stability to Ethereum.

In the second part of the interview, Danny Ryan also talked about the long-awaited future upgrades, including Danksharding, Stateless Ethereum, and security updates related to Miner Extractable Value (MEV). Moreover, he explained how years of effort have led to new methods for researching and testing future upgrades.

a16z: You previously mentioned the possibility of some miners forking and continuing to use the old chain after Ethereum's "Merge." However, most people will participate in "The Merge." As a researcher at the Ethereum Foundation, what role did you play in this process? How did you coordinate such a large-scale upgrade?

DANNY RYAN: I started working on the Proof of Stake (PoS) mechanism about five years ago and knew that Ethereum would eventually take this step. The Ethereum community has always been striving to move forward, do the right thing, and build a long-lasting effective protocol, rather than just being satisfied with the present.

Thus, people sensed early on that PoS would be better and more secure than PoW, and they were very excited about it. About five years ago, when they realized that the PoS mechanism was becoming increasingly feasible, they began to feel both excited and anxious—excited because the Ethereum community's vision was about to be realized, and anxious about sensitive issues.

To smoothly promote the upgrade, the Ethereum Foundation, research teams, and clients have been working hard to explore solutions. Throughout this process, we often encounter challenges, such as whether solutions that touch on gray areas are the right choice. Should we do it now or later? Every decision we make is difficult, and the Ethereum Foundation plays a supportive coordinating role in this, helping developers collaborate, addressing review issues, facilitating dialogue, and setting timelines and priorities.

Ultimately, in most projects, the Ethereum Foundation's role is to help the protocol become more sustainable, scalable, and secure while remaining decentralized. Therefore, in terms of technical work and community coordination, our focus is mainly on promoting better information flow, research, and dialogue, enabling many members involved in development, engineering, and the community to continuously advance project development and participate in decision-making.

a16z: Over the past five years, the Ethereum community has grown significantly, and theoretically, it will become even more decentralized after "The Merge." What are your thoughts on the future upgrade process? Is it possible to coordinate upgrades through some form of Layer 1 DAO?

DANNY RYAN: As far as I know, the Ethereum community will not conduct on-chain voting or any form of token voting for upgrades; instead, users will decide what kind of protocol to run.

Generally speaking, these protocols have broad consensus, but sometimes disagreements arise, such as between Ethereum and Ethereum Classic. Ultimately, the right to run a particular protocol belongs to you, the community, and all users. Generally, we do not interfere because everyone is working hard to make Ethereum better, and there are not many conflicts regarding core content.

So, I do not expect formal technical mechanisms to constrain free development; rather, I hope the entire process can continue to evolve and change within this loose governance. I believe that governance involving researchers, developers, and community members is also a good choice.

Indeed, you have mentioned that the Ethereum community's team is growing, and decision-making and implementation are becoming increasingly difficult. However, I personally think this is a characteristic of the current stage.

From the perspective of application and user reliability, many Ethereum protocols may need to stabilize rather than constantly change. Therefore, governance and decision implementation are becoming increasingly challenging. Sometimes I feel like I am running with a weighted vest, with weights tied to my ankles and wrists. In the coming years, there will be some critical tasks to complete, and doing things well will become increasingly difficult.

Vitalik Buterin proposed the theory of "Functionality escape velocity," where Layer 1 provides security and reliability, and Layer 2 is used for rapid iteration. Once Layer 1 is strong enough, the rest can be left to Layer 2. Ultimately, we will stop making rapid changes to the protocol and use systems like Layer 2 to add more functionality to the Ethereum ecosystem, expanding and utilizing it in countless ways, with enough available data to handle massive transactions.

I believe that doing things well will become increasingly difficult, but it is something we must do.

a16z: The migration of Ethereum from Proof of Work to Proof of Stake has been ongoing for a long time. Sharding should be the first to launch, but according to the development of the ecosystem, the transition to the Proof of Stake consensus mechanism may happen first. What other work will the Ethereum ecosystem need to do during this process? Additionally, is it possible for other upgrades to occur during the transition to the Proof of Stake consensus mechanism?

DANNY RYAN: First of all, there are many reasons why Ethereum prioritizes the migration to the Proof of Stake consensus mechanism. One is that the costs associated with Proof of Work are too high, and this needs to stop as soon as possible; the second is that Ethereum needs to scale, especially now that many Layer 2 solutions have emerged in the market.

So, if Ethereum has 10-100 times the network capacity, it can focus on other things and complete other tasks, such as "unifying" the current two different systems: the Beacon Chain and the current Ethereum mainnet.

However, some other factors have affected the timeline for Ethereum's Proof of Stake migration, as I mentioned earlier—the EVM issue. When you start considering the direction of Ethereum's virtual machine development, issues of centralization and other security concerns arise. Over the past 12 months, Ethereum developers have sought to optimize Layer 1 to alleviate some of these issues, and extensive research has been conducted.

It is also important to emphasize the sharding roadmap and "danksharding." When you assume that these highly incentivized MEV participants exist, the entire network structure is actually simplified. Some external participants not only change developers' views on network security but also alter perceptions of protocol construction.

If you assume that these highly incentivized participants are willing to do certain things because of MEV, then suddenly you have consensus from this third-party participant, which could lead to some negative outcomes but also stimulate new types of designs.

By the way, the Ethereum Foundation's developers have built some new upgrade testing tools, so the next upgrade will focus more on writing tests rather than figuring out how to test it.

a16z: Are developers still actively discussing and researching Stateless Ethereum?

DANNY RYAN: Yes, accounts, contracts, balances, and such actually exist in a state, which is the state of Ethereum. Given their position in the blockchain, there is a real state that grows linearly over time. If the gas limit increases, the growth of Ethereum's state will accelerate—so this is a problem.

If the rate of gas growth exceeds the memory and disk space of consumer machines, it means that users cannot run Ethereum nodes on home computers and consumer hardware, leading to security and centralization issues. Additionally, as some members of the Ethereum client GETH team have pointed out, the fact that Ethereum's state continues to grow means they must constantly optimize the client, which becomes increasingly challenging.

Stateless Ethereum and related research directions are potential solutions. In fact, executing a block on Stateless Ethereum does not require verifying the entire state of the block. For example, when executing block functions, a hidden input can be set.

For instance: I need pre-state, I need the block, and then I get post-state to know whether the block is valid.

For Stateless Ethereum, state requirements (including the accounts and other things needed to execute a specific block) can be embedded in the block, and it can be proven that these states are correct. This means that executing a block and checking Ethereum's validity only requires having the block, which is really good; we can have complete nodes that do not necessarily have a complete state.

Stateless Ethereum expands the range of nodes: I might have a fully verified node without state, a node that only maintains the state relevant to me, or a very complete node (which contains all states).

The construction of Stateless Ethereum is actively underway. As far as I know, a testnet has already started testing, and we will also test some other interesting things. However, Ethereum upgrades will have some priorities, and compared to Stateless Ethereum, the demand for allocation and Layer 1 scaling is higher, so scaling may be prioritized.

Having said all this, it is hard to say which upgrade will be prioritized. We have "proto-danksharding," which seems to be a gradual scaling approach. Perhaps this upgrade will be prioritized, followed by the Stateless Ethereum upgrade, and finally the full allocation implementation. I believe that for the issue of state growth, Ethereum must have a solution path and must fix it. Although the state issue is not an urgent matter and will not undermine Ethereum's influence in the coming years, it must be addressed.

a16z: What upgrades can we expect after Ethereum's "Merge"? Will there be a cleanup upgrade? Will it be separate from the "Shanghai" upgrade? When will Ethereum introduce sharding?

DANNY RYAN: "Shanghai" may be the name of the fork after Ethereum's merge. During the Ethereum merge phase, you may not be able to withdraw the funds you have staked for nearly two years. Although Ethereum initially hoped to support the withdrawal of staked ETH right after the merge, over time and considering the complexity of the merge, developers will likely choose to "take it in two steps," completing the merge first and temporarily not adding additional withdrawal functionality.

I personally very, very, very much hope that the "Shanghai" fork will enable withdrawals—because this is the first upgrade after Ethereum's merge, and it is a commitment made by Ethereum previously, involving a lot of funds. I do not want any issues; it requires extensive testing and a lot of other work.

I believe the EVM (Ethereum Virtual Machine) has many other improvements that can be applied to this system—different mathematical operations, different scalability, better version control in the EVM, and other features. EVM optimization needs a "pressure relief valve." For years, the work of optimizing the EVM has been sidelined, as developers have invested too much time in the merge and other upgrade tasks, but people really want to see Ethereum make some small, scalable upgrades.

Therefore, after the merge, Ethereum will either implement proto-danksharding (which will lay the groundwork for allocation in Ethereum and help Ethereum achieve greater capacity) or reduce gas fees—this is easy but not a truly sustainable solution.

So, I hope to see two things in the Ethereum "Shanghai" upgrade: support for staking ETH withdrawals and scaling for the Ethereum network.

But the question is: where is Ethereum headed in the future? It is hard to say. If Ethereum can indeed make some progress in the "Shanghai" upgrade, it will undoubtedly be a very good complement for Layer 2; or, if it still cannot meet the needs of Layer 2, at least the "Shanghai" upgrade will lay a good foundation for full danksharding.