Dialogue Vitalik: PoW will eventually transition to PoS, token-driven governance is an outdated model

Original Title: “Interview: Vitalik Buterin, creator of Ethereum”

Author: Noah Smith, Noahpinion

Translator: Cookie, Chain Catcher

Vitalik Buterin is one of the most well-known and beloved figures in the crypto world. The Ethereum he co-created with Gavin Wood has become a leader in the entire Web3 space. It is evident that he is a smart and friendly person who just wants to create cool things rather than deceive anyone. Besides being one of the spokespersons for cryptocurrency, Vitalik also has an interesting Twitter account and a blog where he often shares deep original insights on various crypto topics.

As we mentioned, Ethereum, which powers most smart contracts in the crypto world, is undergoing a historic transformation. In a process called The Merge, which is planned to be completed in two weeks, Ethereum is transitioning its method of validating transactions from Proof of Work (PoW) to Proof of Stake (PoS). This will enable the network to significantly reduce energy usage and carbon emissions.

In the following interview, I discussed various topics with Vitalik, including Proof of Work vs. Proof of Stake, the recent crash of the crypto market, cryptocurrency security, decentralized governance, and "network states."

Noah Smith: I think we should start with some current events. Almost all cryptocurrencies have crashed in recent months. Why do you think this happened? What impact will it have on the future of the blockchain ecosystem?

Vitalik: Actually, I was surprised that the "crash" happened so late. Typically, crypto bubbles last about 6-9 months after the peak of a bull market, followed by a rapid decline. This bull market lasted nearly a year and a half, and people seemed to have developed a mindset that high prices for cryptocurrencies were the new normal.

I always knew that the bull market would eventually end and we would experience a downturn, but I just didn't know when. The current feeling is that people do not have a deep enough understanding of the cyclical dynamics of cryptocurrencies. When prices go up, many people say this is a new paradigm and the future, while when prices drop, people say it is doomed to fail and has fundamental flaws.

I do think that the price drop helps to "reveal" certain issues that have existed from the beginning. Unsustainable business models often succeed during boom times because everything is rising, so the funds people have at their disposal are also increasing, allowing these models to be temporarily supported by the constant influx of new dollars.

As we saw with Terra, this model became ineffective during the crash. This is most prevalent in extreme cases like high leverage and Ponzi schemes (crypto players from 2017 will remember "BIT-CONNE-EEE-ECT!!!"), but it is also true in more subtle ways, such as how new protocols need to consider how easy it is to maintain high yields during a bull market; when prices crash, it is often difficult for newly formed teams to sustain themselves financially. Besides the usual advice I give, which is for people to remember crypto history and take a long-term view, I have no way to combat these cycles.

Editor’s note: BIT-CONNE-EEE-ECT refers to the Bitcoin investment lending platform BitConnect, which suddenly announced in January 2018 that it would shut down its lending and exchange services. At its peak, BitConnect had a market cap of over $2.6 billion, promising investors returns of up to 40% per month.

Noah Smith: That makes sense. Now, please talk more about the financial aspects. For Bitcoin—the most widely held and traded cryptocurrency—we have seen this pattern of recurring bubbles and busts, but each time the percentage return during the boom is lower than the last. To me, this looks like a curve— as more and more people hold cryptocurrencies, the economic benefits for new users are getting smaller. Have we reached a saturation point for Bitcoin adoption, where the returns have dropped to gold levels?

Vitalik: I think that in the medium term, cryptocurrencies will stabilize, and their volatility will be similar to that of gold or the stock market. The main question is at what level the prices of cryptocurrencies will stabilize. In my view, a lot of the early volatility was related to the uncertainty surrounding cryptocurrencies: in 2011, when Bitcoin dropped from $31 to $2 in six months, people doubted that Bitcoin was just a passing trend and would eventually collapse. By 2014, that uncertainty was less than before, but it still existed. Then after 2017, the issue of uncertainty shifted to whether cryptocurrencies could rise to higher prices, contingent on whether they could gain the necessary mainstream recognition and legitimacy.

Although we have come a long way, this is still roughly the level we are at in 2022. Over time, these existing issues will become clearer. If by 2040, cryptocurrencies have robustly entered several niches: they have replaced a share of gold as a store of value, they have become a kind of "financial Linux," ubiquitous and always available.

Alternative financial products are the truly important goal for cryptocurrencies, but that does not mean they will completely replace mainstream financial products. Therefore, the chances of them disappearing in 2042 or completely taking over the world will be much smaller, and the impact of individual events on that possibility will be very minimal.

Some die-hard individuals believe that the price of cryptocurrencies is trapped within a limited range (between all wealth in a zero-sum world), and cryptocurrencies can only remain highly volatile within that range for a long time until repeatedly buying high and selling low becomes a mathematically almost certain winning arbitrage strategy.

Noah Smith: Additionally, estimates of Bitcoin's energy usage suggest that the network's energy consumption is closely related to the price of Bitcoin. This is not the case for stocks, houses, or gold—these assets do not require increased energy usage to support higher prices. In the long run, this seems to represent a force that will suppress Bitcoin's price, doesn't it?

Vitalik: I generally view the interaction between the demand and supply curves for Bitcoin, and the issues that arise from supply, as two separate problems. Difficulty adjustments ensure that the number of Bitcoins mined is fixed according to a schedule: currently, 6.25 BTC are mined every 10 minutes, decreasing to 3.125 BTC every 10 minutes starting around 2024, and so on. This schedule applies regardless of total hash power or price. Therefore, from an economic perspective, it does not matter whether the protocol hands these tokens to miners or core developers. That is why I do not agree with the mindset that miners somehow "support" the value of Bitcoin.

A consensus system that consumes a lot of electricity is not only harmful to the environment but also requires the issuance of hundreds of thousands of BTC or ETH each year. Of course, the ultimate issuance will decrease to nearly zero, at which point this will no longer be an issue, but then Bitcoin will face another problem: how to ensure the security of the network…

These security motivations are also a very important driving force behind Ethereum's transition to Proof of Stake.

Noah Smith: Let's talk about those security issues. Many people seem to believe that if a token is called a "cryptocurrency," then the protocol that determines the transfer and ownership of that token must be secure. However, from what you've just said, you seem to be more concerned about security, at least regarding Bitcoin. Can you elaborate on that?

Vitalik: Efficiency and security are not independent issues. The core question is: how much security can you buy for every dollar you spend each year? If a system's security is too low, you can increase security by offering more cryptocurrency as incentives, at which point you have gained security at the expense of efficiency. I delve into some economic reasons in this article explaining why Proof of Stake can achieve about 20 times the security at the same cost. Essentially, the ongoing costs and entry costs for Proof of Work miners are moderate, but the ongoing costs for Proof of Stake validators are low while the entry costs are high.

It turns out that the level of security of the network depends solely on the entry costs, as that is the cost attackers must pay. Therefore, the consensus system of the network should have low ongoing costs and high entry costs, which is what PoS excels at. Additionally, there are differences in how both recover the network after an attack: in PoW, the network can only recover by changing the PoW algorithm, but it will discard all existing mining hardware. However, in PoS, the network can have a protocol and only forfeit the assets of the attacker, so the attacker has paid a lot, but the ecosystem can quickly recover.

Regarding Bitcoin, I have concerns for two reasons. First, in the long run, Bitcoin's security will come entirely from fees, and the network has not successfully reached the required fee revenue levels, making it difficult to become a system that could potentially be worth trillions of dollars. Bitcoin fees are about $300,000 per day and have not really grown that much over the past five years. Ethereum has been much more successful in this regard because Ethereum is more designed to support users in using and building application ecosystems. Second, compared to Proof of Stake, the security provided by Proof of Work for every dollar of transaction fees is much lower, and it seems unfeasible for Bitcoin to migrate away from Proof of Work. If in the future the Bitcoin network reaches $5 trillion, but attacking the blockchain only requires $5 billion, what will users think? Of course, if Bitcoin is really attacked, I hope they will at least show a willingness to shift towards a hybrid Proof of Stake, but I expect that to be a painful transition.

Noah Smith: Regardless, your argument about the security per dollar provided by Proof of Stake is very compelling. The high energy costs of Bitcoin are essentially security costs. However, let's talk about why Bitcoin supporters are reluctant to accept any alternatives outside of PoW. Does the idea of Proof of Stake allow large stakeholders to modify the network protocol for their own advantage and exploit user fees? Has Proof of Work created a class of large miners, and will there still be an incentive for them to protect their ongoing income in the future? Even if these miners' income represents the holding costs for users.

Vitalik: There are some arguments against Proof of Stake. In my view, the most compelling is the "Costless Simulation" problem. The idea is that in a Proof of Stake network, an attacker can contact token stakers from years ago to buy their old private keys at a very low price (since those tokens have already been transferred) and use those addresses to create a different chain that forks the main chain. During the vacuum period after the fork, the historical data of this chain looks valid. A node that connects to the network from scratch, knowing only the protocol rules, will not be able to distinguish between the actual chain and the simulated chain provided by the attacker. However, in PoW, creating such a simulated blockchain requires redoing an equivalent amount of Proof of Work.

Editor’s note: The PoW mechanism follows the longest chain as the correct blockchain, and an attacker must expend a large amount of computational power to forge the longest blockchain. In contrast, PoS validators do not need to perform extensive calculations; they simply take transactions from the transaction pool, package them into blocks, and broadcast them. Therefore, under the PoS mechanism, forging a chain as long as the main chain incurs no cost.

In the PoS mechanism, this issue will be addressed by introducing a weak subjectivity: nodes need to occasionally connect to the network (for example, once a month), and the first node to sync may need to ask some trusted sources (which do not have to be centralized) to transmit the correct version of the blockchain. Stakers must lock their tokens during this time, and if someone sees a staker supporting two conflicting chains, they can send a transaction to "slash" them, burning a large portion or all of their token assets. In the model, all of this is completely reasonable. But PoW supporters feel uncomfortable with weak subjectivity; they prefer a pure approach where validators only need the protocol rules.

In reality, I think pure idealism does not work in practice. Regardless, validators need trusted sources to provide protocol rules, especially considering that validators will receive software updates to improve efficiency or occasionally fix bugs. Moreover, I believe that the attacks that pure idealists fear will not occur in reality: you have to convince a large group of people that the recent block hash is incorrect, while no one other than the attacker can determine which other hashes are correct. Once you start to delve into the details, it seems unfeasible.

Some also try to claim that PoS allows large stakeholders to control the protocol, but I think this argument is completely wrong. They mistakenly believe that PoW and PoS are governance mechanisms, whereas they are actually consensus mechanisms. What they do is help the network reach consensus in the right direction. Blocks that violate protocol rules (for example, a block trying to gain more reward tokens than allowed by the protocol) will not be accepted by the network, regardless of how many miners or stakers support it. Governance is a completely separate process involving user discussions of BIPs and EIPs, calls from all core developers, and coordination with other official bodies to propose improvements. Interestingly, Bitcoin holders (who tend to be the most supportive of PoW) should understand this well, as the Bitcoin civil war in 2017 demonstrated the impotence of miners in the governance process. The same is true in PoS; stakers simply execute the rules and help package transactions.

Translator's note: The Bitcoin civil war refers to the split in the Bitcoin community over the block size issue in 2017, which subsequently forked into another public chain, Bitcoin Cash (whose token was BCC, later renamed BCH).

One possible argument is that PoS has greater centralization pressures than PoW because the nature of digital stakes makes it easier to centralize, or because the PoW mechanism involves stakeholders leveraging local power to obtain cheap electricity. These are definitely concerns of mine, although I think people exaggerate this issue. Especially currently, Ethereum's Proof of Stake has not yet enabled stakers to withdraw ETH. If stakers participate in staking in a liquidity pool, then at some point when they want to withdraw funds, they can sell their stake to others, thus providing a significant competitive advantage in terms of liquidity. However, this will no longer apply once withdrawals are enabled next year.

Another issue with staking at present is that, also due to the lack of withdrawal capability, stakers cannot easily switch liquidity pools (or switch to solo staking), but they will be able to do so next year. As for the decentralization of mining machines, I am just not sure whether these highly decentralized small-scale mining pools are significant enough. Mining is a highly industrialized activity, and large mining farms outside the U.S. (which account for about ~35% of global hash power) seem closely tied to various governments, so the censorship resistance of PoW has strong contingencies in the future. The highly democratized early days of Proof of Work were a wonderful thing, greatly helping to make cryptocurrency ownership more equitable, but it is unsustainable and will be phased out.

Noah Smith: Let's really talk about governance. To me, governance seems to be one of the most promising and interesting aspects of blockchain technology— a way to bypass cumbersome business processes and create fluid, temporary economic cooperation, especially in the realm of cross-national collaboration. I am a huge fan of the sci-fi novel "Rainbow's End," where much of the economy is based on this cooperation. But so far, it seems there are many issues with how people are trying to achieve this— in fact, you've written a lot on your blog criticizing blockchain governance systems that attempt to eliminate all human judgment and trust. Can you briefly outline your views on how blockchain governance should work?

Vitalik: One reason blockchains are interesting is that they share many attributes with things we are already familiar with, but they are not exactly like any of them. Like companies, blockchains offer tokens that can be purchased, and holders hope they will increase in value. But unlike companies, blockchains are more like nations, not relying on external authorities to resolve internal disputes. Instead, blockchains are "judges" capable of adjudicating their own issues; you could even say they aim to be "sovereign states" (of course, blockchains are not truly independent of existing nation-state infrastructures. But in reality, most nation-states also do not achieve true independence).

Like democratic nations aspire to be, blockchains are highly open and transparent, allowing anyone to verify whether the rules are being followed. Blockchains often give rise to something akin to religion, inspiring followers to develop a lasting and fervent enthusiasm, but their economic components are much more complex than what religions typically do. Blockchains are like open-source software projects, with egalitarian ideals, and more importantly, the ability to fork freely: if the "official" version of the protocol goes astray and violates core values that some parts of the community believe in, the community can fork around their own ideals, and then they can compete for legitimacy with the original protocol.

But blockchains are different from open-source software projects: in blockchains, billions of dollars of capital are at stake, and the costs of forking can be much higher. Just as some believe that the scale of DeFi prevents Ethereum from experiencing an Ethereum Classic fork, if the costs of forking are too high, its role in governance becomes more like nuclear deterrence rather than conventional weapons.

All of this means that blockchains are a powerful infrastructure that can carry the governance logic of other applications, but the blockchain itself is also a complex entity that requires a new and different form of governance. We have seen various forms of "constitutional crises" in both Bitcoin and Ethereum, most notably the Ethereum DAO fork and the Bitcoin block size debate. In both cases, both sides had strong and differing beliefs about the values that the protocol should embody, ultimately resolving through a fork. Interestingly, both Bitcoin and Ethereum have avoided formal governance: there is no specific individual or council or voting mechanism to decide which protocols can become officially legitimate protocols. Core developers have made calls, but even then, there are no clearly defined rules, and for any truly contentious issues, core developers often step back and listen to the community's opinions.

Of course, there are often voices saying that this "anarchic" design looks ugly and needs to be replaced with a more "proper" formalized system. But they have almost never succeeded. In my view, our current "unstructured governance" actually contains a lot of wisdom. In particular, it captures the idea that a relatively small core development team should be able to independently make detailed technical decisions that do not truly affect the core vision, but deeply operationalizing certain things requires high philosophical theories, such as saving hard fork events or switching to Proof of Stake.

Governance of applications on the blockchain is a different challenge. There is also a divergence in the "forkability" of applications: whether it is like E Noah Smith, where if governance fails, the community can fork using different rules and persuade all infrastructure to migrate to the new rules. Stablecoin DAI relies on reserve assets like ETH, so you cannot safely fork DAI without forking other underlying assets. If an application is forkable, it provides you with additional support, and you can leverage it (as Hive has done). If the application is not forkable, then you really need some fully formal governance that can be trusted.

For a long time, I have believed that the currently popular token-driven governance model (where governance is completed by token holders voting) is really not suitable, and we need to move towards something better, especially something less "financialized." Token-driven governance naturally favors the wealthy, and many ways will emerge in the long run that could cause the system to collapse. In my post from last year, I described a smart contract that allows token holders to automatically accept bribes in a very user-friendly way to vote in a specific manner from the highest bidder:

This would turn every governance decision into an auction, leading to only the wealthiest participants having any say, with the community pursuing soulless profit maximization, and in the worst case, this could lead to the rapid extraction of community wealth, causing the project to collapse quickly.

I believe the best alternative to token holder governance is some form of multi-stakeholder participatory governance that seeks to formally represent stakeholders, not just token holders. Optimism is doing this with the concept of "citizenship," which is intended to be allocated to contributors and ecosystem participants, and explicitly states that this authority is non-transferable. But we are still in the early stages of figuring out how such things will work.

Noah Smith: Let's talk more about alternative forms of human organization that blockchain might enable. I really appreciate your deep commentary on Balaji Srinivasan's network states. So far, have there been any effective attempts to create "network states" that incorporate crypto?

Vitalik: So far, I think the reason this concept has not been truly realized is that there is a fundamental difference between the blockchain ecosystem and comprehensive network states. The blockchain ecosystem survives economically by persuading many people to participate. You only need a few core developers, and they don't even have to make significant personal sacrifices; they can live in a "normal" real city, looking like any other job. But a network state is something deeper. It requires people to take risks and move to a specific place, which may not be a conventional location, and it comes with significant downsides that can only be overcome by the advantages created by the community itself.

I believe the crypto narrative is deeply permeated by "morality," which was central to the ecosystem from 2009 to 2014, when ordinary crypto users had almost no offline crypto social circles and even legal issues were still uncertain. Crypto, as a grand movement of internet libertarianism, was almost seen as the spiritual successor to PGP, BitTorrent, Tor, Assange, Snowden, etc. These strong ideals served as the glue of thought and morality, prompting people to make significant sacrifices and take risks for the space.

Recently, the industry has matured, and with this maturation, the moral-related condemnations have been somewhat diluted. This dilution is beneficial for the mainstream adoption of cryptocurrencies; in fact, newer blockchain projects often intentionally downplay this eccentricity, aiming for mass adoption. NFTs are expanding the appeal of cryptocurrencies to groups far removed from their original users.

But at this point, this mode of growth has also made existing blockchains too "slim" to form a good network state. Ethereum has so many different user communities, many of which have significant disagreements (for example, there is certainly a debate between the "woke" and rationalists, not to mention international disagreements). There is a strong consensus mechanism regarding defending the integrity and operation of the network, as we have recently seen the community unite around preventing on-chain censorship, but there has not been enough unity to form a state.

Moreover, so far, attempts at offline crypto communities have also been poor. The problem I see is that they all basically use some form of "low tax" as a slogan; while low taxes are an appealing benefit from an individual perspective, if your goal is to attract truly interesting people, communities formed around low tax slogans are really boring and lame. Network effects are about quality, not just quantity. I think a "network state" has a lot of demand for a community oriented around specific values, needing to provide a constructive channel to express those values, not just through zero-sum games or Twitter wars, combined with the need to escape the high cost of living in the U.S. and offline community needs, not just the theoretical authoritarianism of many other major countries. But so far, the projects I have seen have not done well.

This answer is mostly about culture. Whether we have on-chain land registration, smart contract property rights, or other minor things. I believe "network states" should try ideas that are radically different from our habits. For example, I would try to downplay the idea of absolute ownership of specific land, houses, and apartments, and emphasize economic consistency of the community through something like "crypto cities." I think innovations like this have more long-term value, and just this alone is not enough to attract people in the short term. Initially, the concept's combination with cryptocurrencies and blockchain technology is mainly symbolic, and over time, it will evolve into something more practical.

Noah Smith: One last question: in that article, you disagreed with Balaji's description of the importance of dictatorial leaders to "network states." Do you think your role as a founder of Ethereum and a "spiritual pillar" has been overstated by the media and crypto enthusiasts?

Vitalik: From the very beginning, I have hoped that as Ethereum develops, my influence would gradually diminish. Many amazing works have already emerged on Ethereum. In fact, I think this has been happening over the past two years. In 2015, I was basically doing 80% of the research on Ethereum, and I even did a lot of Python coding. By 2017, I was doing much less coding, probably only 70% of the research work. By 2020, I might have only done a third of the research and coded very little, but I was still doing most of the "high-level theory."

However, in the past two years, other people have begun to emerge who can propose high-level theories. We have many great new Ethereum influencers, like Polynya, who has been providing a lot of thought leadership around Layer 2 scalability. The Flashbots team has been leading the entire MEV research field. People like Barry Whitehat and Brian Gu have taken on responsibilities for zero-knowledge proof technology, while Justin and Dankrad, who were initially hired as researchers, are increasingly showing qualities of thought leaders.

This is all a good thing! I think public perception has not fully shifted yet, but I hope that over time, people will begin to understand this.

Noah Smith: Okay, that's my last question. Are there any recent projects that excite you?

Vitalik: I would say what excites me the most is not any single project, but rather the way the entire ecosystem is coming together, composed of many interesting ideas. Technically, this is true; Ethereum is about to undergo The Merge, and significant improvements in blockchain scalability, usability, and privacy will soon follow. On the level of social and political ideas, many thoughts around decentralized organizations, radical economies, democratic mechanisms, and internet communities are maturing.

At the forefront of non-crypto science, advancements in biotechnology and artificial intelligence are astonishing. Some might say this is a bit exaggerated. We are starting to understand what the politics and technologies of the 21st century will look like and how each part we are studying will fit into this scenario.

By 2022, cryptocurrencies have finally become meaningful. Many mainstream organizations and even governments are using it as a way to send and receive payments, and I have a vague feeling that other applications will soon emerge. The future still feels uncertain, but compared to before, we have a better view of how all of this will develop.