Vitalik Buterin summarizes the gains from the seven rounds of Gitcoin quadratic funding experiments

This article is sourced from iNetworkSociety, authored by Vitalik Buterin and translated by ETHPlanet.

This article is based on the speech video recorded by Vitalik Buterin for the fifth Network Society Annual Conference themed "The Practice of Wisdom Network," Forum 1 "Radical Markets," and Forum 6 "Crisis and Turning Point." The original title of the speech is "Ethereum Community Governance and Gitcoin Quadratic Funding." The author, Vitalik Buterin, is the founder of Ethereum.

Communities have always relied on public goods, especially online communities. However, fundraising for these public goods poses a political dilemma: who decides which projects to fund? How do we evaluate the quality of these choices? The innovative mechanism of "quadratic funding" allows the process of funding public goods to be both democratic and market-compliant. This mechanism is also largely applicable to various types of online communities. Although we still face many issues in implementing the scheme, at least over the past year and a half, Ethereum's experiments with "quadratic funding" through the Gitcoin funding platform have shown very optimistic results. Today, I will discuss our findings regarding the results of the quadratic funding experiments within the Ethereum ecosystem.

Today, I want to talk about quadratic funding, a new mechanism for fundraising for public goods. In the following speech, I will introduce what quadratic funding is doing, my views on it, and what we have learned from applying quadratic funding in the Ethereum blockchain ecosystem.

Imagine you have a pool of funds and want to allocate a portion of it to public goods that need funding. You may lack a clear business model to execute this idea and may not be sure which specific projects are worth supporting. In this case, you would need to rely on public opinion to decide how to allocate the available funds. Quadratic funding matches funds to projects, allowing them to receive a certain amount of donations and subsidies. The calculation formula it relies on is quite special: for a given project, you take the square root of each donation amount received, sum these square roots, and then square the result to get the total amount of matching funds for that project.

Illustration of the quadratic funding calculation mechanism

In the diagram, the large square represents the total sum of square roots, which is the total funding the project should receive, while the difference between the smaller squares is subsidized by the subsidy pool. When you do not have enough funds to support all projects, you can calculate the required subsidies for each project according to the formula and then decide how much to invest based on your available funds. For example, if your funds are only enough to pay a 1/4 subsidy to a project, you would pay that 1/4 subsidy.

But why calculate project funding in this way? Why not simply multiply each donation by five? In fact, the benefits of "quadratic" funding are twofold: first, the more people participate in donating, the better the matching effect; second, small donations will be well matched.

According to the matching mechanism of quadratic funding, when there is more than one participant, the total funding amount composed of square roots is much larger than simply adding each donation amount. This means that the more "public" the good, the more funding it receives. As shown in the diagram, when the side length of the small square is 1, the area is 4, and then the area of the square will become 4 squared, which is 16. If a project has n donors each donating 1 dollar, then the total funding the project receives is n², meaning the total amount received by the project increases by n times. When the number of beneficiaries of the project remains n, each participant will receive 1/n of the total value of the public project. At this point, the quadratic funding mechanism will significantly increase the per capita benefit for the project's beneficiaries. This helps alleviate the negative effects of the tragedy of the commons.

Moreover, quadratic funding also gives more importance to small donations. As shown in the diagram, each small square will match to a relatively large square. The larger the small square, the larger the square it matches to. From a ratio perspective, we find that the smaller squares occupy a larger area ratio of their corresponding larger squares. This indicates that smaller donations will be amplified at a higher ratio in quadratic funding. This largely avoids the scenario of "the rich dominating everything"—after all, the matching received from a 1 dollar donation may be greater than that from a 10,000 dollar donation.

The reason for using quadratic funding in Ethereum is that the Ethereum community has a large number of public goods that need funding (in fact, all online communities have public goods that need funding, as one of the biggest differences between physical space and cyberspace is that public goods may outnumber private goods in cyberspace). In cyberspace, any open-source code, articles written and published, and forum posts are public goods. Whether now or in the future, they will be widely read. The Ethereum community is no exception—there are many software projects waiting to be written, a lot of research to be conducted, and many documents to be drafted, in addition to numerous public goods in education, law, translation, etc., that need funding.

However, the current way the Ethereum Foundation allocates funds to public goods is relatively centralized. Although the foundation is trying to allocate its annual budget of 30 million dollars as fairly as possible—such as seeking cooperation with other organizations or encouraging those who receive funding to maintain their independence—it is still controlled by a small group of people. People often complain and wonder why the Ethereum Foundation would rather spend five million dollars on "zero-knowledge proof" research than support some smaller projects that are considered valuable but not prioritized by the foundation. To address this, we are trying to find a funding allocation method that is more diverse, democratic, open, and inclusive. Quadratic funding is the solution we are experimenting with.

Gitcoin is the quadratic funding experiment conducted by the Ethereum community, originating from GitHub. Initially, this platform provided rewards for those who wrote code for new projects. Now, this reward category has been integrated into the Ethereum blockchain ecosystem to support the operation of various public goods. Over the past year and a half, our quadratic funding experiment has completed seven rounds. Most of the subsidies come from the Ethereum Foundation, with others from "Consensys" and individual donations. In the recent seventh round of fundraising, some DeFi (decentralized finance) projects also made donations. Below are the experiences we gained from these rounds of fundraising. First, let’s take a look at the results of the third round of experiments.

Results of the third round of quadratic funding experiments

In my view, the third round of fundraising has reached a high level of application. The white numbers in the diagram represent the donations received by the top projects in the third round of fundraising, while the green numbers represent the total matching funds they received. As shown, "EthHub1" received 4,797 dollars from 131 donors and matched a total of 16,000 dollars. You might wonder why these projects had over a hundred donors, but their matching amounts were only two or four times the original donation amount, rather than a hundred times? This is because some donations came from the same person. To minimize the risk of cheating, we modified the formula in the third round of fundraising. The modified formula is used for "bounded quadratic funding," meaning that when a large number of donations come from the same person, their matching share will be reduced. For example, although the "Lighthouse" project had many donors, its matching amount was divided by at least ten, resulting in a final matching amount formed through subsidies that was less than the calculated matching amount.

Results of the fourth round of quadratic funding experiments

In the fourth round of experiments, we attempted to categorize projects into "technical" and "media" types. Technical projects can solve transaction privacy issues. Among them, "DAppNode" allows users to run nodes more easily; "Uniswap" is a very famous decentralized exchange; "Sablier" is a payment system that can meet subscription payment needs. The media projects include "EthHub," a media outlet that collects Ethereum-related news, "the Week In Ethereum News," and the Twitter account "@antiprosynth."

Although there are many excellent projects in both categories, media projects sparked a lot of controversy at the beginning of the fourth round of fundraising. Opponents of media accounts often argue that becoming a Twitter influencer is not a real job and does not deserve compensation, as tweeting on Twitter is something anyone can do for free. They might say, "I have a Twitter account, and I have also posted some Ethereum-related content, while the subsidized accounts are just retweeting Ethereum-related news, posting supportive messages for Ethereum, and rebutting opposing voices from other large communities." Supporters of the accounts might argue, "This is a person who has been diligently supporting Ethereum for two years, and he is only being rewarded this time," as gathering information to ensure timely understanding and providing various perspectives is very time-consuming and labor-intensive.

Additionally, there is controversy over whether media projects, while generally beneficial, may have potential harms. Some believe that certain media projects may make Ethereum appear chaotic and could provoke disputes among dissenters. However, regardless of people's opinions on specific cases, at a higher level, such controversies reflect the lack of avenues for people to express the view that "certain projects have negative value" in quadratic funding. For example, in the aforementioned situation, if 10 people believe that media projects have more positive value and 10 believe they have more negative value, then while quadratic funding allows the former to donate, the latter cannot do anything. This could lead to a project that faces opposition still receiving subsidies—even if we cannot know whether these projects will generate value or incite public anger in the real world.

Therefore, in the fifth round of fundraising, we conducted a "negative donation" experiment, meaning that people could not only donate to a project but also transfer the donations received by a project to other projects. However, this experiment did not yield ideal results—very few people actually made "negative donations," as publicly "being the bad guy" can be somewhat unsettling. Although "negative donations" were not continued in the sixth and seventh rounds due to unsatisfactory test results, we do not rule out the possibility of trying similar mechanisms in the future. Perhaps next time we will allow anonymous donations, so people will not know who made positive donations and who made negative donations.

Currently, some controversial media projects still receive funding, but the subsidies they receive are less than before. The projects that receive the most subsidies through the quadratic funding mechanism are usually those that are widely recognized and have a significant positive impact on the Ethereum ecosystem. Although the previously controversial situation has improved today, the aforementioned issues may arise again in the future, and we will remain vigilant about this.

Additionally, we discovered the issue of "income stability" in the fifth round of fundraising. While we hope that the funds obtained from quadratic funding can not only serve as people's side income but also become their main income—allowing them to become quadratic freelancers—practically, the amount of funds a specific project receives in different fundraising rounds is not stable enough as a salary. To address this, we attempted to automatically transfer a certain donation from the previous round to the next round in the sixth round of fundraising, but unfortunately, this adjustment did not significantly impact the final results. In summary, "income stability" is also one of the issues we continue to focus on.

The fifth round of fundraising also revealed issues of "collusion" and "fake accounts." Based on the working principle of quadratic funding, a user distributing donations among 50 fake accounts would receive much more matching funds than an honest user. If a user creates 50 fake accounts and pays others to donate for them, the funds matched to the project would ultimately return to their pocket—this is the "collusion" involving fraud. To prevent and address these issues, we improved some user interfaces in the sixth round of fundraising to reduce the difficulty of contributing to projects and increased the difficulty of creating fake accounts through a phone verification mechanism. In the seventh round of experiments, we added second-layer protocol payments, using "Zk-Rollup" to reduce transaction costs and conduct decentralized identity verification. Additionally, we enabled "BrightID," a decentralized mechanism based on social networks, to prove that certain accounts correspond to specific individuals.

In the past three months, many decentralized finance projects also participated in the matching of the seventh round of fundraising and received a lot of attention and funding. Below are the top 10 projects that received the most funding in the seventh round, including Bankless, Week in Ethereum News, RadicalxChange Foundation, Ethereum Magicians (a decentralized finance tutorial project), Zero Knowledge Podcast, and my favorites ETH Meme and @antiprosynth. Overall, projects with a lot of work and those that receive more matching funds tend to be more attractive.

DeFi projects participating in the seventh round of quadratic funding experiments

Results of the seventh round of quadratic funding experiments

The largest single donation received in the seventh round came from "BalanceLabs," followed by significant donations from "Synthetix Optimism," "ChainLink," and "Learn." In addition, some investment funds and individuals also made donations. This indicates that the quadratic funding mechanism is becoming increasingly important, with more and more projects willing to participate in matching and a large number of users actively participating in donations.

In conclusion, I believe quadratic funding is practical; it indeed has the potential to benefit the future. It is not only highly efficient in project funding allocation (eliminating the need for the Ethereum Foundation to allocate funds itself), but it also creates a civilized and open culture of community participation. It provides a channel for people to express their opinions, allowing them to discuss what they consider valuable in the Ethereum ecosystem. This is beneficial for creating a beautiful ecosystem with a collaborative atmosphere.

However, this mechanism also has areas for improvement. As I mentioned, it needs more channels for people to easily identify quality projects and allow for more agency authorization; its security performance also needs to be enhanced. Additionally, the projects that currently receive the most funding from the quadratic funding mechanism usually come from the United States and Europe; for projects in Asia and Africa, obtaining donations through quadratic funding is still quite challenging. Therefore, making fundraising results more inclusive and expanding the distribution of fundraising results is also a direction we are striving for.

It should be noted that implementing quadratic funding in any new community requires several rounds of effort to operate better. In the Ethereum community, at least four to five rounds of fundraising have allowed us to truly understand how quadratic funding works. Therefore, if the first round of experiments does not go smoothly, please do not be discouraged, as experiments require persistent effort. Currently, the "Gitcoin Grants" team is trying to apply quadratic funding to other scenarios; they conducted a "city stimulus" experiment, applying quadratic funding to some local businesses in Denver and Colorado—results seem promising. I am confident that if the experiments continue at a frequency of once every three months for two to three years, the results will be even better.

Overall, the quadratic funding mechanism is still being refined. We are accumulating execution experience in each round of experiments, trying to understand how to execute it correctly—especially regarding "which projects should receive donations," "how to participate," and "how the community should regulate." I look forward to quadratic funding continuing to play its role in the future and even being widely applied. I also hope that over time, quadratic funding can be incubated by the Ethereum ecosystem and even surpass Ethereum in value.

Thank you all.

Review of Radical Markets Discussion and Q&A Session

• Speaker: Vitalik Buterin

• Commentators: Yan Han, Liu Yisi

• Hosts: Huang Sunquan, Liu Yihong

Liu Yihong: Next, let's invite our commentators Yan Han, Liu Yisi, and host Teacher Huang.

Huang Sunquan: Vitalik, welcome to China online! Our two commentators, Yan Han and Liu Yisi, will give brief comments on your recent sharing. Let's start with Yan Han.

Yan Han: I will briefly introduce what Ethereum founder Vitalik just said. Before doing this segment, Teacher Huang warned me not to talk too much about mathematics, or everyone would fall asleep.

Vitalik and an organization called Gitcoin conducted a very important experiment for programmers. He hypothesized a wealthy, successful organization with social responsibility—this organization refers to the Ethereum Foundation, and this society refers to cyberspace—how do we prevent the philosopher king from ultimately becoming a "dragon"? How does young Vitalik, 20 years later, let go of power and allow this money to operate naturally to find the public goods projects that need funding the most? This is akin to how governments in the real world decide which bridges to build, which roads to pave, and how to fund education and healthcare. So, in the organization of cyberspace, how do we collaborate to ensure the success of the experiment while removing human power and centralization?

This experiment has also sparked a lot of controversy. For example, some believe that scientists and engineers should be compensated, but educational and communicative work does not deserve excessive compensation. Another controversy is: can dissenters cast "negative votes"? We know that there are political votes in the real world, so can we create a "negative score" vote to express opposition to a candidate or proposal? With the help of tools like quadratic voting and quadratic funding, can "negative votes" also generate value? Then everyone started arguing.

There is also a deeper question—how long can such a model be sustained? We see that throughout real history, most republics eventually devolve. I actually want to ask Vitalik the first question today as an audience member: we know that Ethereum is about to enter its sixth year, while Bitcoin is in its eleventh year. The infrastructure of the internet has been around for 50 years now. We see that some things that predate Bitcoin and Ethereum have now devolved into dictatorship and centralization. As a founder, how do you ensure that your project will not face this issue in the future? How much do you think quadratic voting and quadratic funding can help?

Vitalik: I understand your question. We can only say: this is an experiment, and we do not know what the specific results will be. Ethereum has been around for 6 years, while quadratic funding has only been around for 2 years, and the world is constantly changing, changing very rapidly, so we cannot promise that quadratic funding will become a solid infrastructure that can equally support public welfare projects in Ethereum for decades to come.

But as an experiment, it has been relatively successful so far, at least more successful than many people expected. It has shown very good results in many aspects and indeed provides us with some alternative methods—such as the Ethereum Foundation—can provide funding and support for projects that are truly beneficial to the Ethereum ecosystem.

So I want to say that quadratic funding has potential and is effective. We can continue to experiment with this across the entire cyberspace. Although whether these experiments are sustainable and can move into the future is still unknown, we are currently very determined to continue pushing this experiment and working hard to expand it to confirm whether it can be extended to broader fields and platforms.

Liu Yisi: Thank you for your recent sharing. I actually think that a lot of what Vitalik just mentioned is that quadratic funding and quadratic voting are both good methods, but there are various issues within them. For example, he mentioned in Round 5 that if someone has 50 dollars, or if I have a funding pool, I can divide it into 50 parts and invest them one by one to maximize the matching of funds in this pool. Or some people create a fake project, colluding with others to fundraise for this fake project, because I will never lose; your Gitcoin platform will help me match a lot of funding.

I am very interested in such issues because they relate to trust. When Vitalik created Ethereum, he actually wanted to create a decentralized community and platform. Such decentralized platforms are indeed difficult to avoid these issues. Therefore, Gitcoin later used phone number verification for identity, and later introduced BrightID as a decentralized identity verification method. But are these methods truly decentralized?

Actually, I am not very clear on how BrightID works. Vitalik, could you explain how we can avoid such fraudulent activities while ensuring decentralization in such an environment?

Vitalik: This question is difficult to answer. The idea behind BrightID is that when we can see who verifies whom and who is verified by whom, it may form a decentralized way for individuals. Whether this mechanism can be effective, I do not know yet, as it is still just an experiment.

Liu Yisi: Are there better solutions?

Vitalik: Our identity verification solutions currently consist of two types: "centralized solutions" and "decentralized solutions." Centralized solutions are familiar to everyone; for example, phone numbers can be verified, calls can be verified, and there are many ways to do so. Decentralized methods have not been done before, and we still do not know whether these methods are secure. Gitcoin Grants uses both methods: verification can be done through phone numbers or through BrightID. Perhaps in the next round of adjustments, we will add more identity verification solutions. Our plan is to continue experimenting with a combination of centralized and decentralized methods. Whether these experiments will succeed is still unknown, as they are all relatively new.

Huang Sunquan: We still have some time left, so we can open the floor for questions from the audience.

Audience 1: I have two questions. The first question is how to evaluate art within the Ethereum framework? The second question is a radical hypothesis: if I initiate a crowdfunding campaign on Ethereum to destroy Ethereum and gain the support of most people, and this is a real project, how would Ethereum execute this? If it truly executes and destroys itself, does this also prove the decentralized nature of Ethereum? How do you resolve this contradiction?

Vitalik: There are actually two ways to interpret this matter.

First, users can only propose to stop the Ethereum blockchain or destroy Ethereum when the Ethereum protocol has its own built-in governance mechanism, and then we can discuss what would happen if such a proposal were to pass. First of all, the Ethereum protocol does not have such a built-in on-chain governance mechanism, at least not at this point. Moreover, I believe that to a large extent, people do not want to add such an on-chain governance mechanism. In other words, adding such a built-in on-chain governance mechanism would encounter a lot of resistance.

Second, we can consider the situation of Gitcoin quadratic funding itself. I believe that the assumption of a project attempting to attack or destroy Ethereum is actually part of a broader issue I mentioned earlier: how does quadratic funding prevent bad projects and allow people to express dissatisfaction with these projects? This question is not difficult to answer. The Gitcoin team is also thinking about corresponding solutions, such as negative voting or having an approval mechanism. There are many feasible options now, and in the future, we will continue to explore various possibilities.

Regarding your first question, I believe that art itself is very valuable, so I think the question is about "who evaluates the value of art." For example, if you have an art project, you can completely use quadratic funding to obtain project funding. Currently, there are no artists trying to raise funds for art projects on Gitcoin; I wonder if anyone will be willing to try this in the future. We can see what results this will yield. As for art, the only evaluation I can currently make is how people value a particular piece of art, which requires further exploration.

Audience 2: I have two questions. The first question is: when media projects attract many fans due to their own influence, will the intense voting of these fans affect the fairness of project subsidies or result in funding that does not match their actual value?

Vitalik: There is indeed a possibility of this (but I cannot say) whether this phenomenon is good or bad.

For example, the democratization process inevitably comes with some controversies and structural issues. Some structural issues may have a larger audience, while others may have a smaller audience. I think this phenomenon is unavoidable. (Those topics with more audiences) may indeed have an excessive influence. If that is the case, is there a solution? Perhaps we need to explore this.

Audience 3: The second question is: as you build quadratic voting and adjust the quadratic funding mathematical formula through continuous practice, are you becoming a new rule-maker and authority? After all, no matter how adjustments are made, the rules will always have biases. In this process, whose interests take precedence over others?

Vitalik: The long-term plan and idea for quadratic voting and quadratic funding is that we hope to set a final rule in a year or two, and then we won't have to do anything. At that point, the mechanism will become automatic.

Currently, adjustments are frequent because this experiment is still relatively young, and we have many unknowns. At present, there are still many details in quadratic funding that can be adjusted, and we are conducting various experiments to seek the optimal solution. But in the long run, we hope not to have to make mechanism adjustments.

Audience 4: What are the similarities or differences between quadratic funding and venture capital?

Vitalik: Quadratic voting and Ethereum's goals are different. Ethereum primarily provides funding for certain projects. These projects are often profit-oriented, at least with plans to be profitable in the future. However, the quadratic funding scheme supports projects that, while important, do not have a clear business model. The two serve different purposes.

But I believe there is indeed some overlap and interaction between the two. On one hand, Ethereum can invest in certain projects, and these projects can also receive quadratic funding; on the other hand, some projects supported by quadratic funding can also support the Ethereum Foundation. In other words, quadratic funding directly invests money into the Ethereum Foundation, which commits to pursuing goals that are more aligned with community demands. Initially, these projects can receive funding from quadratic funding, but in the future, when they find a clear business model, they can seek funding from the Ethereum Foundation.

I hope there can be good interaction between the two. But at the same time, we must remember that their goals are different.

Audience 5: How does Gitcoin address regional issues? For example, very few Chinese teams receive Grants, and Gitcoin Grants can only support projects in the English-speaking world. How do Ethereum and Gitcoin Grants handle localization and regional equity issues?

Vitalik: We do not know whether this is an issue with the quadratic voting mechanism or its promotion method.

Currently, most of Gitcoin Grants' promotions target English-speaking communities. Therefore, members of the English community often know what quadratic voting is. Gitcoin Grants currently does not have a Chinese version, and aside from the English version, there are no other language versions. Although there are indeed some Chinese community projects participating, the number is small. I think we may need to solve this problem through more balanced promotion. Of course, another method is to create a dedicated matching pool. For example, next time, allocate 30,000 dollars to the Asian community, with different pools and funding for each region. In short, this is a particular issue we are paying close attention to. This issue will definitely be addressed in the next round of improvements; we just haven't determined the specific method for improvement yet. This issue is still under discussion.

Yan Han: Indeed, very few people realize that projects from the Chinese community can also participate in Gitcoin Grants—due to reasons everyone knows. Vitalik's proposed method is somewhat similar to the affirmative action in American universities—giving some disadvantaged groups extra points, and then balancing it in the future.

Huang Sunquan: Thank you very much, Vitalik. Next, we will enter the intermission. Let's invite Teacher Liu Yihong. Thank you, Yan Han and Liu Yisi.