*This text is taken from the interview published on our YouTube channel
The Big Whale: You are one of the main contributors to MetaDAO, a Solana-based prediction market protocol directly inspired by the concept of "futarchy", invented by economist Robin Hanson in 2000. Can you explain this concept?
Proph3t: Let's start with prediction markets, where people bet on future outcomes. Polymarket is a well-known example. Historically, these markets have shown great accuracy, often outperforming professional pollsters, particularly in predicting the outcome of US elections. In fact, 74% of the time, prediction markets beat pollsters. They have also proved more accurate than government forecasts in areas such as meteorology. Companies such as Google and HP have used internal prediction markets, where employees bet on outcomes such as future product sales. These markets consistently outperformed traditional forecasting methods.
The idea behind futarchy is to exploit this excellent ability of prediction markets to gather information and make predictions. But beyond simply predicting the future, the aim is to make better decisions. The central concept of futarchy is to use markets not only for forecasting, but also to make decisions directly.
MetaDAO is the first concrete implementation of this concept. Although the idea of futarchy was proposed by Robin Hanson some 25 years ago, MetaDAO is the first live application of the idea. We are currently developing software to enable other DAOs to adopt futarchy as well.
Why might the way MetaDAO works be more effective for decision-making in DAOs and potentially even in the governance of traditional companies?
Let's take a classic example: a company that has to decide whether or not to make its CEO redundant. Traditionally, the board votes on such a decision, but this approach has its limits. Boards can be disconnected from the interests of shareholders, prioritising their own careers or avoiding legal risks over maximising shareholder value.
In a futarchy, the process would be radically different. Let's say IBM is considering thanking its CEO. Participants could bet on IBM's share price according to two scenarios: firing or keeping the CEO. If the market predicts that IBM's share price would rise to $110 if the CEO left, but only to $97 if he stayed, this is a clear signal that the market believes the company is more valuable without the CEO. This market forecast would then form the basis for the final decision - to fire or keep the CEO, depending on the market's verdict.
Do you really think this is a reliable way of predicting share prices? Is it possible to incorporate other parameters into the decision-making process? And how can we ensure that these parameters are relevant in guiding choices?
I think that in the context of a company or a DAO, the value of the token is generally a very clear indicator. The traditional objective of a company is to maximise shareholder value, although that approach is now being challenged. For a government, Robin Hanson, the father of futarchy, often says: "Vote on values, bet on beliefs". This means that in a country, citizens would vote on the parameters they care most about, and betting markets would predict how different actions would influence those parameters.
Let's take the example of a hypothetical country that votes that its priority is to increase GDP per capita - that's their only concern. Citizens would then place bets on the policies most likely to achieve this objective. They might bet on the impact of a new development programme or other initiatives on GDP per capita. If the markets predicted that a particular policy would be effective, the country would implement it.
For MetaDAO, all the proposals are aimed at increasing the value of the META token. But is this really the only interesting metric for assessing the value of a protocol like GDP for a country?
Is the value of the token really the best indicator for MetaDAO? This is undeniably a subject for debate. I've thought long and hard about it, and I've yet to find a better indicator. The value of the token encompasses many aspects, as it reflects the participants' estimates of the future value of the project across multiple dimensions. However, I'm open to suggestions if someone comes up with a better indicator. For the moment, it's the best tool available to us.
"Prediction markets essentially work like bets" Can you elaborate on what MetaDAO does today?
To illustrate how it works, I can give you a concrete example. Recently, we had a proposal concerning the development of a new product that was potentially crucial for the future of MetaDAO. Participants could trade on two separate markets - one for accepting the proposal, the other for rejecting it - using the MetaDAO token and USDC (Circle's stablecoin ).
These markets essentially work like betting. Participants make trades based on their predictions of the token's value depending on whether the proposition is accepted or rejected. In each market, people can sell Meta for USDC or buy Meta with USDC. If the proposal is not accepted, all transactions in the acceptance market are cancelled - buyers get their USDC back, and sellers get their tokens.
In this particular case, the proposal was not accepted, which means that the price in the acceptance market did not reach the required threshold. As a result, all transactions in this market were cancelled: those who had bought Meta in anticipation of the proposal's success got their money back, and those who had sold got their tokens back. Only the transactions in the rejection market were executed, thus determining the final result.
What motivated you to launch and contribute to the creation of MetaDAO?
As an engineer by training, I've always been fascinated by markets. Ever since I was a child, I've been studying the equity and bond markets, trying to understand how assets are valued. This passion naturally led me to DeFi (decentralised finance). When I discovered it, I immediately saw its potential as a new market infrastructure capable of supplanting the old one - I knew I had to make a contribution. So I worked for a while as a smart contracts engineer in the DeFi on Ethereum .
However, I soon realised that the way DAOs currently worked left a lot to be desired. The real trigger came when I came across a YouTube video by Robin Hanson. In it, he presented a revolutionary idea: using the markets to make all the world's decisions. That's when it all really started. We launched MetaDAO in November 2023, about a year ago. What started as a few experimental proposals has now grown into a full-fledged project, supported by a committed community.
Why do you think prediction markets are gaining so much popularity at the moment, especially with events like the Donald Trump-Kamala Harris duel?
It's mainly due to the public attention they've attracted. People discuss probabilities and market predictions because they are deeply invested in events such as elections. For many, myself included, prediction markets offer a benchmark deemed more reliable than traditional polls.
As for the appeal of platforms like Polymarket, part of it is the excitement of betting on election outcomes. Participants enjoy betting money on their favourite candidate. Historically, prediction markets for the US presidential election have been a flagship category, although other types of market have not enjoyed such a boom. This popularity is therefore the result of a combination of interest in the event itself and the belief that these markets offer valuable insights.
But how do you explain the large gap between Donald Trump and Kamala Harris on Polymarket right now?
It's crucial to make a clear distinction between prediction markets and traditional polls. A poll asks people directly about their voting intentions and then uses statistical methods to estimate the population's behaviour. These polls are essentially retrospective, reflecting current opinions.
Prediction markets, on the other hand, aim to anticipate the actual winner. They provide a probability for each outcome, going beyond a simple snapshot of public opinion. This is why their results can diverge from the polls.
With regard to the gap between Trump and Harris, I must confess that I no longer follow politics closely. My interest in futarchy has made me aware of the flaws in our current political systems, which has reduced my attention to the subject. Nevertheless, one possible explanation could be that market participants are anticipating a future event not yet reflected in the polls. Another hypothesis could be that a major investor is betting heavily on Trump, temporarily distorting the market. However, prediction markets tend to self-correct in the face of such manipulation attempts, as there is a financial incentive for others to bet against this distortion and profit from it.
I recently observed an attempt to manipulate the Polymarket market in the other direction, but it was quickly corrected. Currently, the market consensus seems to be about 60/40 in favour of Trump. In essence, market participants are anticipating certain probabilities that are not yet reflected in traditional polls.
"Compared to Polymarket, MetaDAO aims to be more of a decision market" More generally, what do you think of Polymarket? Is it the future of betting, or is there more to it than that?
First, I think it's important to clarify that MetaDAO and Polymarket are not doing exactly the same thing. Both protocols use the principles of prediction markets, but MetaDAO is intended to be more of a decision market.
I'm a bit pessimistic about Polymarket and prediction markets in general. The fundamental problem is that they are zero-sum. When one person makes a dollar, another loses it. This means that in order to have sophisticated traders who make money, you also need less informed participants who lose money. The challenge is therefore to attract enough 'losers' to make the market viable. Historically, political betting has been the main attraction, and the other categories haven't really taken off.
Decision markets, on the other hand, have an advantage: they are not zero-sum. In a decision market, the organisation financing the market can derive direct added value from it. For example, a company like IBM could pay the market to make decisions that optimise its operations, which better aligns the incentives between the parties.
Do you think we could apply the principles of prediction markets to traditional elections like in France between Emmanuel Macron and Marine Le Pen, to be able to make better decisions?
Yes, it could absolutely work, but you probably wouldn't want to use a traditional prediction market in that case. You would need something closer to a futarchy market or decision markets. The reason is that, as a voter, it's not just about who is most likely to win. It's more about what the results would be for France if one candidate won over another.
In the case of France, anyone could technically set up these markets. So you'd have one market predicting whether GDP would rise with Macron's victory, and another for Le Pen. Then people could bet on these markets, giving voters an insight into what the market thinks of each candidate. For example, if the market suggests that electing Macron would lead to economic decline, voters might choose the other candidate.
But in a DAO where decisions are binding - that is, the outcome of the market directly influences actions - you need a mechanism to prevent spam or malicious actors. Without such a system, someone could flood the DAO with thousands of proposals and exhaust its resources. To solve this problem, MetaDAO uses a model based on AMM (Automated Market Maker), which means that in order to create a proposal, liquidity must be provided in the accept and reject markets. This liquidity requirement helps filter out spam. For MetaDAO, this threshold is currently $100,000 of liquidity to create a proposal, but each DAO can define its own requirements.
Do you think prediction markets could play a role in improving democracy? Is it too risky, or do we need to refine the system first before considering it for democratic processes?
Honestly, I'm rather in favour of replacing democracy with a more efficient system such as futarchy, which could be used to create decision markets. My ultimate goal is to prove the superiority of this system. If it proves to be ten times more efficient, it could represent the next great revolution in the governance of our societies. Why cling to an obsolete model if we have a more efficient alternative? Of course, this system needs to be rigorously tested beforehand, and many aspects still need to be perfected.
But in any case, adding a financial stake to the vote makes the process more engaging for participants.
Yes, I think so. It's quite easy to demonstrate that it makes the vote more engaging. In traditional voting, there is only one action - submitting your ballot. For trading to be more engaging, all you need is for the average number of actions per participant to be greater than one. What we have found is that engagement rates on MetaDAO are quite high. On average, each trader makes more than 10 transactions worldwide. We had one proposal, which I mentioned earlier, where traders averaged over 30 trades each over a three-day period. That's pretty remarkable, especially considering that those trades were all manual - no bots. We observed more than 1,300 trades made by around 45 traders, each with around 30 manual trades. This even distracted some DeFi players on Solana from their usual routine at times.
Speaking of Solana, why did you choose to launch MetaDAO on the Solana ecosystem rather than Ethereum?
As I mentioned, I'm a markets enthusiast, and one thing you learn in this business - if you read thinkers like Howard Marks - is that you have to be both contrarian and right to get exceptional returns. For example, if you think Apple is a good buy, but everyone else already thinks so, then the price reflects that, and there's no real opportunity. But if everyone thinks Apple is a bad investment, and you see it as a good one, then there's a potential opportunity if you're right.
When I looked at Solana, it was clearly an upsetting pick. Many of my friends in the Ethereum space had already dismissed it, describing it as a "ghost blockchain" or even a "scam network". I remember one tweet where someone put it in the same category as Cardano. But as I dug into the technology behind Solana, I realised that it was more decentralised than people were saying, and that the ecosystem had a lot of potential. So I gambled that Solana was both vexing and fair, which meant there would be more opportunities for those building there.
And when do you plan to support other ecosystems like EVM-based ones, to go beyond Solana?
I won't decide - the market will. But my expectation is that MetaDAO should spend at least a year consolidating its presence in the Solana ecosystem, engaging all the DAOs there and refining the product. After that, we might consider expansion. Historically, multi-chain expansions haven't always been very successful, so if we decide to do it, we'll have to be very careful and make sure the core product is solid before we launch.
In an ideal world, how would MetaDAO work to help DAOs like Aave or MakerDAO make decisions based on prediction markets? Would it involve voting with the MetaDAO token, or is that different?
Actually, no, it would be with each DAO's native tokens. MetaDAO is just the platform provider. We already have three DAOs on Solana that use this model. For example, there's DeansList, a DAO that uses futarchy via MetaDAO. Recently, someone proposed a change to DeansList's tokenomics, and users placed bets on whether this proposal would benefit the value of the DEAN token. The market estimated that the DEAN token would be around twice as valuable if the proposal was adopted, and this proved to be correct, as evidenced by the spot price after the proposal was implemented. So, for MakerDAO, people would bet with Maker tokens, and for Aave, they would use AAVE.
What I think we need to understand is that DAOs are neither truly decentralised nor autonomous. They are often classic organisations where the team makes the decisions. I'm sure you've noticed this in DeFi too - yes, there's a vote, but it's rare to see the voters oppose the team's proposals, because the team has inside information and people trust it.
Markets can change this by offering permissionless prediction markets. For example, Polymarket is quite decentralised in that the price reflects the aggregated actions of everyone who wants to bet on the market. No central authority decides the price-it is simply determined by the market participants. This allows for a more autonomous approach, and even the automatic execution of certain decisions via smart contracts.
Markets are also often better aggregators of information than teams of specialists. There is plenty of evidence to show that markets outperform experts in their forecasts. For example, when the space shuttle Challenger exploded in 1986, it took the US government 4 months to identify the cause of the accident. But the stock market had already integrated this information in 16 minutes, with the fall in the share price of Morton Thiokol, the company responsible for the faulty part, much more pronounced than that of the other subcontractors. Markets quickly integrate the opinions of experts and informed players.
What improvements would you like to see for MetaDAO? What is your roadmap for the coming months?
Globally, MetaDAO is a two-sided platform: on one side, there are the organisations, and on the other, the traders. It's a bit like a stock exchange, where organisations list proposals and traders participate. Until now, our main challenge has been to attract more organisations. But recently we've made some progress. For example, Flash, a major project on Solana, has announced that it will use futarchy via MetaDAO for its decision-making. Fido, Solana's largest LST (liquid staking token), also has a proposal in the pipeline to use MetaDAO to manage its grants.
Now the focus is on attracting more traders. This involves improving the platform to make it easier to use and potentially through incentive programmes such as liquidity mining. We have recently added features to simplify trading decisions, helping users to place more informed bets. Our aim is to make the platform as user-friendly as possible and promote it through traditional marketing campaigns.
Just what do you plan to put in place to make the platform more accessible?
At the moment, if you go to the MetaDAO interface, you will see two main challenges. The first is to make transactions simpler, and the second is to help people understand how the system works. Surprisingly, these problems are distinct. You can simplify the process by focusing on what users need to do-for example, place bets without knowing all the technical details. But explaining the deeper workings of the system is more complicated.
Decision markets are a fundamentally new concept, particularly with their conditional markets that can be cancelled if certain conditions are not met. This is something that has not previously existed in a liquid, tradable form. We are constantly testing different ways of explaining this, such as the analogy of betting on IBM's share price under different scenarios. But making it all understandable remains a challenge.
What is MetaDAO's business model at present?
For the moment, MetaDAO operates as a marketplace where trades take place. At this stage, we don't take any commission, but it would be possible to activate a fee option in the future, taking a small percentage of transaction volume.
