Original title: From prediction markets to info finance
Original author: Vitalik Buterin
Original translation: 0xjs, Golden Finance
One of the Ethereum applications that excites me the most is prediction markets. In 2014, I wrote an article about futarchy, a governance model based on predictions conceived by Robin Hanson. As early as 2015, I was an active user and supporter of Augur (look, my name is in the Wikipedia article). I made $58,000 on bets in the 2020 election. This year, I have been a close supporter and follower of Polymarket.
For many people, prediction markets are about betting on elections, and betting on elections is gambling—it's great if it allows people to have fun, but fundamentally, it's no more interesting than buying random tokens on pump.fun. From this perspective, my interest in prediction markets seems puzzling. Thus, in this article, I aim to explain why this concept excites me. In short, I believe (i) even existing prediction markets are a very useful tool for the world, but additionally (ii) prediction markets are just one example of a larger and very powerful category that has the potential to create better implementations in social media, science, news, governance, and other fields. I will refer to this category as "information finance."
The dual nature of Polymarket: a betting site for participants, a news site for everyone else
In the past week, Polymarket has been a very effective source of information regarding the US elections. Polymarket not only predicted a 60/40 chance of Trump winning (while other sources predicted 50/50, which in itself is not very impressive) but also showcased other advantages: when the results came out, despite many experts and news sources trying to entice the audience with favorable news for Harris, Polymarket directly revealed the truth: the chance of Trump winning exceeded 95%, while the chance of taking control of all government departments exceeded 90%.
Both screenshots were taken at 3:40 AM EST on November 6.
But to me, this is not even the best example of what makes Polymarket interesting. So let's look at another example: the elections in Venezuela in July. The day after the election, I remember catching a glimpse of people protesting the highly manipulated election results in Venezuela. At first, I didn't pay much attention. I knew Maduro was one of those "basically dictators," so I thought, of course, he would rig every election result to preserve his power, and of course, there would be protests, and of course, the protests would fail—unfortunately, many others did fail. But later, while scrolling through Polymarket, I saw this:
People are willing to put over a hundred thousand dollars on the bet that Maduro will be overthrown in this election with a probability of 23%. Now I start to pay attention.
Of course, we know the unfortunate outcome of this situation. Ultimately, Maduro did continue in power. However, the market made me realize that this time, the attempt to overthrow Maduro was serious. The scale of the protests was massive, and the opposition executed an unexpectedly well-performing strategy to demonstrate how fraudulent the elections were to the world. If I hadn’t received the initial signal from Polymarket "this time, something is worth paying attention to," I wouldn’t have even started to pay attention.
You should never fully trust the Polymarket betting charts: if everyone believes the betting charts, then any wealthy person can manipulate the betting charts, and no one would dare to bet against them. On the other hand, completely trusting the news is also a bad idea. The news has sensational motives and exaggerates the consequences of anything for clicks. Sometimes, this is reasonable, and sometimes it is not. If you see a sensational article but then go to the market and find that the probability of related events hasn’t changed at all, skepticism is warranted. Or, if you see unexpectedly high or low probabilities in the market or sudden unexpected changes, that’s a signal for you to read the news and see what caused it. Conclusion: by reading the news and the betting charts, you can gain more information than by reading either one alone.
Let's review what's happening here. If you are a gambler, then you can bet on Polymarket, which is a betting site for you. If you are not a gambler, then you can read the betting charts, which is a news site for you. You should never fully trust the betting charts, but I personally have incorporated reading the betting charts as a step in my information collection workflow (alongside traditional media and social media), which helps me gather more information more effectively.
Information finance in a broader sense
Now we enter an important part: predicting election outcomes is just the first application. The broader concept is that you can use finance as a way to coordinate incentive mechanisms to provide valuable information to the audience. Now, a natural reaction is: isn't all finance fundamentally related to information? Different participants make different buying and selling decisions because they have different views on what will happen in the future (beyond personal needs like risk preference and hedging desires), and you can infer a lot of knowledge about the world by reading market prices.
To me, information finance is like this, but structurally correct. Similar to the concept of being structurally correct in software engineering, information finance is a discipline that requires you to (i) start with the facts you want to know, then (ii) deliberately design a market to best extract that information from market participants.
Information finance is a three-sided market: bettors make predictions, readers read predictions. The market outputs predictions about the future as a public good (because that’s its designed purpose).
Prediction markets are an example: you want to know a specific fact about what will happen in the future, so you set up a market for people to bet on that fact. Another example is decision markets: you want to know which decision, A or B, will yield a better outcome based on some metric M. To achieve this, you establish a conditional market: you ask people to bet (i) which decision will be chosen, (ii) if decision A is chosen, what the value of M will be, otherwise zero, (iii) if decision B is chosen, what the value of M will be, otherwise zero. With these three variables, you can determine whether the market believes decision A or decision B is more favorable for obtaining the value of M.
I expect that one technology driving the development of information finance in the next decade will be AI (whether large models or future technologies). This is because many of the most interesting applications of information finance are related to "micro" issues: millions of small markets where decisions have relatively small impacts when viewed individually. In reality, low-volume markets often cannot operate effectively: for experienced participants, it doesn't make sense to spend time doing detailed analysis just to earn a few hundred dollars in profit, and many believe such markets cannot operate at all without subsidies, as there aren't enough naïve traders to allow experienced traders to profit from anything other than the largest and most sensational problems. AI completely changes this equation, meaning that even in markets with a transaction volume of $10, we can potentially obtain quite high-quality information. Even if subsidies are needed, the amount of subsidies per issue becomes very affordable.
Information finance requires human distillation
Judgment
Suppose you have a trusted human judgment mechanism, and that mechanism has the legitimacy of being trusted by the whole community, but making judgments takes a long time and is costly. However, you want to access at least one approximate copy of that "expensive mechanism" in real-time at a low cost. Here are some ideas from Robin Hanson on what you can do: each time you need to make a decision, you set up a prediction market predicting what result that expensive mechanism would yield if called. You let the prediction market run and invest a small amount of money to subsidize the market makers.
99.99% of the time, you are not actually invoking the expensive mechanism: perhaps you would "revert the transaction" and return everyone's investment, or you simply give everyone zero, or you check whether the average price is closer to 0 or 1 and treat that as the basic fact. 0.01% of the time—potentially random, possibly targeting the highest volume markets, or a combination of both—you actually run the expensive mechanism and compensate participants accordingly.
This gives you a credible, neutral, quick, and cheap "distilled version," which is your originally highly credible but extremely costly mechanism (using the term "distilled" analogous to "distilled" in LLMs). Over time, this distilled mechanism roughly reflects the behavior of the original mechanism—because only participants who help achieve that outcome can profit, while others will lose money.
Possible model of combined prediction markets + community notes.
This applies not only to social media but also to DAOs. One major problem with DAOs is that there are too many decisions, and most people are unwilling to participate, leading to either widespread use of delegation with the risks of centralization and principal-agent failure common in representative democracy, or being easy to attack. If actual voting in a DAO rarely occurs and most things are decided by prediction markets predicting the voting outcomes with a combination of humans and AI, then such a DAO might run well.
As we see in the example of decision markets, information finance contains many potential pathways to solve important issues in decentralized governance, with the key being the balance between market and non-market: the market is the "engine," while some other non-financial trust mechanisms are the "steering wheel."
Other use cases for information finance
Personal tokens—many projects like Bitclout (now deso), friend.tech, etc., create tokens for everyone and make them easy to speculate on—are a category I call "original information finance." They deliberately create market prices for specific variables (i.e., expectations of an individual's future reputation), but the exact information revealed by the prices is too ambiguous and subject to reflexivity and bubble dynamics. There is potential for improved versions of such protocols, which could address important issues like talent discovery by more carefully considering the economic design of the tokens (especially where their ultimate value comes from). Robin Hanson's idea of reputation futures is a possible end state here.
Advertising—the ultimate "expensive but trustworthy signal" is whether you will buy a product. Information finance based on that signal can help people determine what to buy.
Scientific peer review—there has long been a "reproducibility crisis" in the scientific community, where certain famous results have become part of common wisdom in some contexts but ultimately cannot be reproduced in new research. We can attempt to determine which results need to be re-examined through prediction markets. Before re-examination, such markets will also allow readers to quickly estimate how much they should trust any particular result. Experiments of this idea have been conducted and seem to have succeeded so far.
Public goods funding—one of the main problems with the public goods funding mechanisms used by Ethereum is its "popularity contest" nature. Each contributor needs to conduct their own marketing campaign on social media to gain recognition, making it difficult for those who cannot do so or who naturally have more "background" roles to secure significant funding. An attractive solution is to attempt to track the entire dependency graph: for each positive outcome, which projects contributed how much, and then for each project, which projects contributed how much, and so on. The main challenge of this design is figuring out the weights of the edges so that it can resist manipulation. After all, such manipulation has been occurring. A distilled human judgment mechanism might be helpful.
Conclusion
These ideas have been theorized for a long time: the earliest works on prediction markets and even decision markets date back decades, while similar discussions in financial theory are even older. However, I believe the current decade provides a unique opportunity for the following main reasons:
Information finance addresses the trust issues that actually exist for people. A common concern of this era is the lack of knowledge (worse still, a lack of consensus) about whom to trust in political, scientific, and business environments. Information finance applications can help be part of the solution.
We now have scalable blockchains as a foundation. Until recently, fees were too high to really realize these ideas. Now, they are no longer too high.
AI as participants. When information finance must rely on human participation for every issue, it is relatively difficult to function. AI greatly improves this situation, allowing for effective markets even on small-scale issues. Many markets may feature a combination of AI and human participants, especially when the number of specific issues suddenly increases from small to large.
To fully seize this opportunity, we should go beyond just predicting elections and explore what else information finance can bring us.
Special thanks to Robin Hanson and Alex Tabarrok for their feedback and comments.
Original link
