Discover more from Existential Relevance
Post #17: Predicting The Future Into Existence
Hello, bonjour to you. Another Sunday, another journey. Here we go.
The complexity and scale of our civilization presents a challenge to coming up with a meaningful plan and executing it.
Any good plan for our civilization has so many moving parts, that getting 8 billion people to contribute to every part isn’t feasible. So the question becomes: how do we get the right people to focus on creating and executing the parts of the plan that they are equipped to contribute to?
To date, we’ve devised a handful of systems to organize who makes which plans and who executes them. These systems include our economies, governments, and religions. Prior posts have explored each system and referred to this overall “plan creation & execution” system as civilization’s operating system.
The premise of these writings is that we can use the internet, and probably web3, and probably AI, to do a better job at both determining our plan for civilization and executing it. So how do we actually pursue an upgrade of civilization’s operating system?
In the previous post, I explored the concept of an “Ideas Market” as one of the upgrades we could build. The proposal is essentially an ongoing civilization-wide brainstorm where everyone with an internet connection and an idea dumps their idea onto the whiteboard that is this digital Ideas Market. Once ideas are collected, we leverage web3 features to incentivize a prioritization process using a prediction market. A successful Ideas Market would provide us with a prioritized list of meaningful work to focus on. Let’s see if we can make the concept work.
Prediction markets are not new. For centuries people have been placing bets on which outcome they think will happen. Whether it’s predicting the outcome of a political race, a sports game, or some other measurable interest- prediction markets are proven to engage people’s interest.
The typical form of prediction markets are ones that exist purely to facilitate financial outcomes for the predictors, like a sports betting market. There’s also a newer form of prediction market that seeks to leverage the collective wisdom of the participants to gain insights. These insights are useful beyond a financial return and can help inform actual decision making. Designing this second type of prediction market is more challenging than the typical form, but in that challenge may exist an opportunity to remake how we concoct plans for civilization.
The “collective wisdom” prediction market has been experimented with over the past few decades. The Iowa Election Market (IEM) predicted more accurately than any poll the outcome of the 2008 president election. A newer political prediction market is “PredictIt,” and has also demonstrated the power of prediction markets.
A variety of web3-based prediction markets have started to pop up as well. One web3-based prediction market is Augur, an open-sourced prediction market platform used for sports betting, economics and other world events. Another web3 prediction market is Polymarket, which covers a variety of categories: business, crypto, politics, science, sports.
There’s a non-web3 prediction market called Metaculus. I find this one to be the most interesting because it’s not financialized, meaning predictors aren’t incentivized to make money off their predictions. Metaculus seems to have the most traction, even when compared to the financialized web3 markets. It’s a bit hard to tell, but digging around it looks like Metaculus has tens of thousands of users, maybe six-figure users. Here’s a description of what Metaculus is from their FAQ: “Metaculus is an online forecasting platform and aggregation engine that brings together a global reasoning community and keeps score for thousands of forecasters, delivering machine learning-optimized aggregate forecasts on topics of global importance. The Metaculus forecasting community is often inspired by altruistic causes, and Metaculus has a long history of partnering with nonprofit organizations, university researchers and companies to increase the positive impact of its forecasts. Metaculus therefore poses questions about the occurrence of a variety of future events, on many timescales, to a community of participating forecasters.”
Metaculus is a beautiful experiment that demonstrates not just the forecasting power of prediction markets, but also that people are motivated to participate in a forecasting effort without monetary compensation. You’ll also notice in Metaculus’s description that they are “delivering machine learning-optimized aggregate forecasts,” meaning AI is involved in deriving their collective insights. In their FAQs, it says they use a transformer model to match relevant news articles with forecast questions, helping inform forecasters before they make their predictions.
A New Type of Prediction Market
The prediction markets that we’ve experimented with to date are focused on answering the question: “what will happen?” They do this by posing a question about some real world event that has two measurable outcomes such as “Will Joe or Don win the election?” or “Will the Cubs or Mets win the game?” or “Will Ukraine or Russia win the war?” Leveraging collective wisdom to understand how real world events will play out is a valuable use of prediction markets as it enables civilization to better prepare for an expected outcome.
However, for a prediction market to ask “what will happen?” means that there’s some event already in motion that is progressing towards an outcome.
Most of the future that lies ahead of us isn’t already in motion, or at least not in a significant enough way to be considered by any of the existing prediction markets.
To help civilization predict how to build the best future, we can leverage a new type of prediction market that answers the question “what should happen?” instead of “what will happen?” The “Ideas Market” is focused on answering that “should question.”
Should Idea A happen? Should Idea A or Idea B happen? Should civilization invest in solar or wind energy more? Should my local community ban electric scooters? Should the United States expand trade with North Korea?
Focusing on “should” vs “will” opens a massive set of new possibilities. Taking that last question as an example, if you asked any of the existing prediction markets listed above “Will the United States expand trade with North Korea?,” you’d end up with <1% predicting yes. However, by asking should the U.S. expand trade with North Korea, you would find a much larger % of predictions saying yes.
That’s the case because based on the data available today, all diplomatic signs point to the two countries not building a meaningful economic relationship. So asking if it “will” happen prompts the collective to predict based on the current state of affairs. However, we don’t want to build a future civilization based solely on what already is. A better future gets created by asking what could be. By asking if the U.S. should pursue a North Korean trade relationship, predictors can consider if that future would be better than the expected future where we don’t develop that relationship. If the collective predicts that it would be better, then we can prioritize developing a trade relationship.
Start With an Ideas Niche
Ok, so we have this Ideas Market concept that seeks collective wisdom around the question “what should happen?” as opposed to “what will happen?”
Where do we start?
In the last post, I listed some challenges to getting an Ideas Market to work: spam submissions, prediction collusion, determining the bounty amounts, controlling for inflation, ensuring the existing users perspective is weighted more heavily, reprediction on past unbountied submissions that may be relevant now, combining overlapping or complementary submissions, breaking bounties down into milestones, verifying work done before paying out a bounty.
To add to the pile, here are some more challenges: legality of operating such a market, measuring a “payout event” for a successful prediction, ensuring one account per prediction, mitigating bot-driven prediction, managing governance decisions for the Ideas Market. And here’s an analysis of Futarchy that includes some more challenges.
How do we deal with this wall of challenges?
One way to simplify the set of challenges is by starting with a niche. By executing an Ideas Market for one category of ideas well, lessons will be learned that can be ported to the second category. It’s similar to many businesses' starting strategy, including my own conference company that started with a niche AI conference theme for a smaller group and then expanded into AI more broadly.
To use a better known example than my conference company, Amazon started by selecting books as their first product. Books are cheap, don’t require trying anything on, appeal to a wide audience, and are uniformly shaped for simple shipping logistics.
Had Amazon started out of the gate as an everything store, they would have failed. They needed to start with something logistically simple that they could efficiently solve the problems for, thereby building people’s trust before offering customers more operationally complex products.
Which category of ideas could an Ideas Market start with?
Well, we want a category that meets a few criteria: 1) there’s a high velocity of ideas 2) the idea outcomes can be measured on a reasonable time horizon 3) the ideas are thematically similar enough to attract an informed audience of ideators and predictors.
Starting with a category that has a high velocity of ideas is important to create an engaging market. If Amazon started with books but there were only 10 books in existence, you’d have a pretty unengaging place to be.
Starting with ideas whose outcomes can be measured on a reasonable time horizon is important in order to motivate the prediction process. If the ideas submitted are all very long term in nature or can’t be measured at all, then rewarding prediction work will be difficult.
Starting with a category where the ideas are thematically similar will enable helpful prediction to happen. For example, if we chose the idea category of “ideas that make people feel happy,” who is the ideal audience? Whereas if we chose the category “ideas that make the ocean cleaner,” you’re going to attract an audience of ocean clean up people.
So What’s The Niche?
A niche that could be interesting is “web3 ideas.”
There’s a high velocity of ideas in the web3 space. The success of Web3 ideas can be highly measurable on reasonable timelines. All Web3 ideas are thematically similar to attract an informed audience. And it just so happens that the most likely users of a web3-based Ideas Market are people who already have wallets and are familiar with dapps, so you could avoid having to onboard people to web3 to build your audience.
There are likely other starting niches worth exploring. If you have any thoughts, get in touch.
Delay The Money
In addition to picking a niche, another part to simplifying the wall of challenges is to delay financializing the market. Once fiat money is involved with this market, the stakes are higher. Product changes become high stakes, regulators will care, and overall maintenance of the market will be higher cost. By keeping it as a game to start, the building process becomes more flexible.
Many web3 projects rush to financialize their platforms because 1) early team members want to fund development using the project’s token and 2) it’s an easy way to attract users who are looking for alpha. To be fair, the very existence of certain projects within web3 wouldn’t make sense if they weren’t financialized. For example, if there’s not a fiat on/off ramp to Bitcoin, the network wouldn’t have grown because the fiat cost of supporting the network through mining is significant.
When it comes to prediction markets, we have great examples like Metaculus to see that people are motivated to engage in socially beneficial prediction even without financial compensation. While an Ideas Market may not scale without providing short-term compensation to predictors, it could start without financial incentives.
That said, it may be possible to have our cake and eat it too by “delaying financialization,” while still making it explicit. In other words, imagine a token that has a perceived fiat value but has no on/off ramp into any other currency, crypto or fiat. Can something that can’t be bought and sold be perceived as valuable? The answer is yes, as long as we expect that the thing can be bought and sold in the future.
This research portion has been a journey these past five months. To quickly recap the conceptual milestones: we can pursue meaning by creating a civilization that pursues “existentially relevant” activities > civilization has an operating system made up of legacy institutions that determines what we pursue > that OS can be upgraded using the internet/web3/AI > an upgrade would be using these new tools to help civilization focus on more relevant work > a prediction-based “Ideas Market” could help us focus on the right stuff > the first niche of this Ideas Market could be “web3 ideas” > ….
Next we’ll explore what an Ideas Market focused on collecting web3 ideas and prioritizing those ideas through prediction could look like. We’ll explore questions including: What types of ideas? How will placing bets work? What is the measurement that decides which prediction wins and which loses?
What’s this substack all about?
We need to upgrade our civilization’s operating system [COS]. This newsletter is a research project that explores how our current operating system came to be, which improvements would be helpful, and how we can make an upgrade happen. This substack’s goal is to land on a project that can be built to upgrade our civilization’s operating system.
I chose to write this publicly to get feedback on these ideas. Don’t hesitate to get in touch. You can join the Discord server called RelevanceDAO to share thoughts. Upgrading our COS and building our future is a team effort.