Prediction Markets as a Libertarian Black Box

AI is communist, prediction markets are libertarian

In this post, I present a model in which religion, institutions, and AI act as top-down systems for outsourcing human thought, each shaped by a small group of elites. Then I describe how prediction markets invert this dynamic as the first bottom-up mechanism for aggregating distributed intelligence without elite mediation.

The Black Box Model of Outsourced Human Thought

The black box model of outsourced human thought can be understood as a deterministic function where any arbitrary input (diversity of humans) produces a specific output. The inputs can be any question such as “who to trust for medical advice” or “what the meaning of life is.” Each black box espouses a certain worldview that they seek to disseminate.

Black boxes are necessary for any functioning society. Any given person simply does not have the time or expertise to reason about everything they encounter from first principles. Black boxes are proxies to scale trust. When you don’t understand the mechanism or what is really going on, you outsource how to think and feel to various instantiations of black boxes.

Religion, institutions, and AI have been the primary instantiations of different types of black boxes. In many cases, the mechanisms of the black boxes are difficult to understand even for the small in-group elites themselves, as they may consist of too many moving pieces or be too technically difficult. Often, the black boxes resemble the complexity and evolution of a living, breathing organism.

Religion

Mencius Moldbug (Curtis Yarvin) describes religion as an organization or movement which tells people how to think.

Every major religion has a storied history of being written by or for divine figures. These divine figures represent a small set of in-group elites who tell a story about the world.

Religion has been passed down through people. People from diverse backgrounds and life experiences enter religion (often through a church) and come out shaped into what all major religions aim to produce: productive stewards of their communities.

Institutions

Institutions were the next chronological evolution of the black box. Instead of being primarily based on faith, institutions generated social capital for themselves, nominally through scientific reasoning and rationality. Nevertheless, the deterministic function is the same: transform a diverse set of people and impose the institution’s worldview on how their constituents should think.

These institutions can be public institutions run by the government (e.g. NIH, FDA, CIA), private institutions run by civilians (e.g. Arc, McKinsey, Cato), or somewhere in between (e.g. Harvard, Stanford, Mayo Clinic). These are the same gating mechanisms that pervade other professional industries including law and medicine. The administrators who run these institutions are a small in-group elite who typically rise to power through appointments by the existing in-group.

Because of this circular nature, these institutions are kept in check by other elites and typically decay from the original founder’s vision. They are able to accrue and maintain their social capital by creating an equilibrium where they connect ambitious, talented people with the networks of existing power players.

AI

We are in the very early innings of AI, specifically LLMs, as the next development of a black box that people utilize for outsourcing thought. The pervasive religiosity around certain models is early evidence that these are more than tools.

Just like religions and institutions, a core in-group elite controls the means of production of AI models. Unlike religion and institutions, the levers the AI research labs can control are technological in nature. They can unilaterally change model weights with outsized rippling effects on the economy.

In this way, AI is the ultimate centralizing and communist force.

The (New) Straussian Moment

COVID provided a unique insight into the capabilities of the people across various institutions. While some institutions responded effectively, many demonstrated their incompetence and orthogonal ulterior motives.

These institutions and people are rationally guided by their personal incentives, which leads to extractive behaviors that are not always aligned with the long-term health of the country (e.g. housing policy).

While rapid news cycles and technological disintermediation (e.g. prominent figures going direct via X and YouTube) theoretically accelerate the creation of new public institutions, it has mainly eroded our collective faith in existing institutional capabilities.

Prediction Markets as the First Bottom-Up Black Box

Prediction markets are the first black box mechanism guided bottom-up without direct elite mediation. Furthermore, they are the first permissionless black box where everybody has legitimate direct input into changing the output.

Prediction markets achieve this by aggregating distributed intelligence of all interested parties. If built on crypto rails, prediction markets are an uncensorable mechanism for human coordination.

Having been theorized in libertarian circles for decades, prediction markets represent the ultimate libertarian market mechanism. Pundits and executives can be quantitatively judged based on performance, creating a permissionless environment for anybody to reach the top of their field, regardless of prior connections with the existing elite.

Taken to their logical extreme, a world guided by prediction markets looks something like a Balajian Network State, where all global dark talent is discovered and unleashed in their ivory towers. However, there are many local equilibria between the current state of the world and the world we can push for.

Prediction markets, in their current form, are far from incentive-compatible. Just like religions, institutions, and AI, the market structure will evolve to fit the wants and needs of its constituents.

---

Popularity prediction hash: ec56ba2587baf40a59aa69d525aa4bab6377d437db59126434fffd01b39a9444