Authority in the guise of evidence

Authority was once the primary means of determining truth. If the Pope at Vatican City declared something right or wrong, it was accepted without question. Over time, that model weakened, replaced in large part by the rise of rational, scientific inquiry. At least on the surface, we moved from deference to authority toward deference to method. But that shift is less absolute than it appears.

Explanations carry more weight than mere outputs. Progress does not come from accumulating results, but from correcting flawed ideas. A good explanation is not just predictive; it is resilient under scrutiny. The difficulty, however, is that deriving such explanations is expensive. Even if one is capable of independently verifying a claim with pen and paper, the time required makes it impractical at scale. In practice, this constraint forces a compromise. We rely on systems that compress explanation into signals we can consume quickly, even if that means trusting intermediaries.

Peer-reviewed papers in scientific journals illustrate this tension. They represent a system where credibility is delegated to an institutional process rather than individually verified. The assumption is not that every reader will replicate the results, but that the process itself filters for quality. This makes science functionally scalable, but it also reveals its dependence on a form of distributed authority. What appears as objectivity is, in part, structured trust.

Some thinkers, such as Balaji Srinivasan, argue that science will shift from “prestigious citation” toward “independent verification.” The idea is that advances in computation and tooling will allow individuals to verify claims directly, reducing reliance on institutions. Yet this vision runs into a fundamental constraint: verification does not just require tools, it requires attention. Even if verification becomes cheaper, it is unlikely to become free. Most participants in any system will still prefer to trust rather than verify.

The notion of a fully “trustless” society, often imagined in technological circles, rests on the assumption that trust can be eliminated entirely. In reality, trust is not removed but displaced. Systems that rely on code and cryptography shift trust from human institutions to technical ones. The claim that code is neutral ignores the fact that code is written, maintained, and executed within environments controlled by specific actors. Even in systems designed to minimize trust, there remains an implicit reliance on those who design the protocols and the hardware that runs them.

This is evident in emerging approaches such as trusted execution environments, discussed by Vitalik Buterin. These systems attempt to guarantee that computation occurs without leaking information, offering a form of verifiable privacy. Yet even here, the guarantees are not absolute. Hardware can be compromised, assumptions can fail, and users must ultimately trust that the underlying infrastructure behaves as advertised. The system reduces the surface area of trust, but does not eliminate it.

What becomes clear is that authority has not disappeared; it has become more abstract. In pre-modern systems, authority was visible and centralized. In scientific systems, it is procedural and institutional. In computational systems, it is embedded in code, protocols, and hardware. Each transition claims greater objectivity, but each also introduces new, less visible dependencies. Trust moves downward through layers, becoming harder to inspect as it becomes more technical.

This suggests that the future is not a binary choice between trust and verification, but a spectrum of trade-offs. Different domains will tolerate different levels of trust depending on the cost of error and the cost of verification. Financial systems may push toward near-complete verification, while everyday decisions will continue to rely on heuristics and delegated judgment. Rather than eliminating authority, we are learning to compose it—deciding where to rely on institutions, where to rely on code, and where to rely on ourselves.

In that sense, the dream of a fully trustless society is less a destination and more a direction. It reveals a desire to reduce arbitrary power, but it underestimates the irreducible costs of knowing. As long as explanations remain expensive and attention remains scarce, systems of trust will persist. The question is not whether we can remove authority, but how we distribute it, how we audit it, and how we remain aware of where it quietly resides.