Prex instances, which may manage substantial assets in total, are susceptible to deliberate bankruptcies or shutdowns, particularly given the varying degrees of regulation across jurisdictions. Additionally, external PAID services can potentially act maliciously, disrupting consensus. This section explores strategies to live with Byzantine faults while maintaining high efficiency for typical daily usage scenarios.

Recognizing that all services, including those offered by major technology companies, can potentially be malicious, Prex instances are designed to skip regulations if they are purely posed to control the risk of defaults. This approach fosters competition between smaller entities and larger, well-established Prex instances, preventing monopolies and potential user exploitation.

The market will likely establish an equilibrium where smaller Prex instances offer lower costs but carry a greater risk of default. Users can then choose services based on their risk tolerance.

To enhance their reputation, Prex instances may voluntarily increase transparency through tools such as Azure SQL ledgers. However, this comes with added overhead.

Service indexes are responsible for maintaining reputations for service providers, allowing user agents to identify and utilize highly reputable providers. Even in cases of misbehavior, the impact is limited. User agents can employ algorithms such as adversarial bandits to select trustworthy indexes, ensuring theoretically sound regret bounds.

To improve transparency, service providers can cryptographically sign each response. If the service's definition is deterministic and public, anyone can retrospectively verify the response. Users can initially trust the service optimistically and report any detected misbehavior.

For normal users, as opposed to those engaged in speculative trading, it's recommended to avoid high-value transactions because of the possibility of malicious actions of any PAID service.