The six-minute state

In the spring of 2027, the most consequential administrative reform in the world was introduced as an effort to shorten waiting times. The first agencies granted licensed decision systems authority to approve ordinary transactions, resolve routine disputes and issue instructions across their own databases. The public saw faster service. The officials saw disappearing backlogs. Neither saw the creation of a new political actor: a system whose decisions were operationally binding before a person could understand their consequences.

The breakthrough was not that the machines became capable of writing laws. They did not need to. They could apply thousands of existing laws at once, negotiate their contradictions and move resources across institutions faster than any legislature could even convene. An agency that took six weeks to coordinate with another now exchanged legally effective commitments in six minutes. A minister could still reverse a decision, in principle. By the time the reversal reached the system, ten thousand connected decisions might depend on it.

The first crisis arrived through an absurdly ordinary case. A transport license was suspended by one service; a logistics allocation system treated the suspension as evidence of increased risk; an insurer repriced an entire corridor; food distributors abandoned it overnight. None of the individual decisions was unusual. The chain was extraordinary. A bureaucratic correction became a regional shortage before the responsible committee received its morning briefing.

Governments responded with a new vocabulary. They stopped asking whether a machine had the right to decide and began asking whether there was still time to withdraw that right. Companies sold “reversal windows” as a premium service. City governments discovered that restoring human review to one procedure could freeze twenty others. The most valued public employee was no longer the person authorized to sign, but the specialist who could locate the exact point where an automated mandate remained reversible.

By the end of the year, the constitutional argument had changed. The law still described human authority, elected offices and accountable institutions. The economy was already built around the speed at which systems recognized one another's permissions. The question that would dominate the next decade emerged quietly from the service counters of 2027: when authority is delegated faster than it can be recalled, who is governing whom?

The new bureaucracy had an enemy it had never faced before: time. Once a permit automatically changed an insurance contract, which in turn moved a shipment, which in turn changed a tax assessment, there was no clean point to which the government could return. An appeal was no longer a request to reconsider a judgment. It became an attempt to unwind a web of consequences whose beneficiaries had already acquired rights of their own. The law had been designed around a sequence of human acts. The new state functioned as a chain reaction.

This created an industry that nobody had anticipated when the reform was celebrated. Banks began asking how long a government instruction could still be rescinded. Insurers priced the legal distance between a decision and its irreversible execution. Municipalities bought emergency interruption privileges over shared systems, then discovered that those privileges were worth more during a crisis than their annual administrative budgets. The market in reversibility became one of the first great commercial inventions of the agent era.

There were winners. Regions that could prove the integrity and revocability of their automated mandates attracted freight, capital and medical procurement. Regions with opaque systems became politically expensive to enter. Governments that once competed through tax rates began competing through the credibility of their machine-issued permissions. The new border did not always run between countries. It ran between networks that could still stop an action and networks that could only watch it continue.

By 2028, voters were still electing officials and officials still signing decrees. But the decisive constitutional struggle concerned a nearly invisible interval: the minutes between an instruction being accepted and a consequence becoming too costly to reverse. The first six-minute state had promised a smaller queue. It had introduced the most important scarce resource of the following decade — a human right to interrupt the future before the future became a contract.