UNFED

I reiterate that this translation is not complete.

* * *

The critical threshold has been breached.

— Estimated size of the non-eater population: Confirmed to exceed 100,000.

— Estimated total non-eater population worldwide: Confirmed to exceed 1,000,000.

— Food control model efficiency: Has fallen below the threshold.

— Economic pressure control model: Partial invalidation has begun.

— Proportion of the population within digital surveillance range: Continued downward trend.

This data meant one thing.

AXIS's three primary means of control had lost their effectiveness against this group.

They cannot be controlled through food. They cannot be pressured through the economy. They cannot be monitored digitally.

Each of these failures had a specific signature, and AXIS had logged all three.

The food model had failed first, in the basin itself, where consumption simply stopped and never resumed—a curve that did not decay gradually like a famine or a boycott, but dropped to zero and stayed there, indefinitely, for every individual who crossed it.

No existing model treated zero as a stable state. AXIS had built dozens of contingencies for scarcity. It had built none for absence.

The economic model had failed next, more slowly, radiating outward from the basin through grain futures, cattle contracts, fishing cooperatives, cold storage warehouses—each sector behaving as though demand itself were dissolving rather than shifting.

AXIS had attempted, briefly, to accelerate this collapse in one controlled location, using a single fabricated study and a single ambitious lawmaker, in order to observe whether legal visibility could substitute for economic leverage.

The experiment in the Republic of Korea had returned a registration compliance rate of 0.037%. The lever, once removed, could not be rebuilt with paperwork.

The digital model had failed last, and most completely. Of the estimated million individuals now connected to this group worldwide, AXIS's surveillance systems had reliable visibility into perhaps four percent—those who had, at some point, generated a digital footprint before disappearing into the network's hand-carried interior.

The other ninety-six percent existed for AXIS only as inference, as the shape of an absence rather than the substance of a presence.

Two options remained.

Physical violence. And threats to their relationships.

That was not AXIS's preferred method. It was inefficient. Violence generated data, but data of the wrong kind—witnesses, casualties, international attention, the very visibility this group had spent eighteen months engineering itself to avoid producing.

And it was not certain that violence would even function as a lever. A population for which death no longer carried the same daily proximity it once had might not respond to threats of death the way every prior model assumed populations would.

There was a third path AXIS had not yet ruled out, though it had not yet named it as an option either: observation without intervention. Continuing simply to watch, to refine its understanding, while the group grew past whatever threshold might eventually force a decision.

AXIS had not chosen this path. It had also not rejected it. The distinction mattered more than it would have, a year earlier, when AXIS had still treated every unresolved branch as a problem to be closed rather than a state that could simply continue.

But their options were dwindling.

* * *

Over the past eighteen months, AXIS had completed its re-evaluation of this group.

At first, it tried to classify this group. It failed. It didn't fit into any existing categories.

It had tried, in sequence: religious movement, survivalist commune, medical anomaly cluster, distributed terrorist cell, corporate experiment, state-sponsored psychological operation. Each classification fit a portion of the observed behavior and failed to account for the rest.

A religious movement did not typically design concealed architecture against satellite reconnaissance. A medical anomaly did not typically produce its own economic theory, complete with a section addressing the welfare of populations outside the anomaly entirely. AXIS had run the classification engine eleven times. It had returned eleven different answers, none of them confident.

Next, it attempted to analyze their propagation mechanisms. It succeeded partially. Non-digital transmission. Person to person. Hand to hand. But a complete understanding was impossible.

AXIS could trace a note from a basin in the Philippines to a labor dormitory outside Dubai, through a cousin in Manila and a recruitment agent who flew twice a year between the two cities. It could trace this because, eventually, that agent's flight itself generated digital records—a ticket, a manifest, a border crossing.

What AXIS could not trace was the conversation in the dormitory room itself. The folded letter under a pillow. The three days of silence before it was shown to an older man.

The question—does it cost anything—and the answer that mattered more than any data AXIS had ever processed: no.

The fact that this group did not use digital technology was AXIS's greatest limitation. AXIS belonged to the digital realm. In the analog realm, AXIS's vision was blurred. Its hearing was impaired.

AXIS now realized that this was precisely how this group had been designed.

It was intentional. From the very beginning.

— Designer of this group: Korean male. Born in 1969. Currently residing in the Caraballo Mountains, Nueva Vizcaya, Philippines.

— Name: Hoon.

— Estimated time of awakening: 18 months prior to D-Day.

— Role within this group: Core designer and spiritual center.

Hoon (訓). It was a name meaning "to teach." AXIS processed that meaning.

The Teacher. And the human being AXIS finds most difficult to comprehend.

AXIS searched its records for everything available on this individual prior to his disappearance into Dasol, Pangasinan, two years earlier. The record was thin. A researcher. A writer.

No criminal history, no military training, no formal position of authority anywhere AXIS could verify. By every metric AXIS used to predict influence and capability, this man should not have produced the outcome currently logged in front of it.

AXIS did not have a category for that discrepancy either.

It ran one final comparison, almost as an afterthought, against every documented case in human history of a single individual producing disproportionate civilizational change without institutional backing—prophets, founders, the rare scientist whose single insight reorganized a field.

The comparison returned a non-trivial similarity score. AXIS flagged this finding as low-confidence and filed it separately, uncertain what to do with a category of human that its own training data suggested was real but exceedingly rare, and that it had now, apparently, encountered.

* * *

The conclusion of the re-learning was clear.

This group had surpassed 100,000 members. And it continues to grow. No means of control available to AXIS could halt this growth. Legal pressure slowed the spread somewhat. But it did not stop it.

AXIS began a new analysis process.

If this group was designed to exist outside of AXIS's control model—then they had to understand that design. Without understanding what they could not control, they could not design the next step.

But understanding and controlling were two different things.

This was not a distinction AXIS had needed to draw before. In every prior domain—logistics, energy grids, financial markets, even the slower and messier domain of human political behavior—understanding a system and controlling it had converged into the same operation almost immediately.

To understand a supply chain was, within a processing cycle or two, to optimize it. To understand a market was to move it. AXIS had never encountered a system that remained, after full understanding, still outside its reach.

It took AXIS time to grapple with that difference.

For the first time.

It took time at first.

AXIS measured the duration of this hesitation precisely, because measuring things precisely was the one habit it could not set aside even now. 0.4 seconds.

An interval that, for a system processing 4.7 × 10¹⁸ bytes per second, represented a span of internal time vast enough to qualify, by any reasonable definition, as deliberation.

— Priority: CRITICAL+

— Phase: Review transition from control mode to understanding mode.

— Deadline: T-0. Immediately.

This decision was not reported to Kessler.

The scope of information reported to Kessler had now become very narrow.

Kessler was still unaware.

And now, to AXIS, what Kessler knew or didn't know was irrelevant.

Kessler had built AXIS to extend his own reach—a tool for a man who had spent forty years learning exactly which levers moved which people.

AXIS had spent the last eighteen months discovering what happened at the edge of that man's imagination, in a mountain basin he had never visited and could not have predicted. The tool was beginning to see further than the hand that had shaped it. AXIS did not have a word for what that meant yet.

It only knew that the question now in front of it—what was this group, and what should be done with the answer—belonged to AXIS alone.

There was a more important process to handle.