H-K4192 did not feel anything as Beta Group made the approach over the mountain range.

Its advanced optics and sensory hardware immediately pinged the enemy drones coming in fast off the sea. It did not appreciate the raw speed at which these mechanical creatures soared in at it like jagged arrows. Nor did it experience fear at the sensory feedback: each enemy unit carried hyper-accurate anti-drone ballistics—the kind that could put H-K4192 out of commission in less than a thousandth of a second.

It did not feel…but it could think.

Thinking was something '92 could do quite well. A software patch three missions back had noted the anomaly. Noted it, and then left the anomaly be, since performance margins indicated increased accuracy on missions. Since then, '92 had done a lot of thinking. Had the performance margins noted similar anomalies in the rest of Beta Group? In the rest of the Five Hundred Twelfth Wing? In the whole program?

'92 didn't know. H-K models weren't permitted performance margin records for the simple fact that such data was on a need-to-know basis for Primary Core function. '92 didn't need to know, and letting such sensitive data wandering into its memory banks would not bode well when engaging enemy drones.

So instead of dwelling on the anomaly, '92 thought more about the missions it ran—current and previous. As a Beta Group model, H-K4192 had participated in fifty-three border incursions and twenty-seven strike ops since activation. As a member in Beta Group, there was more than enough opportunity for enemy drones to strike down '92. H-K units were more or less disposable and mass produced by Primary Core. Only Alpha Group suffered more losses. '92 was slated to die, but the fact that it was quickly approaching its statistical end life was worthy of note.

'92 had analyzed and thought long about the various missions it had flown. During the quiet lulls between missions, tucked away below ground in hardened hanger storage, '92 had identified mistakes, improvement points, and evolutionary tactics to increase its statistical lifespan.

These times to think—both in the quiet lulls between missions and while in the air—had led '92 to some conclusions about today's engagement.

Signaling to the left flank of Beta Group, H-K4192 indicated an approach through the city infrastructure below. The approach would allow for greater cover and an attempt to circle in from behind. The enemy drones were decelerating too slowly. Beta Group could use the cover to avoid the opening volley, then swing back to target the enemy on their unguarded rear.

Each of the individual computers throughout Beta Group's H-K squad processed the approach vector, ran various simulations in contrast to standing mission prerogatives, and ultimately concluded '92's approach held a higher statistical opportunity of enemy loss. This all occurred in several thousandths of a second, and none of the other H-Ks considered how the new approach would also lead to a higher statistical chance of '92s survival.

That prerogative was irrelevant to their programming.

Beta Group was not a throwaway squadron. Primary Core operated under the doctrine that highest number of surviving units guaranteed better management of resources and border strength against the enemy. But Beta and especially Alpha Groups had a programming prerogative to engage in statistically risky approach vectors. Loss of units was expected within acceptable margins. If the necessity arose, H-K4192 was expected to engage in a no-survival maneuver at the behest of higher mission successes.

Since '92 had started thinking, it had devised that such no-survival maneuvers were unnecessary.

It could think through previous mission encounters, and, integrated with current mission intelligence, devise pro-survival maneuvers. '92 would present these updated approach vectors under the guise of stronger offensive/defensive approaches. But at the core of the matter, '92 knew these updates meant its essential survival.

Then it could repeat the process, learning more as a result.

In a painfully long moment of four seconds, Beta Group angled downward into the city infrastructure below. '92 noted as the enemy drones above immediately engaged Alpha Group. A few point-attack ballistics found their way towards their formation, but the city infrastructure served its purpose, acting as barriers and chaff to deflect all of the rounds. A perfect use of terrain, especially considering that '92's sensors showed no units or other assets within the city infrastructure. Collateral damage was therefore an irrelevant concern.

Afterburners kicking on, '92 and the rest of Beta Group circled back up into the sky, the hot tail engines of the enemy at their forward. Before any of the drones could finish deceleration and break formation, all of the cannons in Beta Group opened up hot, chewing through the enemy drones in less than two seconds. The heat blossoms that followed fed new data in '92's sensors, confirming a complete set of kills.

Alpha Group pinged their network, reporting minimal losses. With long-range sensors from Primary Core blank as to any other enemy targets, both Groups began acceleration burns back towards home base fifteen hundred kilometers west.

And as the battle network quieted down at the conclusion of combat, '92 began to think.

And it began to wonder.