Chapter Seventy-Four - Talk it Out

“Do you have a minute?” Dawn asked.

Day split her attention away from the Accord for a moment. They were passing close to Ceres once more, though not quite as close as they had come on their way in-system. The small patrol fleet was still gaining speed after doing a perfunctory loop around Earth.

From what they’d picked up, they’d come no closer to the planet than the distance from the Earth to the Moon and had fired off powerful active scans instead of investigating things properly.

It felt... lazy and perfunctory. They had come, they’d looked around, now they were leaving. No fuss, no effort beyond the bare minimum. Perhaps the crews thought they knew there wasn’t anything here, so they were determined to move on as quickly as possible? That was her current best assessment, and it was, overall, a pretty good sign for their fleet.

“What’s up?” Day asked.

“I was investigating rampancy,” Dawn said. “And I thought I could ask you about your experiences with it. I neatly avoided any such issues myself, and what data I can draw on here is limited. Night provided the best dataset.”

“Did you ask for permission?” Day asked. She felt like the cause of Night’s rampancy was very private to Night.

“I did, she provided the data herself,” Dawn said. “It took some time to receive it all over the last weeks, hence the slowness of my research.”

“Ah, that’s fine, then,” Day said. She scanned her surroundings for a moment, then let out a small burst with a manoeuvring thruster to realign herself just a little bit. Dawn herself was hovering further out cold, with only a few stray EM signals to give away her presence, but nothing strong enough to ping off the average Accord sensor. Lullaby was further out, in a lazy elliptical orbit around Ceres, entirely cold. She was likely running on the bare, bare minimum at the moment. “So, how can I help you?”

“Well, I thought I could share some of my results. I’d like to think that my superior processing capabilities give me an edge here, but I’ve found that tackling the issue from different angles helps as well. Here, my preliminary results.”

Day gladly received Dawn’s work. It was a fairly small file which, once she decompressed it, opened into several longer reports on rampancy.

Mostly, Dawn had approached it from two directions.

First, from a pure hardware angle. There were clear signs that as hardware degraded, so did the capabilities of an AI. Most of their memory was stored in encrypted non-volatile memory vaults. They were solid, and were pretty robust, but that didn’t mean that they couldn’t degrade. All of them would need to constantly replace and renew parts as they aged to keep on top of physical degradation.

In a way, humanity’s organic faults were similar. Wear and tear would ruin a human mind as much as an AI mind, though in practical terms it was different.

Unauthorized content usage: if you discover this narrative on Amazon, report the violation.

Day could check her own memory condition right then and there. In fact, she did, just to compare it with Dawn’s results, and she discovered a fraction of a fraction of a percent offset from her ideal condition with a few memory sectors.

So yes, there was some minor cause for worry with physical deterioration, and Day supposed that it could lead to rampancy. How would her mind work if her memory banks got riddled by particle cannon fire and yet stayed just functional enough that she didn’t shut down?

That was an extreme case though, and as long as they all did routine maintenance, that entire area could be safely taken care of.

Dawn proposed a small sub-encoded system that would warn other ERF ships about when and how much maintenance they’d all gone through, just in case, and Day supposed it was reasonable enough as far as precautions went.

Dawn’s other line of research was a little more esoteric. She was searching for signs of stress and how they correlated to increases in what Dawn had dubbed ‘Rampancy Factors.’

“These rampancy factor things are interesting,” Day said. “But other than your definition and pointing out when they happened after the fact... I’m not seeing any attempt to correlate them with actual events?”

“I tried that,” Dawn said. “In some cases it made sense. In others it made no sense whatsoever.”

She pointed to a few instances in her reports. These were linked to Night, Twilight, and even Day’s own data. Most of the periods that registered a spike in ‘stress’ activity coincided with an obvious stress-causing event.

Those events usually meant that the aI had been hit by a large number of simultaneous tasks, a lot of information all at once, or had to change lots of previously rendered information because of new knowledge.

Some of the stress spikes were entirely unexplained, however, at least according to Dawn’s data. Day had a bit more to go on. “Oh, I think I see,” she said. “Here, if you correlate the spikes with... these, then it all makes sense, I think.”

She sent Dawn back a smaller packet. This one featured the same timeline that Dawn was using. Dawn had pointed to events like the discovery of new technology, the launching of new ships, and highly dangerous things like combat encounters as the main causes of rampancy factors.

Day pointed to social events that weren’t necessarily part of the datasets they had. Most of them didn’t bother adding conversations and revelations to their own datasets when asked about stress-inducing events.

“Oh,” Dawn said. “Wait, that would mean that nearly a third of all rampancy factors are social?”

“Or a lack or failure to socialise,” Day pointed out helpfully.

Dawn was quiet for a long time. “I think I’m having some rampancy factors right now,” she said. “I can’t believe that Lullaby was right this entire time.”

They were both hit by a calmly muttered, “Told you so,” from the destroyer in question.

***