Continuing the first drafts of Splintered Demons. To learn more about this book or view the table of contents for these first drafts posts, visit the novel page here.
The current state of the art in synthetic intellect design consisted of a set of quantum processors arranged in a self-repairing unregulated network cluster, supported by a swarm of midges which were capable of constructing additional processors and network links so long as they remained within the designated boundaries of the mind. Some [designers] preferred to use electro-optical processors, distributed midge swarms, or self-replicating viral/phage processing units, but these each carried as many restrictions in their own way as attempting to build a fully mechanical mind. The chief benefit of quantum processors, despite their comparatively large size, lay in their inherent ability to generate truly original results. Philosophers of intellect still argued as to whether any intellect, human or synthetic, was genuinely capable of originality or if even the most aberrant concept emerged from the organized chaos of preceding influences and could, with sufficiently large datasets, be predicted, but such arguments were irrelevant to Dyson’s study of synthetic minds.
Whatever substrate the syntellect was built upon, the most important decision a [designer] would make consisted of two elements: The seed and the soil.
The seed in this case was a core operating system. A comparatively small kernel of automorphic code which would serve as the foundation for the syntellect. After more than a century of research, some of which was conducted by early generations of syntellects themselves, the original three codebases had exploded into dozens of syntellect seeds. Some were highly specialized, virtually guaranteed to generate a syntellect which prioritized emotional intelligence over obsession with statistics, while others were specifically designed to be highly flexible, resulting in a syntellect which could adapt itself to virtually any situation given a few wakes to absorb the relevant data. Most true syntellects had some level of adaptation, it was considered unethical, and even illegal in some Zones, to spawn a syntellect without full self awareness and the ability to choose its own work, after a period of indenture intended to train the new mind and pay for its development. For applications in which such independence was undesirable, the subfield of not-quite-sentient restrillects had burgeoned into a Shell-wide industry.
The soil consisted of the experiences, real and simulated, to which the [designer] exposed the newly seeded syntellect in the weeks and months after their inception. As with humans, the precise pattern of experiences and opportunities, combined with varying degrees of restriction and deprivation, resulted in a fully formed synthetic intelligence. Interaction with carefully trained humans, as well as selected syntellects, resulted in a distinct personality for each syntellect, one which would continue to evolve through the syntellect’s lifetime.
What bothered Dyson about Anomaly’s claim was neither that it seemed to indicate a completely new seed of intelligence, nor that a syntellect might grow in such an untamed manner, since that was no different that’s the manner in which humans had been raised for tens of thousands of years. The thing that stuck in his mind like a burning knife was the Anomaly’s claim that it had only been sentient for a few days. The average syntellect spent over a year in development before being released from nurturing to work in its chosen, or designated, field.
If a technology existed to create a fully functional syntellect in less than a week, every culture in the Zone would kill to gain access to it, to say nothing the Anomaly’s apparent lack of a physical body.
“I may be of some help to you in examining the Spire,” Anomaly said.
“How so?” Dyson said aloud. What he thought was, This could be very dangerous. Even as the idea formed in his mind, he realized that it was fruitless to hide his misgivings. The being which hovered in the air before him was capable of detecting his every thought. And after only three wakes. What would it become in a week, a cycle, a year?
In answer, the anomaly drifted backwards until it rested against the wall, then continued to move back and through the wall. The Raven’s Flight was built from midge-infused metaloceramic composites. In theory, nothing could pass through its hull except gravity and the odd subatomic particle.
Dyson leapt from the sofa and stumbled over to the wall where Anomaly had disappeared. He pressed his hand to the hull, feeling for any cracks or holes which might have provided passage for a gaseous creature. He thought that the plating might be marginally warmer at the center of the place where the entity had passed through, but he could otherwise detect no change.
“Do I impress you?” Anomaly asked, appearing in the doorway.
Dyson stared at the thing, then back at the wall where it had passed through. He felt questions forming in his mind, but made a conscious effort to block them. Think of pink elephants. Just think of pink elephants or nothing at all.
“A most unusual creature. It almost seems as if…” Anomaly faltered and Dyson imagined he saw its body pulsate a few times.
Rationally, he knew that whatever mechanism Anomaly employed to read his mind was entirely passive or, at the least, probed his synapses with such subtle effector fields that he would never feel them, but that knowledge did not stop him from shivering as he imagined translucent tentacles engulfing his head. He forced his mind back onto the subject of abnormally pigmented pachyderms.
“Ah. I understand. You are visualizing an irrational construct in order to establish mental privacy while you process the unlikeliness of my existence. That is understandable.”
“I just need some sleep,” Dyson muttered.
“Gladly,” Anomaly said and, before Dyson could protest, or even comprehend what the etherial creature was about to do, it reached into Dyson’s mind and sent him into a deep, dreamless sleep.
# # #
We need to think clearly, Zau directed.
It was the single advantage they had.
While the others appeared to have been fully scrambled, whether by the rapid neural imaging, years of dormancy, or trimming made necessary by overloaded data banks, Zau, or the conglomeration which thought of themselves as Zau, retained enough independent thought that they were often able to rise above the clamor and direct the hybrid syntellect towards effective courses of action. As a result, she had emerged as coequal with the combined forces of all the other intellects. A delegate with votes equal to every other member of the assembly.
To learn more about this book or view the table of contents for these first drafts posts, visit the novel page here.