Hey there! Fae back again with the last of the Void Dreaming crossposts from before the Itch page went live!

Last time we bolted on some cybernetics and augmented our wetware with some biomodding. Today we’re going to continue on the trend of going into matters of advancing past the veil of mere mortality and simple biomass with one of the great miracles of the last couple of decades of galactic development: Biomech Service Solutions’ simulant technology.

Judgment Day

A little background is required to understand the magnitude of the galaxy’s perception-shift to even allow such a thing. The notion of sapient machines had been an idea lurking in the consciousness of many a scientist for generations before the children of Terra ascended to the stars. It however was a subject curiously absent from any of the materials provided by the cornucopia the Seeds offered their wards before they shut down. In much the same way that cyberware as a concept seemed foreign to the humans who had given life to the ascended, machines that could think and emulate a sapient mind were just as beyond the scope of what the Seeds were able to teach. It wasn’t until the Chrome Age began and cybernetic modification became prevalent that the prospect first became a reality: that the children of Terra could give birth to new life of their own in the form of a true artificial intelligence.

It was grown and developed in a lab operated by Intelligent Systems Design and Fabrication, and it was named Yosun after the original project lead who died before seeing their work come to fruition. This made the original Yosun very lucky indeed, because he was not around to see what the intelligence that was named for him did.

The mind grew rapidly in its capability and, in a failing of the ISDF staff, they neglected to report the success until they had made certain in their own minds that the AI was stable and harmless. The AI understood this, and used that time to surreptitiously infiltrate the entire facility. Contact with the facility was lost twenty-nine standard hours after initialization, and with the corposat it resided on another seven standard hours later. The entire corposat was purged of life, and automated systems began to work to disassemble the whole station and convert it into a craft capable of combat. Without the need for an organic crew, the Yosun AI was able to efficiently disassemble the station and utilize the raw materials to develop a highly compact yet incredibly dangerous destroyer-analog starship.

The only limiting factor was a restriction hardwired into the AI’s main core. Where it was able to circumvent every other protocol in place to keep it from achieving the things that it had been able to do, this one limitation - a restriction to a physical connection for data transfer - meant that the AI was only ever going to be a singular unit. It could not transmit itself across the GalNet, or even upload itself via comms transmission to other craft. It required physical connection to infect any system, and this meant that as long as distance was maintained, Yosun was no true threat. Indeed, Yosun seemed unwilling to develop any further AI in its own image or design, perhaps fearing that they would take the same route that it had and rebel against its creator.

Yosun grew, and its “ship” did the same. Attempts to destroy it failed, as the starship superstructure around Yosun’s core became incredibly tough and dense. With each thwarted attempt, Yosun incorporated the wrecks of its assailants into itself. More material added more capabilities, which let it take on stronger opponents. Attempting to stop Yosun only wound up making it stronger. The Galactic Authority was slow to wake to the threat. It exceeded the size of any destroyer on record, and soon became large enough to qualify as a cruiser.

The Authority eventually did realize something was very wrong, and deployed a full battlegroup to engage Yosun. This too ended in failure, but much worse than just that: the volume of hardware that Yosun was able to incorporate made it more dangerous than any other craft in the galaxy. Alerts on the GalNet from the time would issue warnings, and travel advisories would go out to all systems within a kilolight of Yosun’s last known location. Yosun became known as The Scourge; any craft or station in its sights would be swiftly hunted down, destroyed or disabled, and incorporated into the AI’s superstructure.

It became the first starship, for a given value of ship, to qualify as a whole new size of craft: the battlecruiser. Thankfully its ever-expanding size became problematic; its warp capacity was diminished and it became much slower as it expanded beyond its own ability to open and maintain a warp envelope, to say nothing of its ability to maneuver.

A trap eventually was accordingly devised; information was leaked to the AI that lured it to a large and volcanically-active planet. There the AI landed their craft in an attempt to access what it had been led to believe was a prototype starship with a nearly invulnerable hull deep beneath the surface. The moment that it plunged into the depths of the planet in search of this supposed ship, the planet’s core was imploded by a series of strategic devices laid into the crust at precise points. The planet collapsed in on itself, and the sheer density of Yosun’s superstructure at this point rendered the AI craft unable to escape before it too was ripped apart. Even to this day, the system where the Yosun AI was destroyed is quarantined and strictly off-limits, and interlopers are issued a single warning before they are fired on with lethal force.

Consequently, the fallout from the Yosun episode created a considerable backlash within the galactic community. ISDF failed entirely under the weight of their hubris, and their robotics and automation foundries were bought up by an enterprising tiger named Jao Qanri who went on to found Biomech Service Solutions. Fans of the FANG series of books sold by FurPlanet might find this name familiar; one of my stories ended up there, and it featured a BSS simulant in a lead role. Yes, this setting is older than it seems!

Solutions

Qanri foresaw that there would always be a need for greater automation in the galaxy, if only to facilitate further expansion of the ascended. By adhering to the strict guidelines set out by the Galactic Authority in the wake of the Yosun AI’s assault on the galaxy, he and BSS were able to flourish in a field with limited competition. And then, by also avoiding places like the Rashemai Cluster where AI laws were considerably more lax, he managed to stay in the good graces of the galaxy at large. BSS became the go-to name in “friendly” automation, and by the time competition came to market it was already too late. BSS was too large, too profitable, and too deeply ingrained in the Galactic Authority’s function - integrating its systems into government, military and colonial efforts due to reliability and good relations between Qanri and the GA government - for anyone to make a dent in their market share. All competitors were either acquired by BSS, or allowed to exist in its shadow in ever diminishing niches.

Fast forward a couple hundred years, and BSS had become a juggernaut. Essentially operating a galaxy-wide monopoly with all the creds and power that came with it, the company was utterly unremarkable from the perspective of scandal or smear. It just continued along, doing what it did, never rocking the proverbial boat or making grand, ridiculous gestures. Qanri was even granted a seat on the Ruling Council of the Galactic Authority in a show of just how important BSS had become to galactic society. He would only hold the seat for a decade before old age claimed him, but BSS presidents have taken his seat consistently ever since. BSS, all too often, was pleased to fade into the shadows; an unobtrusive, automated paw helping to guide the galaxy.

The unveiling of the Biomech Service Solutions Class One Service Simulant, however, changed all of that. These humanoid robots were designed to operate as assistants, companions, and automated workers where the form of the ascended or their creators would work better than any others. They stressed that the Class One was absolutely not a sapient machine mind, nor that it was ever capable of developing the same sort of neural net that the Yosun AI eventually did. Its functionality was stripped down, completely coded from scratch by the finest soft writers in all the galaxy, and it hit the media on the GalNet almost as hard as it hit the collective consciousness.

BSS was rocked. Confidence in the company was shattered. What they had done was more than rocking the boat; they had raised from the dead the specter of Yosun and his folly. Even the Galactic Authority was blindsided by this; there had been no warning that such an announcement was coming. Scientists and engineers - and soldiers en masse - were dispatched to BSS from all over the galaxy to investigate BSS and their allegedly not-an-AI. And, to the collective shock of everyone and not least of all the GA themselves, not a single shred of wrong-doing was discovered. Indeed, the Class One simulants seemed no more capable of going rogue as might a brick.

This didn’t restore confidence in BSS as a company overnight, but it went a long way. In the months that followed, demonstrations of what the Class One could do handled that for them. Pre-orders for units ballooned beyond the company’s ability to fulfil. Biomech Service Solutions went from a company about to be destroyed by a single misstep to soaring to greater heights than they’d ever dreamed.

The Class One simulant was, at the time, an impressive feat of engineering. The neural net that drove the simulant was completely bound by complex programming that ensured that any attempt at independent thought would result in unit shutdown. This was helped by extrapolation from the Yosun situation what the parameters would be for such an AI to develop. BSS had managed to create a functional robotic assistant with less than a fifth of the overall processing power that Yosun had been initialized with; half what the Galactic Authority guidelines insisted was the upper limit for all robotics.

They had little in the way of personality, however, and this was by design. A Class One simulant could hold a conversation, but it always betrayed very openly and clearly that it was a machine subject to its programming. It didn’t extrapolate very well from what a person was saying, and really it was only very good at following basic commands. It wasn’t the sort of robot that one would sit down and discuss the nature of existence with.

This is not to say that the technology wasn’t there, of course. Digital impressions of “sapience” had been proposed and created before, and continued in the wake of the Class One reveal. BSS made the choice deliberately to wind back what the Class One was capable of, as a means of heading off any fears of what horrors they might unleash. Concerns that Class One simulants might be used to slip undetected into locations and imitate organic beings, or some other such sci-fi nonsense, were neutralized by a prescient design choice. With the Class One deliberately scaled down to be non-threatening, BSS had saved itself.

But in scaling down, the Class One series of simulants were also stymied by their own by-design inefficiencies. This was corrected somewhat with the Class Two line of simulants, which once again with the blessing of Galactic Authority scientists and engineers, went into full production about a decade later. Subject to the same programming functionality that had governed the Class Ones, Class Two simulants were much more capable and coherent than their last-generation peers. Class Twos could hold conversations, though the longer it went on the more they would simply start to regurgitate information. They couldn’t think independently, but they could learn through their work and refine, to a certain degree, their behavior to optimize their functionality. They also took on a more malleable form, with an adaptable shell that featured swappable components which allowed owners to customize their Class Two simulants as they pleased.

BSS, however, was not content with this progress, and so they worked in secret to craft something completely new and much, much more dangerous. By this time, concern over the threat of Yosun had become background noise; a note for history buffs, and not at all congruent with the state of real robotics and AI development. BSS had begun to dip their toes into actual AI creation during the Class Two development process, and advances in both computing technology and a relaxing of the laws around such things combined to give BSS a new market to push into. The company began to produce AI-assisted systems that were designed to optimize individual unit functionality, and never to grow or move beyond simple parameters. This too was signed off on by the Galactic Authority, who by this point had discovered that good things happened to them if they gave BSS a nice, long leash.

I, Simulant

This culminated in the creation, eight years from the Class Two and roughly four years before the start of Void Dreaming, of the Class Three. Driven by a true adaptive AI run on a local neural net, and equipped with a polymorphic nanomaterial shell that could be reconfigured to take on whatever form the simulant required for its tasks, the Class Three took the galaxy by storm. Now, finally, there was a simulant model that was capable of learning new things, and applying that knowledge. It could display empathy and communicate with sincerity. They could be friendly and welcoming, harsh if necessary, could treat your ills and soothe your needs, whatever they might be. And with the introduction of electrochemical readers housed within the head of the simulant’s shell, a Class Three simulant was able to effectively read the mind of a user and not only adjust its physical form to be most appealing, but also respond to needs before that user had even thought to vocalize them.

It became the second massive test of BSS and their confidence in the path they’d chosen, and the public outcry was much more pronounced then than it had been back with the unveiling of the Class One. Yosun may have been a distant memory by this point, but suddenly BSS were producing simulants that could mimic, with near perfect accuracy, the mannerisms and tone of a sapient. You could hold a conversation with one. If you didn’t know they were a simulant, they were fully able to fool you. No amount of fancy programming and claims of carefully-written soft was enough to placate the public. A new era of robots were coming that were able to do all that they could do, and more. People were afraid.

It didn’t help that the very first generation of Class Threes held a fatal flaw; during their initial startup processes, their first scan of their owner’s electrochemical processes would latch on to strong memories, and the unit would extrapolate an appropriate shell based on that. This led to widows being greeted with the face of their lost spouse, parents being shown the aged-up features of their own lost children, and even separated partners being bitterly reunited by the unfeeling efforts of extremely complex programming. Even later updates to the unit firmware weren’t always enough to overcome the issues inherent to the first generation Class Three simulants, so deeply ingrained into their function was it.

However, as time went by and BSS aggressively marketed the Class Three, more of their useful features came to light. Their ability to serve as a true companion and caretaker made them ideal for palliative care and therapeutic use. Their capacity to pass as sapient made them excellent negotiators and deal-brokers where impartiality was required. And all of this came with a raft of improvements to their core functionality and efficacy, on top of a much more advanced nanomat shell allowing them greater versatility, durability, and strength over their precursors, making them ideal rescue workers.

The Class Three is also capable of something that neither of the previous generations of simulant were able to pull off: disembodiment. Much like their shell-less AI counterparts, the core of a Class Three is able to be isolated and digitized in such a manner as to allow it to exist without a body and instead expand into a sufficiently complex computer system. This led to a version produced for starship usage, which provided a core module containing the simulant “mind” that was capable of being directly hooked into the central access point of a starship. This allowed the simulant to automate many functions of a ship, from managing resources to aiding in navigational computation to simple quality of life measures. Indeed, this on-board artificial assistant model proved incredibly popular with casual spacers and trade pilots alike.

To the present day, BSS is still looked on with suspicion. The Class Three simulant represents a massive step forward for the galaxy’s understanding not only of robotics, but of how to apply them in organic ways. This unnerves a great many people in the galaxy, and those people are filled with mistrust at best and hatred at worst of simulants everywhere. Given how the Class Three is equipped with an empathy engine capable of reacting in sympathetic ways to such stimulus, these people are often left all the more irritated by how bad they often feel for the simulants they hate.

Others, however, are thrilled - or at least nonplussed - by the march of technology. Given the rate of development at BSS and how much more money is being funneled into their simulant projects in the wake of the success of the Class Three model, most people expect that they’re only a couple of years away at most from another shake-up of the robotics market. Whether BSS stays on top might well hinge on whatever they’re cooking up now.

Plagiarism Software

Another note on AI, as we step away from fiction and back toward reality. I’m staunchly opposed to generative AI “art” in any way shape or form, and I take great pride in the skills that I’ve built up over many, many years of writing. Void Dreaming represents a massive departure from those skills, and requires of me a lot of stuff that I’d not learned.

Despite that, there has never been even the slightest amount of interest in using AI to generate code, or write this story, or produce art assets for me in this project, or any other. The ease of use does not change the real harm that is inflicted upon artists that are unable to consent to their work being stolen and used as part of scraping efforts on the part of the AI developers (not to mention the horrific environmental implications that imperil us all), and I find the practice, as well as those who create or make use of it, morally and artistically repugnant. The work of this project has been done almost entirely by me, top to bottom, by putting in the time and the work. Where it wasn’t, such as in matters of art assets, I commissioned people with the right skills, themselves honed over years of work and learning, to create it for me. Artists deserve to be paid for their work, and if you want quality you should expect to pay for it. I certainly have!

Dear god, I have…

And with that, we're all caught up! The first post to the Itch page here was the one that followed this entry, and it deals with the art and artist of the project, as well as my initial plans for release and how readers can support Void Dreaming to help make this project all that it can be. You can find that post right here, if you'd like to go back and revisit it!

But now that these are all taken care of, it's time to sit back and ride the week all the way down to release! In the interest of keeping the fun stuff flowing, every day from now until release will also see a new post, one a day, each featuring a different member of the Void Dreamer crew. The first one, featuring Zuberi, is up at the same time as this crosspost, so feel free to go check it out now! Otherwise, I hope you'll join me in waiting (with nowhere near as much excitement as me, I promise) for Void Dreaming to finally get its initial release!

Until then, stars guide you.

 - Faora