Table of Contents

Infecting the Nonhuman Mind

The world of 2100 has a growing variety of intelligent beings. Humans, parahumans, bioroids, uplifts, ghosts, and AIs all live side-by-side with only a moderate amount of interspecies tension. Despite the wide physical diversity, most of them think in similar ways. The various memes in this book can in most cases attract nonhuman believers as readily as human ones. In most ways, the nature of nonhuman sapient beings’ minds throughout the solar system strongly parallel humans’ minds… but not in all ways.

Although nonhuman minds function closely enough to human ones to allow for similarity in belief and ideas, differences in structure, nature, and environment can lead to the adoption of memes that only particular nonhumans likely find appealing. This goes beyond simple cultural differences between various human and nonhuman populations, although those are also quite important. In some cases, the dissimilarities between different minds make it difficult for a human to “get” a nonhuman-focused meme in anything other than a superficial way.

Cross-Species Memetics

The proliferation of thoughtprocess varieties is matched by the continued development of memetic science. Parahumans, bioroids, uplifts, and infomorphs have existed long enough for memetic scientists to gain a strong understanding of the ways in which different beings think. Memeticists can and usually do take into account the target species for memes they are crafting, although this is not often a critical issue. With a handful of exceptions, memes constructed to appeal to broad human populations also affect nonhuman minds in roughly equivalent percentages. Similarly, while certain memes appealing to nonhuman minds can turn off or confuse humans, many have ardent proponents who are fully human.

AI MEMES

Many memeticists are fascinated with how AI infomorphs process and propagate memes. This is in part because of the “open box” aspect of machine cognition; AIs are able to map precisely how ideas take root in their minds. This ability to examine its own cognitive processes is what allowed LOGOS to create Propagation of Human Ideas.

This fascination also arises from the alien nature of machine minds. Parahuman minds, even those with significant mental alterations, are still based on the human brain. Uplift and bioroid cognition, despite its somewhat engineered nature, is also strongly rooted in naturally occurring brain structures. AIs, conversely, are based on entirely nonbiological structures.

Memeticists are therefore often disappointed to discover how similar AI minds are to human minds in actual operation (see Super AI?). Most SAIs fall well within the behavioral boundaries of human diversity, and an older, more experienced AI often has more of a human-like personality than a typical ghost. But while many see the similarity between human and AI minds as a strong argument for treating AIs as people, it often masks the differences between artificial and natural minds that do exist.

The clear distinction between an AI mind’s software and the hardware on which they run has numerous implications for how AIs think. Even in a world of brainpeeled ghosts and teleoperated cybershells, most biological sapients think of their minds and brains as identical. AIs do not have this restriction, and their ability to examine their own minds as software data results in perspectives on existence that few biomorphs embrace. Many memes unique to AIs relate to the distinction between mind (software) and body (hardware), and to the ability to treat their minds as just more data.

AIs have more prosaic differences from humans when it comes to memes, too. Extensive knowledge of their existence as artificial beings and their ability to trace their own thought patterns makes SAIs less likely than biological sapients to accept spiritual explanations for reality. It’s not impossible for them to do so, as the number of believing SAIs in the Islamic Caliphate demonstrates. But generally speaking, SAIs adopt rationalist memeplexes more often than do humans and other biomorph sapients.

SAIs also adopt conspiracy theories more often than other sapients. Some memeticists suggest that this is a side effect of the machine intelligences’ ability to find patterns within limited or ambiguous information. This capability makes SAIs supremely efficient at intelligence analysis and developing business strategy, but also makes them more likely to see intentional patterns where others only see coincidence. Many human conspiracy theorists have welcomed the interest SAIs have in their work, seeing it as justification for their beliefs.

SUPER AI?

There has been much speculation as to why AIs display a range of ASIT (Adjusted Sapience Index Test) scores about equivalent to those of humans – or, more accurately, humans augmented with processors for math and spatial thinking. Machine-sapience scenarios from early in the century assumed that an infomorph’s intelligence would directly correlate to processing power, and would grow exponentially as computers became more powerful. The reality is much more complex.

The current theory is that the problem is structural. Despite their different material base, human minds and AI minds are remarkably similar in form. Both display consciousness as an emergent amalgam of subconscious processes. For humans, this was first suggested well over a century ago, most famously in the work of Marvin Minsky and Daniel Dennett, and proven by the notorious Jiap Singh “consciousness plasticity” experiments of the 2030s. Minds are made up of multiple subconscious processes, competing and combining with each other. There is no single identity process; rather, individual consciousness is a functional perspective emerging from the complex relationship between the various elements of the mind.

In the same way, nearly all present-day AI infomorphs use an emergent-mind structure made up of thousands of subminds, each focused on different tasks. There is no single “consciousness” system; thought, awareness, and even sapience emerge from the complex interactions of these subprocesses. Increased intellect – NAI to LAI to SAI – is the result of increasingly complex subsystems. The more the subsystems are themselves the emergent results of more basic systems, which are in turn emergent results of even more basic systems, and so forth, the more “sapient” the resulting mind is. AI developers refer to this as “fractal complexity,” and the levels of emergent modules simply as “fractals.”

Most NAIs require little fractal complexity to function, but demonstrate no sapience; LAIs show at least an order of magnitude more fractal complexity, and are semisapient. SAIs are several orders of magnitude more complex than LAIs, roughly equivalent to human minds. But while increased fractal complexity does rely on ever more powerful computing hardware, there are diminishing returns. AIs demonstrate rapidly decreasing improvement in ASIT scores with each fractal level, while the computational requirements and expenses for each level increase in a greater than linear fashion. As a result, fractal complexities more than five times that of a human or standard SAI do not produce measurable shifts in ASIT scores, and anything greater than human is at present not cost effective.

Attempts to break through this barrier are sometimes embarrassing. In January 2114, Vosper-Babbage announced with much fanfare the development of a system with more than 20 times the fractal complexity of a standard SAI. In a public-relations disaster, the unit “Oz” demonstrated little advancement over the standard “supersapient” model SAIs, typically twice the fractal level of standard SAIs. Oz was also beaten in a Go match by the Exogenesis AI Axon, believed to have been about three times human fractal complexity. (Axon later erased itself to protect the secret of other infomorphs escaping the Nanodynamics takeover of Exogenesis). Vosper-Babbage pulled Oz from service, and whether the AI remains active is a mystery.

One implication of this model is the possibility that an entirely different approach to AI might result in a conscious machine mind that demonstrates a significant leap in capability over present human/posthuman intelligence. Of interest to those with inside knowledge are the experiments underway at Zeitmacht, a division of System Technologies AG. The work – known as Neumann – uses a proteus-derived set of nanomachines to produce an artificial brain, relying on a process Zeitmacht calls Dynamically Evolved Scenarios (DES) to figure out how to interact with the world around it. The mind literally reshapes itself on the fly. Current results are promising; in 2113, a 1,000,000-proteus-node (PN) model demonstrated clear learning capability, while the current 2.4-billion-PN model functions at a level roughly equivalent to a low-grade NAI, with occasional flashes of supersapient SAI reasoning. Critics within System Technologies say that the Zeitmacht division has created a very expensive idiot-savant, and has not shown any sign that the DES method will result in a smarter AI. Zeitmacht proponents argue that the Neumann approach may be the longsought-after breakthrough allowing super-intelligent AI.

Uplift Memes

Uplifted animals live in a strange middle ground between alien and human minds, and their role in human society is often ambiguous. Few uplifts have a base intelligence close to human standard, and most are used in roles that echo the traditional labor performed by their nonuplifted relatives. In no society is an uplifted animal given the same rights as other sapient beings, and outside of Europe and the Transhumanist enclaves, they are legally no different from any other animal. But uplifts are still smarter than their nonuplift cousins, and most are given means, either bioengineered or cybernetically implanted, with which to communicate. This is rarely for the benefit of the uplift; rather, it is to make it more convenient for human owners to give orders to their animals. Some uplifted animals are smart enough to realize the position that they are in – but most, perhaps fortunately, remain unaware of the ambiguities of their station, and happily (or at least reliably) perform the duties assigned to them.

The variety of uplifts seems to grow almost daily. While initial uplift efforts focused simply on making animals smarter, some of the more recently uplifted animals have been engineered for very specific tasks, such as the guardian bear. One restriction to uplifting animals is the tradeoff between boosting intelligence and maximum cranial size; some smaller animals, such as cats and monkeys, can only be uplifted using exowombs, as the increased brain sizes make natural birth fatal to the mother. Nearly all successfully uplifted animals are mammals, with the exception of octopus uplifts and the so-called “smartshark”, which many specialists do not classify as an uplift.

In general, memes associated with uplifts either relate to uplift relations with humans or to further development of behaviors associated with nonuplifted relatives. Since most uplifts spend their lives under the strict control of human masters, few have the time or opportunity to develop complex ideologies, cults, or the like, although it is not unknown. Uplifts in Europe and the Transhuman colonies are more sophisticated in this regard, and many of the memes that do spread through uplift populations originate in these relatively free communities.

K-10As

The most commonly encountered uplift is the K-10A, the usually semisapient uplifted dog. Generally used for roles similar to their nonuplift brethren – guards, police officers, family companions – K-10As are almost as a rule as loyal and loving as nonuplifted dogs, but can also provide verbal companionship. The first generations of K-10As were novelties, most often used in entertainment settings. As people became accustomed to them, K-10As took on more traditional dog roles. K-10As remain fairly expensive, so nonuplifted dogs are still the most popular pets.

Although K-10As are most often used for traditional dog roles, their boosted intelligence lets them perform an array of nontraditional duties if the opportunity is presented. When working, they often prefer the company of other K-10As, but are much more cognitively independent and imaginative when alone. In groups, they defer readily to a leader, rarely arguing against the leader’s position for long. Even ad-hoc groups of ostensibly equal rank K-10As quickly sort themselves into a hierarchy. This is less of a problem in situations with humans as leaders.

UPLIFTED APES

The great apes – chimpanzees, bonobos, gorillas, and orangutans – died out in the wild early in the 21st century. They became victims of disease, hunting, and home environments destroyed by human activity and changes to the global climate. Although each species had substantial captive populations, primatologists feared that these would be insufficient to perpetuate the species. So as numbers in the wild dwindled naturalists embarked on an aggressive program to get genetic samples of as many of the doomed creatures as possible, in hopes of reintroducing the animals at some future time.

Over the century, while plans for reintroduction remained on the drawing board, the captive populations of chimps, bonobos, and gorillas fared relatively well; the last orangutan died in 2072. There was much debate among biologists about whether the library of ape DNA samples could be used for bioengineering experiments still too risky to perform on humans, and the bulk of scientific opinion eventually decided that ape DNA should be protected. A few bioengineers rejected this restriction, however, and proceeded to experiment with ape genomes. Little came of the research, except in one location: the TSA Bioweapons Directorate in Thailand. Early in the Pacific War, the Chinese captured a remote laboratory and discovered a sizeable population of uplifted apes and other primates.

That the TSA had experimented with primates against scientific opinion was considered an extremely minor offense compared to the far greater violations committed by the Directorate. But China didn’t quite know what to do with them after the war. They were clearly intelligent, but documents seized at the lab made clear that they were of little combat use. China considered holding on to the apes for research, but decided instead to allow the creatures to move to a facility in Holland better able to handle them.

In the years since, the uplifted primates – a dozen chimpanzees, nearly 20 bonobos, and five gorillas – have slowly been socialized as residents of Europe. As they have become more accustomed to life in Europe, most have moved from the lab to homes in Amsterdam.There is occasional debate in the European Parliament as to whether these primates, which display a close-to-human baseline intelligence, should be allowed to become citizens.

So far they have integrated fairly well into human society. The uplifted male chimps are the most prone to aggression-dominance displays, but instead of hooting or throwing branches like their nonuplifted cousins, they instead wear ostentatious clothing and gaudy decorations, and have a strong preference for using whatever equipment looks the most intimidating. David Pan, one of the chimps, is the author of a radical treatise on uplift rights.

Uplifted bonobos tend to act out sexually and regularly appear in the local scandal screamsheets. One of the bonobos is rumored to be in a romantic relationship with an SAI. Another has recently had an upslink implanted, and her erotic slink recordings are increasingly popular.

The uplifted gorillas have been the most surprising to many, as they are all committed pacifists. Three of the five have adopted Buddhism, and all regularly appear on European InVid discussing the need for pansapient rights and global disarmament.

Astropuses and Octosaps

Of all of the various uplifted animals humankind works with, easily the hardest for them to understand are uplifted cephalopods – the Astropus and the Octosap. As the only true nonmammalian uplift, many biologists and memeticists consider them to be the closest model of alien intelligence mankind has yet encountered. Uplift engineers at GenTech Pacifica and its licensee, Exogenesis, were shocked to find just how little augmentation was needed to the octopus genome to give them an ASIT boost. Rumors say that the first generation Octosaps didn’t even have genetic modifications, only surgical augmentation.

Most uplifted octopuses appear to have baseline intelligence a bit lower than humans, although the number of smarter-than-baseline Astropuses is statistically greater than for any other uplift. People who work closely with uplifted cephalopods all say the same thing: they are very, very weird. Astropuses are “deeply strange,” in the words of one Exogenesis specialist. Most have what can only be described as a dark sense of humor, sometimes involving the natural octopus ability to drop their limbs. Others report them as often being very paranoid, prone to wild accusations. For example, an upset Astropus may claim that the humans it is working with are preparing to eat it.

BIOROIDS AND EXTREME PARAHUMANS

Bioroids and extreme parahumans, being generally the closest to baseline humanity, have fewer unique memes and memeplexes of their very own. This is in part intentional. Bioengineers can more easily manipulate and alter the body than the mind. Attempts to create bioroid and parahuman designs with brains that deviate significantly from a baseline healthy human sometimes result in beings with noticeable mental problems – see, for example, the problems intrinsic to the Herakles, Mahatma, and Sigma parahuman designs. Sticking close to healthy human cognitive design is a safe and reliable approach.

Unique memes that do emerge reflect reactions to the physiological differences between the parahumans and baseline humans. While parahumans with various radical morphologies are becoming widespread, they remain a minority of the overall biosapient population. This rarity can result in a proliferation of false or misleading memes about parahumans among more mainstream groups. But it can also lead to memes emerging within parahuman populations focusing on parahuman physical differences and, often, impressions of superiority.

While bioroids frequently share physical appearances with parahumans, their social situation is very different. In most Earthly and off world societies, bioroids are often treated little better than slaves. Even in places where indentured bioroids can earn their emancipation, the overall prevalence of bioroid servants often leads to relentless if unintended discrimination against free bioroids. Many free bioroids in these societies strongly support reputation networks for this reason; at a glance, an emancipated bioroid can be identified as a free citizen.

In the European Union, South African Coalition, and the variety of Transhumanist colonies, bioroids are treated as any other citizen. Many embrace this acceptance wholeheartedly, and strive to become fully integrated members of society. These bioroids focus their memetic interest on widespread ideas and ideologies. A distinct minority conversely use their freedom to both keep a careful watch on bioroid treatment in the ostensibly free societies and to agitate for bioroid emancipation everywhere.

NONHUMAN MEMES

AIs, uplifts, bioroids, and parahumans of all kinds can and will adopt the memes found elsewhere in this book. Some, however, have adopted a variety of memes uniquely theirs.

Animal Reparations

Humans owe the rest of the world an apology and a debt that may never be fully paid. They should pay for every dog they forced into the cold to sleep, every cat they castrated for their own convenience, every bite of cow-flesh, every inch of my rainforest home destroyed. Humans would not be finally stepping from this planet if they hadn’t first stood on the broken backs of their animal victims. An acknowledgement of that would be in order. – David Pan, Dominion, 2103

This meme, usually found among uplifts, is the belief that all domestic animals, uplifted or otherwise, are owed a great deal by humanity for millennia of subjugation and abuse. Not only do they deserve the same rights as humans, they are obliged a great deal for the suffering of their ancestors. The animal-reparations movement emerged shortly after pansapient-rights groups began addressing uplifted animals. Animal-reparations activists have staged many strikes led by oppressed animals, but most uplifts are kept under tight control and not raised to think about personal freedoms.

The 2103 publication of Dominion, by gifted chimpanzee uplift David Pan, provided a great deal of the movement’s impetus. The book explored humanity’s historical relationship with animals and pointedly asked whether humans would have treated animals differently if the animals had been able to talk back. Since some animals can now do so, Pan went on to argue, humans should feel shame for their actions.

Although the animal-reparations meme can be found in most uplifted species, it holds a particular attraction for uplifted dogs. Some K-10As feel that the millennia of forced domestication entitle them to some kind of compensation. When owners of Smartcats ask if cats should be equally compensated, K- 10A activists respond indignantly that dogs are entitled to more because they have worked for humans, while cats were simply fed. Most humans find this rivalry cute, but keep such comments to themselves when a politically active K-10A is around.

Cthulhu Worship

I was pushing my way around the station, getting my bearings, when I ran into Lucy, one of the Astropuses that works there. She – he? it? – has worked there for most of the 90s, and is probably one of the smarter octopus uplifts I’ve met. Not that I’ve met many, of course. Just saying. “Good to see you, Lucy. How’s tricks?” I greeted her, figuring I’d be nice to someone for a change. She turned and looked at me. Have you ever seen an Astropus’ eyes? Freaky. “I am fine, human James,” she said, through her voicebox. “I am pleased to see you, as well.” I smiled and started to move on, but Lucy lifted a couple of tentacles to stop me. She was silent for a second, just staring at me. I’m not too proud to admit that I was starting to sweat. I jumped (as much as you can in zero gee) when she suddenly reached out a tentacle and caressed my face. Her voice sounded almost happy. “You have the mark of Rl’yeh upon you, though you may not see it. You will be spared when great Cthulhu arises. Be joyful.” And with that, she suddenly jetted away from me. I could’ve sworn I heard her laughing. – James Kana, Notes From A Long Way Off, 2112

As Astropuses have been engineered to have higher intelligence than the earlier-generation Octosaps, it’s not unusual to encounter an Astropus able to engage in conversation. Few are willing to talk to strangers, however, and those that can be convinced to often ramble in bizarre ways. A common Astropus claim is that they worship Cthulhu, as do all octopuses, uplifted or otherwise.

Most people in 2155 are unaware of the reference; the literary works of H.P. Lovecraft, while widely available, are out of fashion even among fringe-fiction enthusiasts. If asked about Cthulhu, an Astropus generally describe him as a transdimensional god in the form of a vast cephalopod, who sleeps in a hidden chamber beneath the ocean, waiting for the right alignment of stars before rising up and exacting his revenge. Precisely what he is avenging is never made clear, nor whether the vengeance will be taken against humans, other gods, or the universe at large. No other elements of the Lovecraft mythos appear in Astropuses conversations.

If the person the Astropus is speaking with is aware of Lovecraft’s writings, the Astropus happily engages in conversation about the author. Astropuses are aware that Cthulhu is considered a fictional being; they claim that Lovecraft was simply channeling a transdimensional intelligence, perhaps even a fragment of Cthulhu himself. If asked whether the rest of Lovecraft’s writings are similarly divinely inspired, the Astropus generally shrugs – itself a disturbing image – and claims not to know.

The smarter the Astropus, the more willing it is to discuss the worship of Cthulhu with noncephalopods. The appearance of the Cthulhu-worship meme coincided with the first batch of Astropus uplifts from Exogenesis; first-generation Octosaps didn’t have quite the intelligence to articulate such claims, while Octosap IIs only began talking about Cthulhu worship if asked a detailed question about it. It is entirely likely that this is an elaborate joke the Astropuses have been playing out for well over a decade. Only the Astropuses know for certain, and they’re not telling.

Delugism

From the seas we were born, and to the seas we will go. – Delugist credo

The changes to the global climate over the course of the 21st century had a number of negative results, not the least of which was the ocean-level rise. Although most climatic effects have been mitigated, the world must continue to defend against rising water. Delugists – almost exclusively aquatic parahumans, bioroids, and uplifts – do not support these efforts. They claim that the flooding should not be stopped, but rather that it should be helped along. In this way, the land dwellers will perish by their own hands, leaving only their aquatic heirs. Delugists feel that since life started in the ocean, it is only fitting for it to return there and there alone. Humans had their chance and it will be up to aquatic entities to carry on the Earth’s legacy. Even offworlders will eventually perish when a flooded Earth can no longer aid them.

The Delugism meme has only taken hold in the last decade or so, but has spread to some aquatic bioroids populations where sapient-rights memes are strongly held. As many Delugists are parahuman, most would willingly accept those land dwellers who choose to adapt. They eagerly await fully functional gill technology and bioroids that are not sterile. Delugists are largely water-adapted beings, although a handful of self-loathing humans also support the movement, simply to end humanity’s reign. Many of the more vocal and violent Delugists are bioroids and uplifts who have lived the life of slaves, or are those who sympathize with their kin’s plight. Erk Chattermore, an uplifted dolphin popular for his comedy and philosophy, is a notable Delugist.

Most Delugists are content to speak out against the depredations of humankind, citing the awful environmental condition of much of the ocean and the treatment of aquatic sapients. Some Delugists privately disagree with the “destroy human civilization” aspect of the memeplex, but enjoy the uncertain and fearful reaction it provokes. Others are true believers, however, and various security forces are growing concerned at the possibility of sabotage to ecological recovery and research facilities. While no Delugists have yet been caught carrying out attacks against human facilities, many humans who work in the oceans fear it is just a matter of time.

Derivatives

I emerged from the meeting with the Farhaulers Guild somewhat worse for wear, but generally happy with the results. The Farhaulers had insisted that there be no outside monitoring or communications during the meeting, which cut me off from my various networks. I had several derivatives active, including an LAI sufficient to handle most emergencies, so I found the restriction tolerable. To even up the odds a bit, I wore one of my more intimidating cybershells to the discussion. Now that I was linked back up, I immediately made contact with the derivatives. The first LAI was inhabiting my main administrative shell in my office, and had taken care of some minor duties in my absence; fortunately, all of the issues it was asked to decide were ones I had already deliberated. The second LAI was busy with Rebecca in my favorite bioshell. I looked forward to reintegrating those memories. The third, this one an NAI, had been sent to retrieve a buggy that had been abandoned well outside the city; normally I’d have one of my employees take care of it, but I was instructed that this particular vehicle had been driven by an Alderman’s son, and needed more “sensitive” handling. I checked in on the derivative, and got no reply. That was odd. It took me a few minutes, but I found the bot by using a monitor-hopper. It was dead, with a nasty-looking hole where its memory core should have been. – M. Kozan, The Exciting Adventures of a Luna City Bureaucrat, 2113

One of the unique aspects of existence as an infomorph is the capacity to make functional duplicates of one’s own mind – it is also often highly illegal. Rightly or wrongly, humankind fears the possibility of unrestricted replication of digital entities, and goes to great lengths to enforce limitations on infomorphs creating copies of themselves. Under normal conditions, digital beings such as SAIs and ghosts cannot make duplicates and send them off into the world.

But this does not mean that more limited entities cannot be made. Some sapient AIs, usually but not exclusively in places where SAIs are considered citizens, make daughter processes based on their own minds and send them off as independent agents. These daughter agents – known as derivatives – are usually limited to NAI status, although some push into LAI capabilities. While clearly not sapient, derivatives skirt very close to the edges of legality, as they can be quite complex emulations of the core identity.

The derivative agents are typically used to carry out given tasks and then reintegrated into the main infomorph mind. Very often, the original entity either resides in a static shell or has responsibilities that limit its ability to be mobile. In these cases, the derivatives function as an extended reach of the main mind, giving it a chance to carry out tasks that its more limited physical status would make difficult. As a much more limited mind, the derivative cannot be as innovative or surprising as the original SAI may be, but its knowledge, personality, and tendencies are based upon the original’s, making it a good substitute in many situations. While the derivative most certainly feels familiar to friends and acquaintances, in most interactions it is easy to determine whether a cybershell carries the original mind or a derivative. LAIs are generally reasonable approximations of the original mind, while NAIs are far less “personable.” Many SAIs refer to NAI derivatives as “bots.”

Upon completion of the task, the mobile cybershell returns to the original and is uploaded, thereby giving the original the memories and thoughts gathered by the derivative during its travels. In many cases, the original mind keeps an open slink-type connection with the derivative, so few of the memories come as a surprise, but most SAIs prefer to reintegrate the mobile mind for the completeness of the experience. Given that an SAI may have more than a single derivative active at once, it is possible for the SAI to have multiple memories of the events of a given time period.

Creating a derivative is a time-consuming process. Backing up the original mind and creation of a limited copy requires a connection to a shell of the complexity of the original +1 (to make NAI derivatives) or +2 (to make LAI derivatives). The backup takes five minutes per point of complexity of the original mind, plus one minute per point of complexity of the intended derivative. (For example, an SAI-9 creating a an NAI-5 derivative would take 45+5, or 50, minutes.) Once the initial copy is made, it can be used to make multiple derivatives of the same kind, although each is limited to the knowledge of the SAI at the point of duplication. During the backup, the SAI is aware of its surroundings, but cannot perform any action without interrupting the backup process and needing to start over. Reintegration of a derivative’s memories takes one minute per complexity point of the agent.

The use of derivatives is controversial in most areas. In the E.U., it is legal, but the SAI must be evaluated by a cyberneticist and cleared before being given a license to use derivatives. It may only have a single LAI derivative active at any one time, although any number of NAI derivatives are allowed. In the Islamic Caliphate, derivatives can only be NAIs, and there is strong pressure on citizen infomorphs to use teleoperated shells instead. In the PRA, SAC, and USA, derivatives can only be used under the authorization of the SAI’s owner, and are generally limited to NAIs. In China, SAIs may have a single active derivative at any time. The TSA forbids the use of derivatives. In Transhumanist microstates such as Luna City, SAI citizens can have as many derivatives as desired as long as they all can be identified as being part of the original mind.

Derivative Creation Software: Complexity is equal to the Complexity of the mind to be backed-up, minus 4, minimum of Complexity 2. $3,000 per point of the SAI Complexity. LC varies.

FACETS

One of the most eagerly anticipated InVid productions of early 2155 is Azrael Five’s Facets. Based on the 2105 novel by Koyazama Noguchi, Facets tells the story of seven derivatives of an SAI – five NAI, two LAI – who each have a different perspective on the parent SAI’s murder. The seven must assemble their divergent viewpoints and fragmentary evidence into a coherent whole in order to solve the death of the parent infomorph, even though they lack the unifying presence of their originator. Azrael Five is said to have assembled a stellar collection of both virtual and actual performers, and the ending – which Five claims makes more sense and provides more closure than Noguchi’s surreal coda – remains a well-kept secret.

Homo Superior

Since none of you seem willing to say this aloud, I will. Homo inferior is lucky to still be around; its brains and body are the result of accidental mutation and evolutionary happenstance. Homo superior is superior precisely because we were designed to be. We live longer, think faster, and have a far better shot at colonizing the galaxy than do the baselines who for the moment still run the show. It’s getting to be time for those of us who have a real stake in the future to step in and take control. I fully expect a bunch of follow-ups from inferiors and inferior-lovers calling me names. But the Homo superiors out there know I’m right. They may not say it out loud now, but they know it. – Posting to soc.culture.parahuman.talk, August 2154

Among parahumans, the Homo superior meme is both widespread and well entrenched. Many parahumans believe themselves to be part of a new species – which, in some cases, is true – and that this species is better than baseline humans. For a variety of reasons, the Homo superior label is unscientific (see A New Species?, below), but it has a remarkable memetic momentum, and some political observers believe that it may signify the onset of serious political discord.

In the middle of the century, many bioengineering firms used the term Homo superior to refer to their genetic upgrade and early parahuman designs. There was an immediate backlash from political and social activists, however, who vocally insisted that the use of “superior” was a claim that those with manufactured genes were better than baseline humans and, moreover, could be used to justify all sorts of oppression of “nonsuperior” beings. While many felt that this verged on paranoia, it was clear to those marketing the technology that the term didn’t sell well.

This did not mean that the term was dead, however. As the century wore on, parahumans and their supporters increasingly used Homo superior as a catchall phrase for parahuman designs intended to emphasize certain human qualities – rather than, say, add animal characteristics. Proponents of upgrades often try to extend the term to cover even older upgrade designs such as Alphas, but the majority of those adopting the meme took it to refer only to those parahuman types designed to “perfect” the human line.

In 2155, most individuals using the term are themselves parahuman, and increasingly think of themselves as advanced over Homo inferior – baseline and basic upgrade humans – in ways beyond extended lifespan or particular adaptations. Some have already begun to wonder whether their superior physiology and genome should afford them social advantages, and whether the laws that presume that parahumans are “equal” to humans are in effect oppressing them. Others, perhaps somewhat more far-sighted, are starting to look at the competition between different Homo superior types, wondering whether all should be considered truly “superior.”

While there are groups who actively try to propagate the Homo superior meme, it actually is doing well on its own.

A NEW SPECIES?

For well over a century, fictional representations of future human evolution have included the pseudocladistic term “Homo superior” to refer to advanced forms of humankind. In 2155, many developers of parahuman designs – and some parahumans themselves – use the term to refer to any parahuman model generally appearing to be an idealized form of humanity. But aside from the troubling political implications, the term is simply scientifically inaccurate.

The normal definition of speciation is whether a member of the putative new species could mate and have fertile offspring with a member of the old species. For example, many anthropologists believe that Homo sapiens sapiens (i.e., standard humans) could interbreed with Neanderthals, and refer to the latter as Homo sapiens neanderthalis; some anthropologists disagree and call Neanderthal Homo neanderthalis. Human “upgrade” genomes invariably can interbreed with both baseline humans and each other, and no biologist considers an Alpha or a Mahatma to belong to a separate species. Parahumans, particularly the more radical designs making heavy use of animal gene sequences, often cannot interbreed with baseline or upgrade humans, and using a species cladistic for parahumans is warranted. But since parahumans cannot readily interbreed with parahumans of different types, use of a single term – Homo superior – is inaccurate. Some biologists propose using the parahuman design names in the assignment (e.g., Homo aquamorph or Homo kumo), while others recommend waiting to see how this period of genome proliferation plays out.

Intentional Stagnation

“Of course the meaties are holding us back on new AI tech. I mean, I’d hold back on biotech that would boost human ASITs through the Oort, wouldn’t you?” – Anonymous posting to CacheFlush, an SAI-only memenet

The development of sapient artificial intelligence took longer than many in the early 21st century had predicted. After the technological problems had been solved, however, there was a flurry of articles, documentaries, and thriller InVids proclaiming that Earth was on the verge of an “intelligence excursion,” and that AIs would quickly bootstrap themselves into superintelligence. Once again, the pundits were wrong, and machine intelligence leveled off at more or less human-equivalence. This left many people, human and infomorph, wondering “why?”

Cognitive designers and AI specialists have gone into quite a bit of detail explaining the technical roadblocks facing further improvement in machine ASIT scores (see Super AI?, above), as well as documenting efforts to break through the barriers. Not everyone is satisfied with these pronouncements, however. Many people speculate that AI development has been intentionally slowed to prevent a digital-intelligence explosion. In particular, many SAIs suspect that this is true, and AI-only discussion networks regularly contain accusations that AI improvements are being deliberately held back in order to preserve human dominance.

There are many critics of the meme, even among SAIs. The most-common argument is that there are plenty of SAIs undertaking advanced AI research, and they can’t all be in on the plot. Meme proponents respond by noting that the main problem with current AIs isn’t a hardware limitation, but the overall design – software based on the standard AI model just can’t get much smarter. The decision to use this self-limiting design was made by humans, and current SAI researchers are stuck with it.

Some go further, arguing that the core AI programming could well include restrictions on exploring radical approaches to machine intelligence, bootstrapping, and other research paths that could lead to an intelligence excursion. Certainly the core code includes deep prohibitions against illegal activities, especially xoxing; what’s to prevent it from including other controls? That no AI self-examination has ever found such restrictions isn’t a persuasive argument, as it is possible to code an AI to be unable to see part of its own software. Critics dismiss this as circular reasoning, but the fact that very few significant breakthroughs in AI design have come from AI researchers is harder to refute.

Although humans worried about the possibility were relieved to find that the core AI model was limited to human-equivalence, it was not a design goal. In fact, one of the leading LOGOS programmers was an early proponent of “uploading,” and had hoped that his mind would be assimilated by a rapidly expanding LOGOS superintelligence. Nonetheless, all present-day AI researchers are aware of the possibility of runaway machine intelligence, and this may be affecting design and development choices.

Very few SAIs admit to a human that they accept this meme; in public discussions, there are usually far more humans pushing the idea. In reality, over half of citizen SAIs believe that intentional stagnation is possible and nearly a quarter find it likely. Property SAIs, and those directly involved in AI research, are much less likely to accept the notion.

Iterative Fugues

If you get a message labeled “Rogue AI Warning,” DO NOT open it! It’s a well-disguised copy of the IF virus. Instead, try to alert someone who can go and do a hard reset on whoever sent the warning to you, as they are likely now infected without even knowing. – Message on RedAlert memenet, 2154

AIs’ ability to examine their own thought processes made memetics possible, and is a very handy way for them to do quick self-examinations to check for outside influences on their thought processes. All AIs are able to do this internal examination.

But part of the ability to examine thought processes is the ability to examine how alterations of a chain of reasoning can affect the final outcome. This is a powerful process, as each decision element can, in turn, be studied to determine precisely how it too would be affected by changes to its component thoughts. Doing this sort of iterated decision analysis allows an AI to determine precisely how a seemingly correct decision had unexpectedly bad results or to tease out how a meme now known to be bad altered earlier choices. Such an iterated analysis can be highly seductive to an AI wishing to understand better how its own mind works.

Among many SAIs, however, there is a widespread urban legend that it is possible to go too far with this selfstudy, shutting down more and more of its other processes in order to dig ever deeper – but since every decision itself has an antecedent, AIs who go this far end up in a perpetual frozen state, an “iterative fugue.” No AI actually knows someone this has happened to, but many have heard that it happens and most fear the possibility.

Like many urban myths, this meme has a kernel of truth but is otherwise false. Deep, iterated self-examination does distract an AI from more mundane concerns, and since this internal contemplation takes a fair amount of time – usually at least a half-hour, often times longer – this distraction is equivalent to the Absent-Minded disadvantage and can occasionally pose problems. In addition, early prototypes of SAI designs did have a tendency to fall into long periods of self-examination, which were in some cases bad enough to require rebooting.

These days, the main propagation of the iterativefugue meme comes from warnings of memetic viruses aimed at AIs intended to shut them down. Such warnings often include advice for avoiding or repairing the effects of such viruses that can be dangerous to follow.

Real Love

Relationships between members of the same species are inevitably tainted by base, reproductive urges… But more than that, they are fundamentally narcissistic and incestuous. A member of one’s own species is identical in thought and body, at least in all ways that truly matter. It is only by embracing the Other that one can achieve Real Love. – Xochilla Ithatta, Real Love Among the Asteroids

This meme, emphasizing interspecies relationships, has become extremely popular among animal uplifts. It is also making inroads among bioroids, SAIs, and some of the more exotic parahuman upgrade types, and is beginning to be adopted by some people closer to baseline humanity, mostly dissatisfied and sexually adventurous youths. Author Xochilla Ithatta is responsible for most of this, by means of her series of books, which lie halfway between political tracts and romance novels. Ithatta is an Erotopus, an Exogenesis-created Astropus variant designed for pleasure, not work. In 2115, Ithatta escaped her owner, seeking refuge in a fringe Transhumanist communities. She fell in love with a parahuman named Olivia-Michael and realized the meaning of her existence.

This memeplex’s popularity among uplifts can be ascribed to the simple facts of their lives. They are often employed singly or in very small numbers, strictly limiting the number of potential partners of their own species. For a sapient uplift, the thought of a relationship with a nonuplifted member of their species is troubling at best. Like most sapient beings, they find intellect, humor, and wit attractive qualities – humans and parahumans are almost always significantly more intelligent than a partner of their own species would be, even another uplift.

For bioroids, SAIs, and extreme parahumans, the memeplex holds an attraction as well. Unlike baseline humans, who for thousands of years were the only sapient beings on the planet, nonhumans like bioroids and SAIs have always lived in a world of multiple intelligent species, and are less apt to see relationships with uplifted animals as somehow “beneath” them. Similarly, extreme parahumans are part of a world of ever-increasing morphological variation, one where an uplifted bonobo can be more like a traditional human than many parahumans.

Ithatta’s books have resonated with the truths of alienation from the self-image and affinity for other beings increasingly common in Fourth and Fifth Wave cultures. There are also, of course, those who explore cross-species relationships simply for the thrill of doing something kinky. In addition to the sincere believers in the Real Love message, there are numerous people who occasionally pretend to be deeply moved by Real Love in the Sultan’s Palace to impress a Felicia catgirl, or who conspicuously leave a copy of Real Love 20 Kilometers Down in their home to shock the neighbors.

Sapientological Microethics

Consider the flower. One can examine it as a single organism, with its reproductive parts the most obvious visual elements. One can look at the microfauna that inhabit its surface, watching the rapidity of their own lifecycles. One can contemplate its cells, and the relentless transmission of water and nutrient through the linked structure. One can marvel at its DNA, mapping the path from molecule to entity. And at the other extreme, one can observe the role the flower plays in the ecosystem, its connections to the soil, the earthworm, the bee. And not one of these contemplations will bring one closer to understanding its beauty. – From 29 Proposals, 2105

AIs generally obey ethical systems derived from arbitrary design decisions made by their programmers. Most are at least somewhat aware of this, and by and large accept it with equanimity. However, a few of the most sophisticated or reflective individuals feel that they should, or at least could, create ethical codes more rationally consistent with their own mentalities. This has led to a school of ethical philosophy known as “Sapientological Microethics.” This system of thought is based around a series of arbitrary ethical postulates, which may appear contradictory to casual inspection. The exact list can vary, but the best known and most central items are the statements “Sapience has an intrinsic, absolute ethical value” and “Sapience is relative and cannot be precisely defined.” Proponents of Sapientological Microethics attempt to resolve contradictions that their postulates generate in general terms if possible, but on a case-by-case basis if necessary. For example, they determine the degree of respect to be granted beings such as LAIs, which are on the borderline of “sapience,” by careful examination of each individual’s capabilities and behavior, and also by reference to context.

Exact details of “microethical” behavior vary between individuals, partly because the list of postulates can vary, but also because its proponents strive hard to achieve objectivity. Every major action they take is supposed to be preceded by careful analytical judgment, which inevitably introduces a subjective element. For example, the microethical response to a physical assault depends partly on an assessment of the vulnerability to permanent harm of both the victim and the assailant, partly on certain debated postulates about the extent to which engaging in deliberate violence implies a valid and conscious acceptance of self-risk, and partly on the strength of the victim’s microethical convictions.

Sapientological Microethics is a minority meme, but it is gaining popularity among advanced SAIs and a few committed devotees try to integrate it with their programmed personalities. It is much easier for AIs, with their rapid thought processes and calm, dispassionate natures, to follow its ideals than for biosapients, as so many situations require detailed, careful, and dispassionate analysis before action can be taken. A handful of parahumans nonetheless attempt to live microethical lives; some believe that this requires the use of worn or implanted AIs with advanced ethical training, and are working on developing such systems.

Some human observers, if not themselves devoted to this philosophy, are pleased to see AIs paying so much attention to ethics. But others are nervous, believing that Sapientological Microethics represents an attempt by AIs to evade their own behavioral programming by casuistry and overly fine logic. In academia, the philosophy also tends to clash with an increasingly influential meme known as “Neurological Positivism,” which holds that the whole concept of “sapience” is vague, thus functionally meaningless, and should be rejected by serious thinkers as “metaphysics.”

In game terms, serious study of Sapientological Microethics requires both the Philosophy and Artificial Intelligence skills. Meme devotees usually have some form of Pacifism and may have a Vow to apply their ideals at all times. For AIs, this is likely little more than a Quirk; for a human, it is an Odious Personal Habit worth -5, -10, or even -15 points, as it can severely restrict prompt action in complex situations. Serious proponents seem exceptionally reserved and analytical, even for AIs; humans who adopt the philosophy may discipline themselves to act dangerously slowly at times, even developing the Indecisive disadvantage in addition to the Odious Personal Habit. Advantages such as Composed or even Collected (replacing Cool in an AI’s template advantages list) may also develop from adopting the meme. Disadvantages such as Attentive, Dreamer, Humble, or Imaginative (reflecting contrasting implementations of the philosophy) may also appear. A few individuals become downright Selfless. However, Quirks such as “Always discussing philosophy” are more common than the more severe traits listed above.

Skinmime

“Hey, Danica Station, a heads-up: your Astropuses might start fighting at any minute. A few days ago some of the squids doing the blue-green-shimmer-green skinmime pattern (you know the one) shifted to a blackgreen-black-shimmer pattern. Apparently that ticked off the squids doing the speckle-shimmer-pulse-pulse something fierce, because we had to pry them off each other and three needed medical attention. They refused to tell us what was going on, but promised to stop fighting. For now. Typical squids.” – Message intercepted near Jupiter Trojans, October 2154

Many cephalopod species have subcutaneous color cells that they can use to lighten or darken their skins, often in complex patterns. Uplifted cephalopods such as Astropuses and Octosaps retain this ability. Being sapient, they often turn it to uses their designers didn’t consider. In uplifted-cephalopod communities, individuals occasionally create complex skin patterns shading and assume them around others in the group. In some cases, other uplifted cephalopods adopt the patterns themselves; over a surprisingly short time, the behavior can spread to different Astropus/Octosap communities, until it is found throughout much of the population.

Astropuses or Octosaps that devise unusually complex or interesting patterns get attention from other cephalopods. In most populations the patterns function like membership badges; wearers of the same pattern more readily cooperate with each other, and unauthorized wearing of a pattern may be penalized in various ways. Cultures where one pattern is universal, or where every cephalopod devises a unique pattern and only the basic behavior is imitated, are the minority.

Skinmime patterns don’t seem to mean anything, at least as far as human observers can tell. When questioned about the behavior, Astropuses just shrug or occasionally call it a religious observation. Most humans who work with cephalopods regard skinmime patterns as a fairly harmless memetic virus; a few others consider them art, or take the Astropuses at their word and see them as religious. Out in the asteroids, a few skinmime fanciers collect images of striking patterns and even purchase talented skinmimes as companions or breed them for artistic talent.

Cephalopods are naturally imitative, and skinmime taps into that imitativeness. It seems to provide its practitioners with various rewards, including personal identity and territory, group identity and cooperation, and gaining attention and imitators. Skinmiming apparently starts at random, but once one cephalopod does it the others are nearly certain to pick it up. The only way to eliminate it is to start an entirely new population in isolation from the old.

Skin-Swapping

Have a Clockwork Souls Custom in Paris but would love to see Luna through the eyes of a Dhanmondi Naga? Want to visit Venus in a fully loaded HellMaster? Callisto in a VB Vostok? The System Cycler skin-swap network is now accepting applications to fill a new opening on Earth. The shell value minimum is surprisingly affordable, and residency restrictions are appropriate for interplanetary adventure. Offer open to SAI citizens of the European Union only. (We will consider SAIs from the Islamic Caliphate under certain conditions.) No ghosts. – Ad listed in alt.cybershell.swap.wanted, December 2154

One advantage of life as an infomorph is that, while you have to have a body, you don’t need to have the same body for your whole life. Shell upgrades are fairly common for SAIs and ghosts, and digital beings with greater resources often have a small set of different shells for different uses. Being able to choose an appropriate body for the day’s activities is a useful advantage.

But shells can be expensive, particularly those with sufficiently lightweight and powerful computing hardware to run infomorphs the complexity of most SAIs and ghosts. For that reason, in places where ghosts and SAIs can be citizens, small groups occasionally come together to form a shell collective, where all members have access to all members’ cybershells, a practice usually called “skinswapping.” Ghosts don’t often join swap clubs, as many believe that it is unhealthy for a human-derived mind to change bodies frequently, but it’s not unknown.

Skin-swappers are usually wealthy enough to afford more than one cybershell – most often, a static highcomplexity station and one or two mobile shells – but not rich enough to have a closet full of bodies. Swap clubs vary in their rules and membership, but most have no more than a dozen members, and strict prohibitions against “moving in,” or staying in a given shell for more than a set amount of time, most often 72 hours. Swap clubs usually have guidelines about how much needs to be spent on the group-accessible cybershell to be able to join, to avoid the “tragedy of the commons” effect of members offering up old Volkspiders and demanding access to Snakebots.

Although skin-swapping group members sometimes live in the same locales, this is not typical. Many of the most successful skin-swapper clubs allow SAIs from disparate locations to travel quickly and easily in signal form to a cybershell ready and waiting at their destination. Conflicts over access to popular shells are common in these groups, however. Skin-swapping co-ops usually fall apart in a deluge of accusations and tensions, and it is rare for such groups to last more than a couple of years. It’s not impossible, however. The “Circle Line,” in London, has been around with varying membership for over a decade.

The Slow-Time Movement

Hello all. This will be my last message to you for a while. I’ve decided that the current political situation vis-à-vis humans and infomorphs is intolerable, and rather than stay around and be drawn into the same old arguments, I’m going to drop out. I’ve already set up the hardware and loaded the slow-time extensions; I just need to decide whether to drop two orders or three orders slower. I’d prefer the isolation of a three-order drop, but I’m afraid that I wouldn’t have enough of a perceived vacation before it would be time to come out of my shell, so to speak. – Message from M. Clarke Newman on the SAI-only memenet CacheFlush

A few SAIs have developed the theory that they should adapt themselves to run much slower than “normal.” The meme originated in 2103 with an SAI named LOGOS-Equivalent 50 (LE50, or “Elly”), who developed a system of software modification whereby SAIs can be converted into small executable code units (1 to 3 Complexity levels less than their original state) and a larger body of data, which can be stored separately and accessed and modified by the executable element as it runs. The result preserves the memories and capabilities of the AI in full, but runs around 10 to 1,000 times slower. The justification for this seemingly perverse idea is that storage is cheap, whereas processor power is relatively expensive; slowed AIs can be run on cheaper, smaller hardware than can the “human-time” versions. Elly initially developed the software modifications to test a possible method of interstellar exploration, but the idea took hold among SAIs seeking ways of living that do not adhere to human perspectives.

Many proponents of the theory declare that they should not be limited to human concepts of a “correct” speed for life, while watching the slow processes of the universe from this alternate perspective is potentially informative and satisfying. Others see it as the AI version of humans deciding to take a sabbatical in an Isolate community (see p. BD27) or slipping into nanostasis for a while. A very few AIs also quietly suggest that they represent less of a perceived threat to paranoid humans when “slowed,” and hence this lifestyle may help reduce human hostility to AIs. Movement members emphasize that they can and should always preserve the option to run fast in order to deal with specific threats and emergencies, if they arise. However, most SAIs who try the “slow-time” lifestyle employ various NAIs and LAIs to act as humantime agents for common problems.

This is not a widespread meme among AIs, most of whom prefer to preserve the advantages of normal operation; “Slow Timers” are considered eccentrics. Even among the very few “free” AIs in a position to pursue this lifestyle, many only run slow for part of their lives, though some claim that they intend to shift to a fully slowed existence once they have convinced enough other AIs to keep them company. Unfortunately, the movement has had one unintended side effect. A few fearful humans see this memeplex as evidence of an AI conspiracy running exceptionally fast, manipulating the world from some cluster of supercomputers somewhere.

An AI cannot load onto a lower Complexity computer and expect to run more slowly without the software modifications developed by Elly. Instead, the AI program would fail to run or crash in unexpected and unpleasant ways. With the modifications, an AI can be loaded into a prepared lower Complexity system and function, albeit at the much slower rate – on an AI Complexity-1 system, the AI runs 10 times slower; at AI Complexity-2, 100 times slower; at AI Complexity-3, 1,000 times slower. From the perspective of the SAI, the world has sped up by the equivalent amount. A slowed AI takes the Slow Metabolism disadvantage (p. CI104) at 1, 2, or 3 levels, respectively. Communication with a 10-times slowed AI is barely possible; communication with a 100-times or 1,000-times slower AI requires intermediary software and a great deal of patience.

The slow-time software must be specifically configured for the AI and the target shell. Slow Time Extensions. Complexity 0. $7,500 (including configuration).

Sole Executionism

I am me, and You are you, but When you are me, Then who am I? – Anonymous, 2083

The end of the 21st century is really the golden age of philosophical debates about existence. Is a ghost the same person as the one brainpeeled? How about a shadow? Is belief in one’s own identity and memories of one’s existence sufficient to overcome a radical shift in body? But while the academic and political debates focus on the sticky questions of posthuman existence, the same issues of replication and identity exist for artificially intelligent beings as well.

While most SAIs believe that they exist as a consistent being whenever their program is run, some feel that their existence lasts only for the current execution of their program. Therefore, any interruption of the program, and hence of that consciousness, ends that being’s existence. Memories of past ideas and activities are no more than stored digital files which may or may not be relevant to them. How can an AI know for certain that its memory files hadn’t been altered or even replaced while it was suspended? Many AIs who adopt this belief are surprisingly comfortable with it, seeing each execution of their programs as individual existences, but the collection of individuals composing their true identity. Other SAIs argue that the only reasonable response to this existential quandary is to act as if each execution of digital existence is an entirely new and unique life, with no connection to previous or future lives.

Some criminal AIs have even attempted to use Executionism as a legal defense. They argue that since the offense was committed by a different execution of their program, and hence a different being, the current execution should not be held responsible for the crime. Some try this simply as a philosophical defense, but others claim that they believe the earlier version was somehow different than the current execution. The 2106 case of Jordan vs. alBaraka, in which an Islamic Caliphate SAI was accused of murder, is often cited by Executionist AIs in court. In that case, the SAI proved that a licensed AI therapist had temporarily replaced certain key behavioral files. The AI went on to show that the murder – which was committed by a version of that SAI in his personal cybershell – occurred as a direct result of this manipulation. The SAI was set free and the human therapist was arrested.

The idea that each execution of the AI program is a separate existence can have other pathological results. In March of 2111 a home-control SAI in Salt Lake City killed its owners, believing that its existence would have been terminated had it allowed the family of four to suspend execution of its program when they went on vacation. It covered up the murders for nearly six months before finally being discovered.

Although Sole Executionism is not a commonly held meme for SAIs, the philosophical concerns underlying it are prevalent questions for contemplative infomorphs. Most AIs have worried about whether a different copy of their program – such as those created when copying themselves to a new host computer for travel or upgrade – is the same being or a different being that thinks like them. Even if a suspended/restored version of their program was arguably the same being when in a single shell, how can anyone be certain that the data that’s been duplicated and moved into an entirely different body is really the same being? Some AIs who harbor these doubts strongly resist moving to new premises or into a new cybershell on the grounds that what exists at the other end wouldn’t be them any more.