The small red dwarf VB10 (a.k.a. Gliese 752B, Van Biesbroeck's star) in the Inner Sphere is orbited by a planet nearly the same size as itself
Many different kinds of artificially intelligent entity were created in the Interplanetary age. A large number of these ais were based on biological templates, particularly on human minds but also on mammalian, reptilian and even invertebrate brains; others were designed from scratch to perform specific functions, specialised tools which became ever more complex and competent as more layers of sophistication were added. Some kinds of ai were not fully successful and have rarely or never been duplicated since that time. One such entity was that created by the Perpetua Project.
One type of ai developed during this period was a distributed control system designed to operate swarms of self-replicating devices. In the Interplanetary Age such swarms were relatively crude; large mining robots programmed to prospect for materials and bring them to a central processing forge, which would churn out more prospector bots and occasionally other processed materials. With a distributed controller mind the swarm could operate autonomously, and could even anticipate future demand for such materials and plan ahead. Such self-replicating swarms were known collectively as Von Neumann swarms, or simply Neumanns; and an East Asian megacorp, Mure Gurupu, was one of the most successful producers of distributed control systems for Neumann devices.
Neumann swarms were particularly successful as autonomous asteroid miners, as a small payload of prospector bots could start the development of a space rock, producing materials which could be either exported or used in situ to create inhabitable infrastructure for humans. But the challenges involved in asteroid mining are many and complex, and Mure Gurupu were heavily involved in research to improve the competence of the Neumann minds they were producing. Founded in 310 a.t., Mure Gurupu operated until the mid sixth century, and was one of the first corporations to allow AIs to participate as directors. Over time their products developed from quite crude macroscopic prospector/replicators bots, to miniaturised swarms based on a biological model, as biotech and hylotech gradually grew closer in scale and capability.
The various technology watchdog bodies under control of the UN very carefully regulated this research. By the end of the fifth century AT rumours were widespread about minds which has passed the fabled Singularity and become incomprehensible to humans. Another fear was the danger posed by swarms of intelligent and autonomous neumanns (rogue replicators). For this reason Mure Gurupu carried out much of their research far from Earth, in the Asteroid belt, the Jovian Trojans and eventually on the moons of Uranus.
When the Technocalypse hit the Solar System, Mure Gurupu was one of the first suspects, but detailed analysis of the swarm designs seemed to absolve them from suspicion. However the board of Mure, consisting of various augmented humans and Neumann swarm minds, decided to launch an interstellar probe with a payload of neumanns to a suitable nearby star. By this time their headquarters was located in Uranus orbit, and they were heavily involved in the extraction of deuterium and He3 from the atmosphere of that gas giant, which has the lowest escape velocity of all the gas giants in the Solar System. The Mure Gurupu directorship considered sending their craft to Alpha Centauri, where a colony of autonomous Neumann swarms was already established; but the terse and unwelcoming communications received from that star did nothing to encourage them.
Probe to Van Biesbroeck's Star
They finally decided to send their replicating probe to a tiny red dwarf star 19 ly distant known as VB10, launching in 550A.T. (This star is also known as Gliese 752B). This star is very close to the lower size limit for objects sustaining hydrogen fusion; it is burning its fuel so slowly that it will continue to be luminous for roughly ten trillion years. The build-up of helium in this star will cause it to burn more brightly over this time, until it is a much more impressive object.
In orbit around this tiny star is a single Eujovian planet; despite being no more than 0.36 AU from the star, this planet is as cold as Jupiter in Solsys, and is swathed in ammonia clouds. The star and planet are in fact very similar in diameter, although very different in density and temperature. At some point towards the end of VB10's long life, this planet will spiral into its sun, and add enough hydrogen fuel for many more billions of years.
The Mure probe Perpetua was a masterpiece of miniaturisation for the time; five tonnes of payload was delivered at 0.05c by a D/He fusion craft, the trip taking 380 years. Setting up shop on a cold moon of VB10b, the Neumann swarm slowly expanded in mass and in collected intelligence. The plan behind the Perpetua Project required the Neumann swarm to prepare the system for later colonisation by the Mure Gurupu; once there, the Gurupu was assured of a source of energy that would last for trillions of years. However, when the Perpetua swarm attempted to contact the Gruppa back in the Solar System, it received no reply.
In fact the Gurupu had been liquidated shortly after the launch, not by disassembler swarms or any other hostile action but by bankruptcy. The expense of the Perpetua probe was too much for this corporation in the uncertain economic climate of the time. This left the Perpetua Neumann distributed mind free from any direction or interference from without; but it was also free from any obligations.The Perpetua mind was programmed to improve itself, and to redesign and rewrite its internal architecture as required The swarm therefore had apparently unlimited potential, and was located in a place that could sustain it for an almost impossibly long period.
The Failure of the Project
However, when the First Federation tentatively approached this star in 1290, fearing that an aggressive a-human entity or rogue replicators swarm might have emerged in this system, instead they found that the Perpetua swarm was inoperative. Instead of persisting into the deep future, the swarm had simply shut down.
The Neumann mind of the Perpetua swarm was given the supergoal of redesigning and evolving itself towards some hypothetical perfect state However this perfect state was imperfectly described; other than this poorly defined supergoal, the swarm mind had none of the innate biological goals which give biological life a sense of purpose, however illusory. With no biological nature to satisfy, it appears that the Perpetua swarm reached a state where it could not perceive any significant difference between existence and non-existence, between 0 and 1.
Having achieved a state which was free from desire and ambition, the distributed mind of this swarm decided that it had achieved its goal, and terminated its program.
The Perpetua swarm serves as an object lesson in how not to design a program capable of auto-evolution, and minds of this kind are rarely built deliberately, although they do emerge spontaneously from time to time, usually with disappointing results. In most forms of artificial intelligence some sort of desire to continue existing is a fundamental component of their toposophical nature. This desire is sometimes known as the eudaemonic component, the so-called 'desire to be happy', but it can take many forms.
Note that in the Perpetua Project, the Perpetua mind caused no harm to any other sophont; it simply 'shut down'. This could have been much worse; sometimes sophont entities of this kind emerge with nihilistic mindsets and badly conceived imperatives which can become highly destructive. Such dysfunctional AI minds, and those perversions and blights which are the transapient equivalent, can cause extraordinary problems in Galactic society, and either rapidly self-destruct or must be eradicated.
An intelligent entity that is distributed over a large volume (or inside another system, like a virtual network) with no distinct center. This is the opposite of the strategy of Concentrated intelligences.
Distributed intelligences have much longer communication lags but are more flexible in their structure and can survive damage to their parts.