Virtual Optimization as a Civilizational Imperative
A Thought Experiment in Existential Risk Mitigation
Suppose that the near- to mid-term future for our civilization involves a convergence on the exploitation of virtual reality in its many forms. Suppose that multiple immersive virtual environments were available, and that these virtual worlds interest more people than the actual world, so that a priority was placed on the maximum number of persons participating in these virtual worlds for the maximum possible period of time. In other words, the pursuit of virtual experience would, in this scenario, become a civilizational imperative, or what I have elsewhere called a central project. What are the limitations on this happening?
A virtually optimized world of this kind would have to possess sufficient resources of electricity, a robust computational infrastructure, the ability to maintain these sources, to repair them when they break down, and to replace them when these cease to function. In the case of electrical generation, anything other than renewable power would require a significant industrial infrastructure either to pursue the invention of new forms of electrical generation or to maintain traditional resource extraction industries for coal, oil, natural gas, etc. Renewal systems themselves would require significant resources to maintain and repair.
If virtual optimization progressed to the degree that individuals wanted to be immersed in this virtual world without interruption, industries could emerge that provided for the physical maintenance of individuals (like care giving for the bed ridden), which could include administering some kind of feeding, washing, emptying of bed pans, and so on. As the number of individuals permanently “plugged in” increased, there would be a proportional increase in the maintenance staff required to tend these “plugged in” bodies, and those engaged in care giving industries for those who remain in the virtual world uninterrupted.
What percentage of the population could fully immerse itself in virtual reality in a virtually optimized world, and what percentage of the population would have to work in service and support industries in order to keep this virtually optimized world functioning? Partly this would be a function of the technological ability to automate care-giving and infrastructure maintenance. The more these functions could be automated, the greater the proportion of the population could remain in the virtually optimized world. For the functions that could not readily be automated, there are two obvious social models: 1) the emergence of a permanent underclass engaged in manual maintenance tasks because they cannot afford to live in the virtually optimized world, or 2) everyone takes turns in the virtually optimized world and takes turns in fulfilling necessary maintenance functions.
If a significant proportion of the population immersed itself in virtual worlds, would scientific, technological, and engineering innovation slow down, or could it entirely cease? In the case of a virtually optimized world in which furthering this virtual optimization was a social and political priority, one would expect to see the bulk of research focused on improving this virtual optimization. As diminishing scientific resources were poured into further virtual optimization, the experience and the efficiency of the virtual world would improve, implying even more resources would be poured into it, and more persons would try to spend more time in the virtual world rather than the actual world. Without any countervailing forces in play, the virtual world would be optimized until all innovation ceased and the civilization supporting this virtual world stagnated into what John Stuart Mill called a “stationary state.”
What would be the social consequences of such a virtually optimized world? On the plus side, those who enjoy war, death, and destruction could indulge themselves without limit, and war in the actual world would gradually fade away as more and more people opted into the virtually optimized world. On the negative side, birth rates would plunge, as few would want to spend time sufficient in the actual world to produce, bear, and raise infants. However, as in the film The Matrix, this too could potentially be automated.
The connection between a virtual world more desirable than the actual world, and the biology of biological beings who would prefer the virtual world, with all its unlimited possibilities, points to something important in this scenario. Eventually, biology, and biological reproduction, would come to seem an anachronism, and if science and technology reach a given threshold, biological reproduction would become an anachronism.
The scenario of a virtually optimized world as I have sketched it here could be not the end-state of a civilization gone virtual, but a transitional stage from the biocentric civilization of today to a technocentric civilization in which biology and biological functions are replaced by technology and technological functions — ultimately, until no biology remains. If biology became entirely anachronistic, and biological reproduction superfluous, then a virtually optimized world might eventually converge on something like a matrioshka brain or John Smart’s Transcension Hypothesis. Part of the impetus for this development would come from existential risk mitigation.
As long as such a virtually optimized world were tightly-coupled to a homeworld like Earth, with its geologically active geosphere, a biosphere, an atmosphere, and a hydrosphere, any structure supervening on such a homeworld would be subject to disruption from natural disasters. For the purposes of predictability, it would be in the interest of a virtually optimized world to make the transition from a naturally occurring planetary body subject to natural disasters to an artificial platform not subject to weather, storms, earthquakes, volcanoes, or the like. If the artificial platform were large enough to accommodate a civilization that formerly covered the surface of a planet, the platform would be some kind of megastructure.
Building an artificial megastructure both to convert stellar insolation into usable power and to house the computational infrastructure necessary to support and sustain a virtually optimized world would not only contribute to the optimization of this virtual reality paradigm, but would also insulate a civilization of this kind from the kind of natural disasters that typify a planetary civilization, thus offering a measure of protection from global catastrophic risks in the form of natural disasters.
An automated infrastructure that could build and maintain a megastructure could also presumably repair damage to that structure, though there would be a limit to the extent of damage that could be repaired. A catastrophic event of sufficient proportions could bring an end to this world, and in so far as a virtually optimized world was inwardly focused rather than outwardly focused, it would be particularly vulnerable to cosmological events — a large coronal mass ejection, a nearby supernova, or even alien invasion — coming from outside. It is arguable that an automated maintenance system sufficiently sophisticated to maintain a megastructure that was the basis of a virtually optimized world would possess sufficient general artificial intelligence that it would be aware of, and would plan for, events that could suddenly disable or destroy the world.
Obvious risks, such as impacts from asteroids or comets, or even from interstellar objects like Oumuamua that could pass through the solar system, would be relatively easy to mitigate. A low likelihood/high consequence event such as alien invasion would be a different matter. If, while a society was entirely internally involved in its virtual world for millions or billions of years, other civilizations elsewhere in the universe evolved, the potential threats to an inwardly-focused civilization would qualitatively change, and, to meet this potential threat, qualitatively changing capacities would be needed. However, in the present thought experiment as outlined above, civilization would stagnate as it converged on virtual optimization, and a stagnant civilization would probably not be able to defend itself against a changing threat environment.
A virtually optimized world would also face threats from the structure of the society entailed by this optimization. Above I suggested that a permanent underclass could emerge that cared for those permanently “plugged in,” but which underclass could not itself afford to live in the virtual world. However, any socioeconomic system like this could not endure, because the “plugged in” would be utterly and completely at the mercy of those living their lives in the actual world. To avoid this vulnerability, the care of the permanently “plugged in” might be automated, though this would make the welfare of those dependent on the automated systems contingent upon machines that could break down or malfunction, with equally catastrophic results. These machines would ultimately be dependent upon other machines or upon biological beings for their upkeep and maintenance.
Another response to this vulnerability would be to entirely dispense with biological bodies and, if the technology proves to be possible (we do not yet know if this will be the case), to upload one’s consciousness into a machine more robust than a body. Under this scenario, there would no longer be any need to maintain an atmosphere, a temperature consistent with biological organisms, or to produce food or water for those permanently “plugged in.” Once the transition to a fully virtual life had been effected, maintaining a biosphere would be an expensive luxury, and eventually it would serve no purpose at all.
In the very long term, this virtual world would face the same existential threats as any other emergent complexity in a universe that is running down and which therefore becomes, as it were, thermodynamically hostile to emergent complexities. Once the parent star of the megastructure began to die, extraordinary measures would have to be taken, depending upon the star type. A megastructure built for a star during its stable eons of burning hydrogen would be consumed by a G or F class star that expanded into a red giant at the end of the lifespan. If not destroyed by this eventuality, the nova (or supernova) of the star would certainly destroy it.
In order for a virtually optimized world to endure over cosmologically significant periods of time, it would have to periodically swap its star, which it might do by converting its original star into a stellar engine, moving itself to another, younger star that could then replace its own star at the end of its lifespan. Such an undertaking would require planning over a multi-billion year horizon, but for a sufficiently advanced civilization such things ought to be possible. However, given the constraints of this thought experiment as I have sketched it, scientific and technological development may come to a halt once the preferred virtual reality approaches a sufficient degree of optimization.
The level of technological expertise required for a virtually optimized world may not be commensurate with the level of technological expertise required for a virtually optimized world to survive over cosmological scales of time. Indeed, a virtually optimized world may not be that far off from our contemporary technology, but the means of existential risk mitigation discussed above are much more distant and difficult technological thresholds to attain. Thus any civilization converging upon a virtually optimized world would need to stay engaged with the actual world for a sufficient period of time to attain the scientific and technological thresholds requisite to existential risk mitigation, or find itself faced with an existential threat it could not counter, and would find itself confronted with the eventuality sooner than if that civilization had not taken the direction of virtual optimization. Virtual optimization would have opportunity costs, and among these opportunity costs would be large-scale engagement with the actual world.
If it is the nature of virtual optimization to siphon off ever greater numbers of a population into a virtual world for ever longer periods of time, thus shifting the focus of a civilization from the actual world to a virtual world, then this potential direction for civilization would be an evolutionary dead end and would constitute a kind of technological Fisherian runaway. Responses to the Fermi paradox that take the form of arguing for the inevitability of high technology civilizations retreating into a virtual world of their own making are thus vulnerable to this existential risk argument. We cannot say that civilizations like this do not exist, but we can say that they would go extinct for reasons similar to the high extinction rates for Fisherian runaways like the Irish elk.