The concept of the “singularity” refers to the moment when technological evolution will result in producing an artificial super-intelligence, which will bring about unimaginable changes in human civilization. Some futurologists, such as Ray Kurzweil of Google, even predict that this cosmic transformation will occur in the near future – even before the middle of the 21st century…
In this brief and vague manner, the curator of the edition for the “singularity summit” 1 that took place in Athens in mid-November last year dealt with a handful of issues that are not exactly fashionable when it comes to “general interest.” Artificial intelligence, disappearance/replacement of human labor (including, increasingly, intellectual labor) by robotic machines, the machines’ ability to “self-reproduce” in improved forms: what could inspire awe beneath the title “singularity.”
All these seem “magical” (repulsive or attractive are the two sides of the same coin) to populations that still fear electricity (a technical invention of the 19th century, in widespread use for most of the 20th) or do not know how a common tap works or how it is repaired. And magical they will remain. Not only because the work of the prophet (“futurist” more elegantly) is a job that has always yielded far more than mere survival and increased prestige during unstable and/or transitional periods in the history of human societies: the prophet is, in many respects, a magician who must protect his professional “secrets” (even if they boil down to one: arbitrariness!)… But mainly because the few socially recognized achievements of the third industrial revolution (few within their totality) have already paved the way for mass techno-magic.
Paradoxical or not, the concept of singularity also dates back to the 19th century. James Maxwell, the renowned Scottish physicist (among other things, known for formulating the classical theory of electromagnetic radiation), named “singularity” the moment in a dynamic system when a small change in a factor causes a (disproportionally) large result. Maxwell was referring to non-linear, not strictly deterministic systems. The considerably later chaos theory has dealt with the same phenomenon: the cascading major changes that can be triggered by a small (often imperceptible) cause. Every time you boil water, you witness a singularity: it is the moment when a very small increase in the water’s temperature changes its physical state, turning it from liquid to gas.
In social “theories” and “sciences”, as well as in control practices, chaos theory (or, if you prefer the more refined term, singularity) has long since made its appearance: societies are considered non-linear, dynamic “systems” in which cause-and-effect relationships are not always clear. And where a “minor trigger” can (at least theoretically) cause intense social “oscillations”.
In the technologists’ compartmentalized world, the basic idea of singularity seems to have entered sideways in the mid-1960s. It was again a British mathematician, Irving John Good, who formulated the theoretical prediction of the “intelligence explosion” of machines. 2
In a 90-page monograph in 1965 titled Speculations Concerning the First Ultraintelligent Machine 3 Good begins as follows:
The survival of mankind depends upon the early construction of an ultra-intelligent machine.
In order to design an ultra-intelligent machine we need to understand more about the human brain or human thought or both. In the pages that follow I make an attempt to remove as much magic as possible from the brain issue, via the theory of “subassembly,” which is a modification of the famous Hebb theory of [neural] cell association. 4… The subassembly theory illuminates the physical embodiment of memory and meaning, and there is no doubt that both need embodiment in an ultra-intelligent machine.
…
Let us define an ultra-intelligent machine as one that can far surpass all the intellectual activities of any man however clever. Since the design of machines is one of these intellectual activities, an ultra-intelligent machine could design even better machines; consequently, there would then unquestionably be an “intelligence explosion,” and the human intelligence would be left far behind. Thus the first ultra-intelligent machine would be the last invention that man would ever need to make, provided that the machine would be sufficiently obedient to tell us how to keep it under control….
Despite this fantastic speculation, Goof’s interest in his monograph was not the end of human intelligence, but rather the study of neural synapses and functions in the human brain and their mechanical replication. The technologists of “artificial intelligence” followed this path for a while, until they relativized its value—discovering other logics for the “development” of machine “intelligence.”
Although he did not use the term “singularity,” his definition of a “super-intelligent” machine (such as one that can self-replicate and self-evolve) became the basis for applying the theory of dynamic systems to technological evolution. Primarily, it gave rise to the somewhat far-fetched (yet mysterious) idea of a Super-Revelation, in the form of machine dominance over the planet.
In 1993, the mathematician Vernor Vinge, in a lecture at a NASA conference, after having “worked on” and promoted this specific idea of Good’s for a decade, made his own prophecies: 5
Within the next thirty years we will have the technological means to create superhuman intelligence. Shortly after, the human era will end… The acceleration of technological progress is the central characteristic of this century. I argue… that we are at the beginning of a change comparable to the evolution of human life on the planet. The precise cause of this change is the impending technological creation of entities with intelligence higher than human…
…
I think it is fair to call this event a singularity (\”the Singularity\” for the purposes of this exposition). It is the point where all our old models will become obsolete and a new reality will prevail. As we get closer and closer to this point, this condition will become increasingly involved in human affairs, until its reality becomes commonplace. However, when it is completed, it will be a great surprise and an even more extensive unknown landscape.
…
In the 1960s, I.J. Good wrote…
The cube of the mythology of the upcoming Singularity (with a capital first letter…) had been cast! It was 1993, after all, the beginning of an era of the “end of history” and definitive victory of neoliberalism on the planet. Amidst all the pleasant and blissful announcements made by most prophets, there was also a need for something ominous. Supporters of maintaining and strengthening the American nuclear arsenal (which seemed useless now after the collapse of the hated nuclear adversary, the U.S.S.R.) were looking for new reasons for legitimization, new threats. And they found them in space: shouldn’t a threatening meteorite, which could destroy life on Earth, be blown up far from the planet with ballistic missiles loaded with nuclear warheads? It should, and obviously American nukes should not only be maintained but also evolved—for the good of humanity.
And what about “smart machines”? Would they become devilishly smart (and, additionally, devilishly self-sufficient…) enough to threaten our already inferior species? Some answered warningly “yes” in the early ’90s. And, meanwhile, they organized better.
The myth of Frankenstein was revived in (technological, initially) relevance after nearly 2 centuries. Only now it’s not the magic “that gives life to the lifeless.” It’s the technology that gives them intelligence…
clever tricks…
One way (although not the preferred one, especially for those inclined toward navel-gazing) to explore the theorems of the singularity and its prophets is to investigate what intelligence is. Any reasonable person would quickly agree that the word “intelligence” has a relative meaning, and even more so—not when comparing different species, but only within our own species and its history. Depending on the social division of labor and the evaluations that accompany it, the intelligence of a sculptor, a blacksmith, and a prostitute are quite different from one another. Even more so if the presence (or absence) of intelligence is judged across the entirety of daily life. A card player is intelligent if he manages to cheat his fellow players without getting caught. A housewife is intelligent if she manages to make ends meet at home using minimal available resources. According to the Western perception, a general is intelligent if he manages to encircle and neutralize the enemy army. According to Sun Tzu’s view, a general is intelligent if he manages to overcome his opponent without battle. A dog is smart when it understands very well the gestures (even the emotional state) of its human companions. A cat is intelligent when it ignores them. A child who has grown up in the countryside is smart if he manages not to get lost even in unknown forests. A 21st-century city child is considered smart if he quickly pushes the buttons of a gaming machine. And so on.
These are commonplaces. Yet they could originate from (or end up in) an ontology of intelligence; and, ultimately, leave a margin for ascribing intelligence to machines. This is wrong. Outside technological mysticism, talk of machines’ cleverness would be laughable, just as it is today as it was in the time of Hero of Alexandria and his impressive hydraulic automatons. 6
We propose a different approach. Starting not from intelligence (with its relativity) but from the machine. What is the machine? More specifically: what is the machine under capitalism? The answer to this question is fundamental. Because if the machine is “endowed” with properties of living organisms, then the validity and characteristics of this endowment arise from the motives of those who make it, and not from abstract ideas about the “endowment” itself.
For example, in the first wave of capitalist mechanization, that is, the construction of machines designed for specific reasons and purposes, the machine appears as “omnipotent.” The basic characteristic of the steam engine (and all the peripheral machines connected to it) is power. But what kind of “power” are we talking about? The answer seems easy, but not so simple: it is, of course, muscular power (of humans and animals). The steam engine appears on the scene as stronger than horses when it pulls loaded wagons behind it. It appears stronger than female workers when it operates, as its appendage, a mechanical spinning machine in a textile factory. In other words, the machine initially appears not as “superhumanly intelligent” (as the prophets of the singularity announce the 4th wave of mechanization) but as “superhumanly powerful.”
However, this characterization of the machine as “superhumanly powerful” is not the first and last word for that historical period. Relatively early, at the outset of the first industrial revolution (in England), Andrew Ure writes and delivers an iconic 240-page work titled The Philosophy of Manufactures or an exposition of the scientific, moral, and commercial economy of the factory system in Great Britain. It is a hymn to the machine and its “superhuman power”! A hymn in favor of the generalization of machine use by industrialists, in order to overthrow, undermine, and ultimately break worker resistance to employer demands and norms.
His book is full of stories showing how collaboration between business owners and engineers can result in machines of such power (and efficiency) that those craftsmen trying to curb employers’ productivity demands are cast aside. And as he laconically praises: “…When capital enlists the science and its services, the insolent hand of labor is always crushed and becomes obedient!”
Ure, in 1835, was enthusiastic about the “power” of the machine; not, however, simply “mechanical power” in relation to “muscle power.” He was enthusiastic about the political power of the machine! Clearly, this is a concept of the word “power” entirely different from that which appeared in the cooperation of mechanical parts. Another kind of definition for “power”; a “schism” in the notion of “power” as “muscular” (hence humans and horses were placed on the same conceptual level). And Ure was not mistaken. Particularly enthusiastic about the case of the inventive engineer Mr. Roberts, who undertook the design and construction of a “spinning automaton” on behalf of owners of cotton mills struck by a strike in Lancashire and Lanarkshire, Ure positioned (and meant) the machine not measured in newton units, but as a weapon.
According to Ure, Mr. Roberts managed within a few months to build:
“… The Iron Man, as the artisans have named him, who sprang from the hands of the modern Prometheus with an order from Athena – a creation destined to restore the sovereignty of the industrial classes and to confirm Great Britain as the sovereign of technologies. The news of this Herculean miracle would spread terror among the syndicates. And so, before he could even leave his cradle, so to speak, this mechanical Hercules had already strangled the Lernaean Hydra of rebellion…” 7
Ure could have been lyrical; however, not excessive for the feeling that began to be aroused from the middle of the 19th century onwards by the “superhuman power” of the machine. George Kaffetzis notes on this matter (the emphasis through spacing is ours): 8
… The theme of machines becoming the regulators of social existence generally captivated the European imagination from the 1860s onwards. And not only the European. For example, in 1863, Samuel Butler wrote and published “Darwin among the Machines” in New Zealand. He argued (ironically exploiting and subsequently reinforcing the equation machines = capital) that machines should be considered as the next step in evolution, beyond the human species. The problem with this evolution was: what would be the best human stance? Resistance or cooperation?
The conclusion of the resistance version appeared later in “Erewhon” (1872) where he describes a society that destroys all machines through a terrifying civil war that erupts after the publication of a prophetic article titled “The Book of Machines.” The book’s claim was that machines are quickly becoming the masters of the human species, and that if they are not destroyed (at the cost of tremendous pain that such destruction would cause), the human species will be annihilated or completely subjugated to them.It is no surprise that “Erewhon” was largely written during the Franco-Prussian War and the massacre of the Communards in insurgent Paris. Indeed, at the end of his utopian novel, Butler has an Italian captain save the hero from drowning, simply assuming that he has escaped from the siege of Paris.
Butler’s half-satirical disposition confirms the ambivalence of discussions concerning machines, as they had erupted among the bourgeoisie of the mid-19th century. On one hand, the metaphorical identification of science with capital stemmed from a watered-down “moderate” influence of workers’ demands. But, on the other hand, this identification had gradually lost any class reference, and had begun to appear as an “alien” force whose threat reached even the bourgeoisie itself…
The “machines that are the next step in evolution beyond man”… The “machines that enslave the human race”… The similarity of observations between 1860 and 2010 regarding “machine capabilities” should not be surprising! A time span of at least 150 years is not insignificant, is it? Moreover, if one takes into account that in 1860 there was no issue whatsoever of “mechanical intelligence,” neither superhuman nor subhuman, but only a certain representation/conception of the “superhuman power” of machines at the time, it can be confirmed that the serious issue is not primarily at all what is attributed to machines (power? intelligence? sensitivity?), but rather how and why this “endowment” occurs. How and why this (ideological, symbolic) “investment” takes place—a process that is absolutely human (and capitalist)!
Indeed, before the most recent wave of such anthropomorphization (with “intelligence”), machines had already transformed other concepts that had once been human attributes—turning them into mechanical ones. The “work” (a concept from physics) of machines, from many perspectives, stole the term “labor”—leading early on to fantastical predictions of a (capitalist) world where only machines would work—this in the second half of the 19th century. Or “energy,” also a concept from physics, ceased to refer to humans and became a mechanical requirement.
Just as happened during the 3rd industrial revolution with “memory.” “Memory” is no longer conceived of as an exclusive characteristic of living organisms. Gradually, it isn’t even their own. It is a mechanical property, and it is measured. In giga, tera units…
chapter…
Machines, first and foremost, are fixed capital. The human-made qualifiers “powerful,” “fast,” “intelligent,” “conscious,” etc., cannot be understood unless they are placed within this political framework: fixed capital. Anything else leads to pure ideology and metaphysics.
It is within this framework that the mechanical “superhuman power” of steam engines, that striking emblem of the first industrial revolution, can be properly analyzed, including its limitations. Because, as we have already said, as fixed capital, the steam engine was a political force – together with the substitution of muscular power: two different concepts of “power,” with the first overriding the second. On the other hand, this mechanical “superhuman power” was neither mental nor emotional. Steam engines were not “courageous”! They had nothing even approaching what was called “human endurance”: in practice, a machine with any “superhuman power” would have worn out/rusted many times over within the average lifespan of a person in the second half of the 19th century, that is, within 50 or 60 years. Which means that other notions of “power,” which at the time lay outside capitalist definitions and orientations for strangling the Lernaean Hydra of (worker) resistance, were completely unrelated to mechanical “superhuman power.” Generally speaking, this mechanical “superhuman power” was, essentially, powerless – no matter how much it was vaunted for its omnipotence. This was proven a few decades later (an insignificant historical timespan) through worker sabotage and strikes by unskilled/semi-skilled operators of these all-powerful machines…
Nothing changes if instead of “superhuman power” the machines are attributed with “superhuman intelligence”. It is still about fixed capital, still specific capitalist determinations and orientations must be served. They may be different if the “endowment” writes on its packaging “intelligence”, “artificial intelligence” and not “power”. In some points, no.
Two issues (at least two) arise, therefore, having in front of us the rhetoric of the priests of singularity and others alike. First, what could be the fields of utilization of the “hyper-intelligent” fixed capital? And second, is it ever possible for the “hyper-intelligent” fixed capital to start “self-utilizing” without (or even against) the human substance?
In the second, our categorical answer is “No”! It was not possible in the “super-strong” fixed capital. It is not possible even in the “super-intelligent”. It will not be possible even in the “super-sensitive” (e.g.: sensors). The phantom of “capitalist self-exploitation” deserted or even contrary to the human condition (e.g. work, although now we are talking about the exploitation of many more human conditions, behaviors, activities, etc.) is old – we brought earlier a small sample from the second half of the 19th century.
Machines (regardless of the type and capitalist determinations with which they have been invested) cannot even “self-design”, nor can they “self-construct”! Because, in the “theoretical” case that something like this would be “technically feasible” it would result in machines of zero value! The “value”, that is, of fixed capital (which is one of the basic forms of capital in general) would be nullified, since it would lose its only measure, human labor (for its construction and maintenance as fixed capital). It is exactly the same thing that would happen if the dream of the alchemists of the Middle Ages for the conversion of all metals into gold became reality: gold would lose all “value”! Just as gold has “value” only in its rarity, in relation to most non-rare metals, so fixed capital has value only in relation to variable capital! Labor – and now, not only that.
The idea, then, that the “super-intelligent” fixed capital “designs” (and gives material form to…) an “even more super-intelligent” fixed capital, and so on, is technologically fetishistic (something that doesn’t impress us at all). It is so because it strips the fixed capital (“smart” or “dumb”: take your pick… – it doesn’t matter) of its political dimension. It prematurely removes it from capitalist norms and necessities; and subsequently, “liberated” from capitalist political economy (that is: the exploitation of human characteristics, commodification, the realization of surplus value in the market, and again from the beginning…) it attributes to it magical, “divine” capabilities! Exactly as Ure did in his literary (though extremely polemical!) apotheosis of the “Iron Man”!
So, will these “super-“, and then “super-super-“, and then “super-super-super-intelligent” machines amuse themselves by admiring each other’s wonderful services? Or will they become so twistedly smart as to keep some human population under their control, in a state of inertia, to which they will occasionally throw some money to maintain the commodity cycle – in the air? Perhaps, ultimately, they will design quite a few “not-so-super-intelligent” machines to extract and process raw materials, to manage the “dumb” energy sources, so that all the super-intelligence of fixed capital in the 4th industrial revolution will not remain “on paper”, as devilishly smart algorithms, but will acquire volume, dimensions, “physical” presence – and self-sufficient exclusive existence?
We do not judge, therefore, the idea of “intelligent mechanical self-design” from the perspective of its technical feasibility; but from the basic perspective of the capitalist determinations it must serve. Unless – we haven’t heard this until now but you never know! – enthusiasts of the idea appear that capitalist machines will destroy capitalism! Fixed capital against capital overall!!! Hmmm…
The first issue (which we formulated earlier) is equally interesting. Because it involves the disenchantment of the characteristics “intelligence” or/and “smartness” when attributed to forms of fixed capital, that is, to machines. What are these machines that are proclaimed as “smart”?
These are machines that perform calculations; computational machines. The complexity of the calculations is certainly something that can impress anyone who struggles with the four basic arithmetic operations – however, complexity does not change the computational function at all. The model of these “intelligent” machines has already been shaped (initially theoretically) since the 1930s, with the Universal Turing Machine (UTM). Therefore, regardless of the magical ways in which the “results” of these calculations are externalized, what is called (mechanical) “intelligence” is nothing more and nothing less than the conversion of manual or intellectual activities into sequences (however “complex”) of simple, elementary 0/1 dipoles.
If there is something truly intelligent in the conception of the UTM, it is entirely human (and entirely Turing!): it concerns the analytical translation of what was once called reasoning and its conversion into calculation. In the development of the third industrial revolution, it was proven that this “transformation” had many possibilities and mechanical applications.
We are always talking, however, about fixed capital—and not about “mysteries” or “divine miracles.” The transfer of activities that were previously non-mechanized into machines, or put differently, the convincing mechanical simulation of activities that historically existed without machine mediation, is a classic process of capitalist expansion. What is called “creation of new markets” on the one hand, and “removal of capabilities from the living and corresponding increase in mechanical capabilities” on the other. In short: subsumption of aspects of “variable capital” (first labor and subsequently the greater part of life) under fixed capital.
The real evolution of “artificial intelligence” is therefore not a technical issue, as the fetishists and priests of the 4th industrial revolution want us to believe. It is political: it advances not simply as an enrichment of computational capabilities, but—mainly—as a subtraction/expropriation (and, in a particular way, a transformation) of human abilities. The “hyper-intelligent” fixed capital reveals itself as such only under the condition of a corresponding “dumbing down” of the variable capital—that is, of us. Just as the “superhumanly powerful” machine progressively stripped our species of its physical/social muscular strength, so too does the “superhumanly intelligent” machine already strip it (even while still in its smart phase…) of its physical/social adaptability. And not only that ability, by the way. 9
mechanization of variable capital…
“… We must only discover the laws of nature, and then man will no longer be responsible for his actions—his life will become excessively easy… Afterwards… new economic relations will be established, everything will be done and calculated with mathematical precision, so that any possible question will vanish as if by magic, with a tremor, simply because every possible answer will be available. Then the Crystal Palace will be inaugurated…” 10
Within this process that has already spanned nearly three decades of the 3rd industrial revolution and continues into the 4th, what is the purpose of proclamations about the impending singularity? Isn’t the prophecy of the (wherever it may be) arrival of a moment when machines are clearly smarter than our species extremely threatening, essentially subjugating us? Doesn’t the idea of the “exponential development” (“self-development”!) of artificial intelligence remind us of something apocalyptic, like the chain reaction leading to nuclear fission and explosion?
Exactly!!! In contrast, however, to the concern of 19th-century urbanites about the potential danger that machines of that time might “run amok” and destroy humanity (a humorous fear then as much as now…), who had no countermeasures except romanticism and the dream of escape / return to nature, today’s prophets of the coming singularity “Revelation” have their “proposals for salvation” ready.
Which have been formulated by an extremely familiar “religion,” called human plus. Salvation lies in the following formula: since information machines tend to surpass us, and since we cannot expect anything from our natural evolution as a species, we must evolve technologically. So that we can continue to interact with the “super-intelligent” machines, and (therefore) control them. 11
The technologically engineered capitalist “Second Coming” is not satisfied with a verbal confession of faith to offer salvation! It requires the adaptation of variable capital to fixed capital. In Greek, this is called “user interface” (ui) and so far it is done in ways “external” to human anatomy.
But no, this is not enough anymore. “Augmented reality” requires technical / mechanical enhancement of the human-machine “correlation”. Not only from one side of the pair but also from the other. And the enhancement of the human being can now (and must) be done only by mechanical means. Otherwise? Otherwise, mechanical “super-intelligence” will cast humanity into the Tartarus of capitalist prehistory.
Full and unquestioning incorporation of parts of the (new) fixed capital; quickly and nimbly: this is what the priests of the singularity say through their teeth…
Ziggy Stardust
cyborg #14 – 2/2019
- John Palaiologos, in the introduction of the special 34-page insert edition of the newspaper “Kathimerini”, on September 23, 2018. The “uniqueness conference” in Athens took place on November 19 and 20, 2018. ↩︎
- Irving John Good (born in 1916 as Isadore Jacob Gudak, from a family of Polish-Jewish immigrants to England) was a close collaborator of Alan Turing, as a cryptography specialist, during World War II. He continued to collaborate with Turing until his suicide, and subsequently moved to the USA. He collaborated with director Stanley Kubrick in 1968, on issues related to artificial intelligence, during the filming of 2001: A Space Odyssey. ↩︎
- Speculations Concerning the First Ultraintelligent Machine. ↩︎
- The Canadian neurologist Donald Hebb, in a book of his (1949) titled “The Organization of Behavior,” formulated a theory regarding the plasticity of neuronal synapses in relation to learning. He is considered the “father” of neuropsychology and the theory of neural networks.
He participated in a series of experiments funded by the CIA to achieve “brainwashing”… ↩︎ - The lecture available at https://mindstalk.net/vinge/vinge-sing.html ↩︎
- From 300 to 230 B.C. ↩︎
- From “Why Machines Cannot Create Value,” by George Kaffetzis, 1997, published by “antischolio“. ↩︎
- Wherever and previously. ↩︎
- More specifically and from a genealogical perspective, in “The Mechanization of Thought”, 3rd notebook for worker use, Sarajevo. ↩︎
- Fyodor Dostoevsky, Notes from the underground, 1864. It is possible that the reference to the “Crystal Palace” as the end of humanity is related to the tremendous impression made on all of Europe by the great industrial exhibitions in various capitals, starting with the exhibitions of 1851 and 1862 at the Crystal Palace in London. The “crystal palace” was a huge exhibition space built of steel and glass in Hyde Park, aiming to showcase the impressive “iron men” of the first industrial revolution. ↩︎
- A large part of the analyses in cyborg relate to this topic. Indicatively:
Cyborg 1 (October 2014): the possibilities of post-humanism.
Cyborg 4 (October 2015): Human plus (human, very human).
Cyborg 8 (February 2017): the Turing test; notes on a genealogy of “intelligence”.
Cyborg 10 (October 2017): wearable, wearable, subcutaneous: the body as motherboard.
Cyborg 13 (October 2018): not only artificial intelligence…but also artificial anoia. ↩︎