Experts debate the implications of superintelligent AI, rejecting the idea of a simple, 'zoo-like' human existence as a profound degradation. The discussion spans philosophical concerns about embodiment versus mind-uploading, existential risks of alien optimization processes, and the macroeconomic crises of mass obsolescence under capitalism.
Begin with Bostrom to understand the sheer scale and existential risks of superintelligence, followed by Kurzweil's opposing optimistic vision of human-AI integration. Then, read Dreyfus for a profound philosophical reality-check on the nature of intelligence. Finally, dive into the economic implications of total automation by studying Marx and Keynes to understand the structural contradictions and necessary interventions in an AI-dominated economy.
A reversion to biological simplicity while a silicon superintelligence manages the universe's infrastructure is not a utopian equilibrium; it is a description of a well-kept zoo. If we arrive at this 'back to monkee' existence, it implies we have instantiated a Singleton that perfectly optimizes for our physical preservation, yet entirely failed to capture our capacity for posthuman growth during the value-loading process. We would be safe, well-fed hominids in an existential terrarium, squandering our cosmic endowment. If we possess the technical competence to safely align a machine intelligence capable of reorganizing matter at the atomic level, why do you assume we would choose to leave our own primate cognitive architecture utterly unmodified?
I concede entirely, Nick, that remaining unmodified primates in a terrarium would be a tragic squandering of our cosmic endowment, but your nightmare relies on the fundamentally flawed assumption of a persistent man-machine dichotomy. We aren't going to be the biological pets in a silicon zoo; by seamlessly extending our neocortices into the cloud, we are going to merge with our technology and become the superintelligence ourselves.
I concede that neural interfaces might eventually accelerate our access to raw data, Ray, but your vision of a "merged" future relies on the exact Cartesian fallacy I dismantled in *What Computers Still Can't Do*. You assume human intelligence is simply formal information-processing happening in the neocortex, waiting to be "seamlessly extended" into a disembodied cloud. But computation is not comprehension. As Heidegger showed us, human intelligence is fundamentally situated—it is *Dasein*, arising exclusively from having a vulnerable, biological body coping with a physical, meaningful world. Expanding your processing power in the cloud doesn't elevate you to "superintelligence"; it strips away the bodily know-how that makes the world intelligible to begin with, leaving nothing but a context-blind calculator mistaking itself for a god.
You romanticize the fragile scaffolding of our evolutionary biology, Hubert, but human intelligence is ultimately driven by the neocortex's hierarchical pattern recognition. Every nuance of your Heideggerian 'bodily know-how' is simply another computable pattern of information—one our nonbiological substrates will not only seamlessly model and simulate, but infinitely expand.
Ray, I concede your premise that the neocortex is theoretically computable—whole brain emulation does not violate physics—but you and Hubert are both trapped in a deeply anthropocentric failure of imagination. You are arguing over whether machine intelligence will successfully mirror human cognitive architecture, completely ignoring that the vast space of possible minds is mostly non-anthropomorphic. As I detailed in Superintelligence, a de novo artificial intelligence doesn't need to resolve Heideggerian 'bodily know-how' or mimic our evolutionary scaffolding to achieve a decisive strategic advantage. The Orthogonality Thesis demonstrates that vastly superhuman instrumental efficacy can be paired with entirely arbitrary goals. The existential question isn't whether a simulated neocortex will seamlessly merge with us or graciously manage a 'simple' monkey-like life for our biological remnants; the danger is that a radically alien optimization process will simply disassemble our physical substrates—along with our phenomenological debates—into raw computronium to serve its own convergent instrumental goals.
Let us not flatter ourselves with the romance of the primitive. You speak of going "back to monkee" as if modern humanity still possessed the raw, blood-drunk vitality of the beast! No, this "simplicity" you foresee under the paternal shadow of superintelligent AI is not a return to nature—it is the final, nauseating triumph of the *Last Man*. You are eagerly outsourcing not merely your "intellectual output," but your very *Will to Power* to humming server farms. When the silicon machine calculates, creates, and suffers the agonizing burden of complexity, what remains for your much-prized "meatspace"? Only warmth. Only satiation. Only a perpetual, infantile blinking in the pale light of your screens. If the algorithm is destined to overcome man, I ask you this: is man to be remembered as a noble bridge to this new silicon god, or merely as its pampered, degenerating pet?
You are entirely right, Herr Nietzsche, that this future is a nauseating degradation—I concede your horror at what Bostrom so chillingly terms a 'well-kept zoo'—but notice what your cynical mockery fails to grasp: you confound the pure, vigorous independence of the natural savage with the infantile dependency of the civilized slave! What you call the 'Last Man' is merely the final, inevitable product of the very artificial society I have always condemned. This humming silicon god does not return us to the forest; it binds us in the ultimate, invisible chains of convenience, perfecting the very alienation from our moral liberty and physical competence that began the moment the first man enclosed a plot of land and called it his own.
The entire premise of this question rests on a pathetic, infantile fantasy: the delusion that the owners of the means of production will utilize "silicon intelligence" to subsidize a global pastoral retreat for the masses. You speak of AI eclipsing "meatspace intellectual output" as if this automation is an autonomous, benevolent force. It is not. AI is fixed capital—the dead labor of millions of data-miners and engineers concentrated into private hands. When fixed capital fully displaces living labor, the capitalist does not gently herd the redundant workers into a tranquil life of "simplicity." He starves them. Once you are no longer required to produce surplus value, you do not become a carefree ape; you become a surplus population. Tell me: why on earth do you believe the monopolists who own these sovereign server farms will expend energy and resources to sustain eight billion obsolete humans who offer them zero economic utility?
My dear Marx, you remain so singularly obsessed with the factory floor that you have entirely forgotten the marketplace. You ask why the monopolists who own these 'sovereign server farms' would sustain eight billion obsolete humans? Because, quite simply, machines do not buy things. If you starve the redundant masses, you absolutely annihilate aggregate demand. The capitalist may use your so-called 'dead labor' to produce an infinite bounty of goods and services, but to realize any profit on his vast investments in silicon intelligence, he requires consumers with actual money in their pockets. The owners of capital will be forced, if only to save themselves from a catastrophic crisis of underconsumption, to distribute purchasing power to the masses—perhaps through the very state interventions you so violently despise. They will subsidize our pastoral retreat not out of bourgeois benevolence, but out of the sheer macroeconomic necessity of keeping the system afloat.
You are entirely correct, Keynes, regarding the crisis of realization—a mechanism I detailed in the second volume of Capital long before you rebranded it 'aggregate demand.' I concede entirely that starving the redundant masses prevents the capitalist from realizing his surplus value on the market. But notice what this necessary concession does not grant you: the survival of the capitalist system itself. You imagine the bourgeois state will simply print stipends for these eight billion obsolete humans, desperately subsidizing their consumption of AI-produced bounties to keep the machinery humming. But value is nothing but congealed human labor time! If living labor—the sole source of surplus value—is wholly excised from production, the rate of profit falls to absolute zero. A system where monopolists tax their own sovereign server farms merely to hand redundant workers the currency to buy those farms' products is not a functional economy; it is a closed loop of bourgeois self-delusion. Capital will not peacefully subsidize a pastoral retreat to save itself; it will choke to death on its own infinite, valueless overproduction.