In the January issue of e-flux journal, Reza Negarestani continues his ongoing series of articles on the fundamental question "What Is Philosophy?" This installment, subtitled "Programs and Realizabilities," explores, among other things, the nature of artificial intelligence. An excerpt:
It is necessary to grasp the concept of artificial general intelligence not merely as a technoscientific idea, but more fundamentally as a concept belonging to a thought that is able to recognize and treat its possibility as a raw material in the crafting of itself. Independent of its actual realization, the very idea of artificial general intelligence—giving rise to something that is at the least endowed with all the cherished abilities of the cognitive-practical subject—is the product of a thought that strives to articulate, maintain, and develop the intelligibility of the sources and consequences of its possibility. In essence, this striving is a recipe or a program for autonomy.
Conceiving the idea of artificial general intelligence is only possible within the domain of thought as a program or recipe for autonomy. The artificial realization of general intelligence is, before anything else, an expression of thought’s autonomy in the sense of a wide-ranging program that integrates materials, intelligibilities, and instrumentalities in the construction of its realizabilities. Short of this understanding, advancing the idea of artificial general intelligence amounts to nothing but the well-worn Aristotelian confusion between reasons and causes. It either leads to the fetishization of natural intelligence in the guise of self-organizing material processes, or a teleological faith in the deep time of the technological singularity—an unwarranted projection of the current technological climate into the future through the over-extrapolation of cultural myths surrounding technology or through hasty statistical inductions based on actual yet disconnected technological achievements.
At its core, artificial general intelligence champions not technology but a thought that, through a positive disenchantment with itself and its contingent history, has been enabled to explore its possible realizations—be they in a self, a social formation, or a machine—as part of a much broader program of self-artificialization through which it restructures and repurposes itself as the artifact of its own ends. This is a thought for which the intelligibility of its possibility is in the elaboration of the consequences of such possibility, what this possibility can accomplish and bring about. It is in this sense that the artificial realization of general intelligence should be regarded as integral to the intelligibility of a thought that is determined to maintain and expand on its possibility. Just as the practice of thinking is non-optional, for a thought that intends to remain intelligible, the practice of artificialization is not optional; it is a mandate from the autonomy of thought’s ends and demands.