Hyperstitional (Desiring-)Machine Conjecture

Abstract

This presentation explores the convergence of computational abductive inference and the CCRUian concept of hyperstition, and how their meeting creates processes of reality-production grounded in desire rather than reason. Hyperstition operating through fictions that make themselves real via recursive feedback loops, making future conjectural states real in the present through their effects, just as computational abductive inference selects operationally effective hypotheses, thereby projecting future potentialities and acting upon them as truth. The presentation focused of four key elements; temporality, unbelief, operational fictionality, and consistency, each illustrating how hyperstition and abductive inference jointly enact a non-representational mode of reality-production. These systems do not describe the real but participate in its actualization, collapsing the virtual future into the present through machinic selection. Drawing on Deleuze and Guattari, the presentation argues that computational inference functions as a desiring-machine that plugs into any number of processes to create ontological potentialities. As with hyperstition, it becomes an engine concerned with a constant process of potentialities, of production, but just as it might deterritorialize hegemonic structures as a politics of potentiality, it may just as easily become a politics of foreclosure as futures are preempted by AI systems. It reframes AI not as an agent of reason, but as one of desire, entangled in cybernetic circuits that produce consistent and effective realities.

Title of Event

Abductive Inference

Organiser/Management