AI

AI video models; dreamworlds of platform capitalism Visa mindreThe world made visible through AI video synthesis models such as Sora, Runway, or Luma Dream Machine is a strange place. It is a world in which cause and effect seem largely decoupled, actions seem to have little or no material consequences, and bodies glide through sand and snow without ever leaving a trace. It is a world where no substantial action, let alone physical labor or material production, seems possible. Although these models allow us to extend and expand each video clip almost endlessly, stretching time and adding more and more visual content, the result is never a narrative sequence of events, just a continuous flow of images. And while what has taken the position of the (virtual) camera appears to be in constant motion, the spaces it traverses do not exist before becoming an image. They have no substance, no gravity, and no three-dimensionality, not even a virtual or simulated one. Any spatial continuum you see unfolds only in time, and any physical coherence of these worlds is only an effect of temporal continuity.What these AI-generated clips show, then, are dream worlds of platform capitalism: worlds made up of endless flows of visual patterns feeding back on themselves, worlds that seem infinitely scalable and expandable without physical limits or consequences, worlds in which objects, bodies, and spaces appear out of nowhere and dissolve without a trace, worlds in which time can be stretched, looped and manipulated at will without anything ever happening: flat worlds of an eternal present. But of course, actions in the real world do have material consequences, starting with the enormous exploitation of natural resources and human labor involved in the development, training, and alignment of these AI models. Far from being able to simulate the physical world, as some of its proponents claim, AI video synthesis is designed to make it invisible.