Two scripts unfold simultaneously.
One emerges from memory — a structure formed through repetition, experience, and learned perception. The other extends forward — a projection shaped by anticipation, attention, and forms yet to emerge. Together, they mirror the logic of predictive processing: a continuous interplay between past impressions and future expectations, shaping how we perceive the present.
Priors is a 16 minute, 4K vertical moving image work grounded in the neuroscience and philosophy of perception. It explores how our internal models (built over time and constantly updated) guide the way we engage with our environments. Perception, in this context, is not a passive reception, but an active inference; an act. We encounter the world not as it is, but as our accumulated models anticipate it to be.
The work is built with a custom generative system that layers various machine vision algorithms and personal visual data. The final piece consists of 16 chapters rooted in the same foundational algorithm. After a shared starting point, each diverges into a different classification environment, resulting in uniquely encoded outcomes. These are the residue of many perceptual cycles—a recursive field shaped by repeated attempts to stabilize, predict, and make sense of the world, each defined by the habitat in which the latent was “bred” within.
Priors brings these systems of sensing into conversation. It does not separate human from machine, but reflects on a shared understanding: the way vision is constructed through what is already known, and what is predicted (imagined). It is an unstable perceptual field, formed as much by looking as by remembering.
The work can be exhibited as a continuous, multilayered looping narrative or as 16 individual sequences, each standing on its own. The curatorial impulse behind it lies in rethinking perception as an active, generative act. It invites reflection on how we — and our technologies — don’t just observe the world, but co-create it through models that are always in motion.
Generative video, 4K, 16:00min (duration is variable based on screen dimension)
Edition of 1 or Collection of 16
Back to Top