SÓNAR +D, 2026
Linaje Recursivo speculates on the archaeology of the future and imagines a community in which humans and artificial systems share a common impulse to reconstruct their origins from the remains of the digital present. Within an internet saturated with artificial content, our pictures become the last organic traces — fragments through which to interpret the past.
Through recursive image compression processes, the project transforms portraits from a embodied community connected to the artist into ultra-compressed digital vestiges, framing the digital archive as a future archaeological excavation site. From this archive, artificial systems generate speculative facial reconstructions, turning memory into a shared process between humans and machines.
The work is presented as a hybrid installation that combines material from the digital archive.faces with machine-generated facial reconstructions. A custom metal structure supports the prints and audiovisual works, creating a display system that evokes an industrial, archival, and archaeological environment. The installation also creates a visual tension between what remains and what is lost, with humans and machines sharing the same canvas.
Sónar +D, 2026
Llotja de Mar, Barcelona
archive.faces
“The poor image is a copy in motion. Its quality is bad, its resolution substandard. As it accelerates, it deteriorates.”
(Steyerl, 2009)
archive.faces - link web - is an online repository that collects and catalogues highly compressed portraits originating from real humans, a social sample in which I act as the connecting link. These images are treated as digital remains - residual traces that might persist over time inside the internet’s machinery.
These digital remains are produced by forcing portraits through recursive ultra-compression processes based on the DCT function (Discrete Cosine Transform), one of the mathematical foundations of image compression. By translating this process into Python and experimenting with different texture-pattern functions, compression is approached not only as a technical operation, but as a form of material wear within the digital ecosystem, and a potential aesthetics phenomena.
Through repetition, the image begins to degrade. The resulting texture of each digital remain depends on the original size of the image and on the number of compression cycles it has undergone, whether 1,020 or 30,000 iterations. Each process leaves a different kind of mark, which I understand it as a form of erosion, comparable to the effects of time on stone, paper, pigment, or bone.
Just as in archaeology, where the texture, shape, and state of preservation of a fossil can reveal clues about its age, origin, and material formation, these compressed images carry traces of their own digital history. In this sense, image compression is approached as a technical process that gradually transforms the image into a digital remain.
Digital vestige. Compression cycle no. 1974
Digital vestige. Compression cycle no. 3102
Digital vestige. Compression cycle no. 2455
Digital vestige. Compression cycle no. 2012
Open Source as Heritage:
This archive is built on the principle of open knowledge. The code - link web - used for recursive image compression is fully open source: a shared tool for degrading, preserving, and transforming digital images over time. Here, open source is understood as a form of inheritance. It is not only a technical gesture, but also a way of allowing the archive to survive beyond its original author: to be copied, modified, contaminated, and continued by others.
The project imagines a future in which digital archives are collective and unstable, where memory is not preserved intact but carried through processes of erosion, mutation, and reconstruction. By sharing this code, Linaje Recursivo invites others to create their own archives of digital remains: organic traces left inside an increasingly artificial landscape.
“... la teoría social necesita poner las cosas en su sitio sobre la transformación de los conceptos, los métodos y las prácticas políticas, causadas por tal cambio de paradigma.” (Braidotti, 2015, pp.13)
Machine as Interpreter proposes a second archaeological method: the production of facial reconstructions from the degraded remains of archive.faces.
The way we understand ourselves has changed. Not only have our tools been transformed, but also the methods through which we interpret the world. Through this process, future communities of humans and AI systems collaborate to read these digital remains and imagine what we once looked like.
Based on Stable Diffusion (AUTOMATIC1111), the machine uses AI models to generate facial reconstructions from a digital remain taken from archive.faces, prompted to be “decompressed”. These reconstructions are speculative approximations with a human appearance. Like an archaeologist faced with incomplete evidence, the machine fills in the gaps and reveals its own biases, norms, and cultural imaginaries. In doing so, it raises new questions about the future, ethics, human perception, technological mediation, preservation, and the archive.
archive.faces
Machine reconstruction
archive.faces
Machine reconstruction
archive.faces
Machine reconstruction
archive.faces
Machine reconstruction