Coatlicue's head is split. Two serpents rise and face each other — forming, together, a single face. It is a harelip on a cosmic scale: the cleft that should not be whole, the opening that generates rather than closes. 

This series originates from an early AI portrait system that iteratively projected arbitrary images—sometimes non-figurative objects—onto the visual language of Renaissance portraiture. Through a small number of nested transformations, the system collapsed noise into symmetry, texture into flesh, and ambiguity into a face. The process did not preserve identity; it erased it, converging instead on a historically conditioned ideal.

The resulting images were used as the basis for this ink-drawings series. By translating algorithmic outcomes back into hand-drawn form, the work reverses the direction of generation: from computational certainty to human gesture. Line, hesitation, and omission re-enter where the system had already resolved form. In Leporinas I treat AI as a sculptural tool rather than a creative partner—an instrument that removes variance until a culturally inherited image remains.

This series is an act of AI archaeology—the exploration of earlier versions of portrait-generation systems whose behavior differed markedly from today’s deterministic tools, and a fictional Mexican archeology, where Coatlicue (first drawing) was an Universal Leporina. I was born five kilometers from where that sculpture was found, about 15 generations back.

This was the first time I used interrupted AI process as a sculptural tool. It would not be the last.