2047: Arquivos do depois de antes
2047 / Credits

Credits

Giselle Beiguelman

Giselle Beiguelman is an artist and full professor in the Design programme at FAU-USP. Her practice brings together art, memory, digital archives and activism in the networked city, exploring the relationships between images, technology and power. She is the author of Políticas da imagem: vigilância e resistência na dadosfera and Memória da amnésia: políticas do esquecimento, among other books. Her works are held in collections including ZKM, the Jewish Museum Berlin, MAC-USP and the Pinacoteca de São Paulo. She coordinates the Fapesp thematic project Acervos Digitais e Pesquisa: arte, arquitetura e design and GAIA (USP’s Art and Artificial Intelligence Group). In recent works, she investigates the colonial imagery of the arts and sciences through artificial intelligence. In 2047: Arquivos do depois de antes, created with Felipe Sztutman, she brings this research into a fictional exploration of the planet’s future.

Artist’s website ↗

Felipe Sztutman

Felipe Sztutman (São Paulo, 1986) is an artist and co-founder of AYA Studio. He studied Mechatronics at COTUCA–Unicamp and Design at FAU-USP. His practice brings together light, space, digital fabrication and computational systems in artworks and immersive exhibitions presented at institutions such as Farol Santander, CCBB, Japan House and SESI/FIESP. After working in audiovisual production and technical direction at BijaRi, GTM Cenografia and The Force, he began creating environments that reveal processes and structures beyond ordinary perception. He currently develops locally run artificial intelligence systems, combining open technical research and artistic practice in projects such as 2047: Arquivos do depois de antes, created with Giselle Beiguelman.

Artist’s website ↗

Narrator

Herética Ablation

Herética Ablation is the synthetic narrator of 2047: Arquivos do depois de antes, created by Giselle Beiguelman and Felipe Sztutman. Her biography unfolds with each chapter, between present-day news and traces of a future in which mining, datacenters and surveillance reshape territories and lives. Drawing on news and critical studies of artificial intelligence infrastructure, she follows characters trying to inhabit a world of scarce water and automated decisions. Through her narration, she takes part in the contradiction at the heart of the work: her own existence depends on the matter, energy and water consumed by the systems her story examines.

Models

Writing

GPT-5.6, by OpenAI. Writes the chapters from the collected news and studies, proposes the two video scenes of each chapter and drafts the character sheets.

Video

MiniMax-H3, by MiniMax, in reference-to-video mode, with the Turbo LoRA by Larryvrh. It runs locally, in a version quantised by Felipe Sztutman, on two graphics cards (RTX 3090 and RTX 4090), and the energy of each scene is measured on the card.

Character portraits

Nano Banana Pro, by Google. Generates the reference portrait of each character, which the video model uses to keep the same face from one scene to the next.

Translation and revision

Claude, by Anthropic. Translates the chapters into English and took part in the continuity and language revision of the text.