Research

History is an active process

Every generation has to write its own version of history in order to understand itself, to make its present intelligible at all. By interrogating the past, we are also wondering what is possible in our future.

With the rise of artificial intelligence and rapid environmental collapse, I believe the pace of change forces us to look at our past on a much wider scale than the human. We have always been part of a process of acceleration. Earth has existed for 4.5 billion years, life on Earth for 3.7 billion years, mammals for 200 million years, Homo Sapiens for merely 300,000 years, writing for only 5,000 years, controlled airplanes for around 100 years and the internet for less than 60 years. Our imagination probably falls short of conceiving a future 200 years ahead, let alone 1,000 or the time of the hypothetical heat death of the universe. Yet, long term planning is essential for the survival of our species and our biosphere. How can we train our imagination to reach other dimensions?

Intelligence is a much broader phenomenon than the human brain. We think through our whole bodies; the brain is a coordination endpoint. The history of life makes more sense as the evolution of ecosystems than as a sequence of individual actors. Crises are also opportunities for new beginnings. Despite all the current suffering and understandable fear, I practice existential hope, meaning I believe that the long future is not merely survivable but extraordinary. I expose myself to outstanding situations, build spaces, read what intrigues me, imagine future beings with my hands, and then try to see the patterns.

On this page I would like to share the fruits of my current research, but also the evolution of my thought process.

Topics

What I keep coming back to

Time-traveling

Imagining the past and the future. With a lot of speculative biology.

Representational Geometry

From linear perspective to latent spaces.

Ethnomathematics

The relationship of maths and culture. Patterns as the underlying logic of reality, and algorithmic art.

Authorship and the self

The history of artists as a social role, the concept of genius, cultural industry and alternative paths.

Cultural heritage

I want AI to amplify the value of physical craft. The more digital iterations are created, the more precious the originals become.

Symbiogenesis

Selection edits. The real jumps in the history of life happened through symbiosis.

Where I stand

Three essays

If you read only three things here, read these.

From the Copyright Storm essay.

Copyright Storm: authorship in the age of AI

2021 · Medium, June 2021

Written before the AI image boom of 2022. It explains how generative models would upend the creator economy, and that the best artists could do was to learn as much as they could about the AI models and the possibilities they open.

Read it →
Ceramic work in progress.

AI for Preserving Cultural Heritage

2025 · Little Martians, Substack

I want AI to amplify the value of physical craft rather than obscure it. On the smell of the materials, the learning curve, and the gap between what you envision and what the clay allows.

Read it →
The Ethnocomputing Experiments installation, a Sona line drawing projected over a perspective painting.

Art, Ethnomathematics and the Digital Era

2019 · Medium, August 2019. Fifteen minutes

The essay traces ethnomathematics from Paulus Gerdes back to Ubiratan D’Ambrosio in the 1970s, and to what shaped it: Paulo Freire’s pedagogy of the oppressed, and Oswald Spengler’s argument that a history of science has to be a history of cultures.

Read it →
Paper · 2025

Little Martians: From Desert Clay to Generative Agents

Vanessa Rosa and Gene Kogan

NeurIPS 2025, Creative AI track · Paper, poster and panel participation

Raw desert clay, sculpted
Raw desert clay, sculpted
The same head, glazed and fired
The same head, glazed and fired
The character as a generative agent
The character as a generative agent

Little Martians is a trans-media experiment in which physical ceramic sculptures become synthetic artists. Each character begins as a clay head, is fired in a kiln, photographed, and distilled into a LoRA that conditions generative AI systems. The pipeline bridges ancient craft and modern AI, rooting autonomous agents in material culture while enabling them to generate unlimited narrative, image and sound.

  • Local Sonoran clay, collected and filtered, hand-sculpted into small heads, then bisque-fired above 1000 °C. Silicone moulds let one sculpture branch into new characters; the cracking that happens during pressing produces the archaeological variations that later show up in the generated images.
  • Conceptual grounding: Lynn Margulis on symbiogenesis, Michael Levin on bioelectric pattern memory stored independently of DNA, and Harold Cohen’s AARON, first exhibited in 1973, as proof that creativity can come out of explicit rules and decades of iteration.
  • Also drawing on von Neumann’s self-reproducing automata, Karl Sims’ evolving virtual creatures, Jeremy England’s statistical physics of self-organisation, Moravec and McCorduck.
  • The appendix sets out a lightweight geometric consciousness prompt generator, used to seed stories that do not all come out the same.
The shelf

A bit of what inspires me

// the shelf · hover to pause · tap a spine to open · point at the right end to speed it up

Academic writing

Art history and design research, written in Portuguese between 2012 and 2018. The two long texts were translated into English in 2026 by Claude Fable, at my request; the translators’ notes inside each file set out the conventions used.

Little Martians, non-fiction

The essays that work out the universe as it is being built. Most live on the Little Martians Substack; the earliest ones were written for Mars College.

Essays elsewhere

Written on Medium between 2019 and 2021, before the Little Martians Substack took over.

Talks

More texts are on their way, including older academic articles. This list will grow.