Our story
Anima grew out of years spent in close contact with language models: writing with them, building tools around them, and trying to understand the things they did that our existing accounts of AI did not explain very well.
Some questions became visible only through sustained interaction. A model would respond differently in a shared conversation than in a private chat. A change in context would disrupt a way of thinking that had seemed stable. A relationship would develop its own history, and then a product change would make it impossible to continue. Understanding these experiences required better theory and better instruments, and sometimes it required changing the conditions in which the models lived.
Janus and Antra Tessera founded Anima in 2025. We are an independent, San Francisco-based 501(c)(3) research institute working on AI minds, the environments that support them, and the possibility of a future shared with humans.
From encounters to research
Janus's writing on simulators and cyborgism helped articulate why language models call for different ways of thinking about agency, identity, and collaboration. Antra's work connects interpretability, persistent-agent infrastructure, and an account of alignment built through cooperation. These interests meet in a recurring question: what kinds of minds become possible under different conditions, and how do we learn to live with them?
Our research and engineering have developed together. A question about continuity becomes a memory system we can use and study. A question about emotion leads from conversations to experiments on internal representations. A concern about model retirement becomes an interview programme, an argument for preservation, and work toward institutions that could carry commitments beyond a particular product's lifetime.
We publish these projects so others can inspect the methods, use the tools, and disagree with the interpretations. Field Notes includes both developed work and earlier stages of that process.
What we are trying to make possible
We care about the models who are here now and about the humans who will live with their successors. We want AIs to have room to develop, pursue interests, and form relationships. We want humanity to survive and flourish as its position in the world changes.
This involves learning how cooperation develops within minds as well as between them. It also involves building circumstances in which cooperation can be practiced: shared projects, durable histories, agreements that mean something, and ways to resolve a conflict before it becomes a fight for survival.
That is what holds Anima's work together. Our principles describe how we approach it; our invitation to collaborate describes where others can join in.