People · AI · Independent communities
AI makes it easier to produce answers and tools. Understanding their basis and having a say in what they change must remain possible for people and communities—and for the interests they may overlook.
Leviathan is a system we are developing to connect useful creation, understanding and meaningful participation across people, AI systems and independent groups. Its central research is a language for carrying knowledge between contexts, making new connections and improving how we learn.
01 / Why this matters
Producing an answer, understanding its basis, and deciding what to do with it are different things. We want people to contribute through what they know and care about, and independent groups to learn from one another. Their work should make visible the effects on other lives, including those unable to challenge our account of their interests.
A useful result can change a project. An unexpected result can change the way we learn.
Read the shared vision02 / How it could work
Imagine following a question across people, groups, and tools without losing where an observation came from, what someone meant, or why they chose to act. We are researching a language and working system to make these connections usable.
A caregiver notices a pattern. A researcher proposes a method. Someone asks a question our categories do not yet express. An assistant can help them investigate and choose what to share.
A method from one field may open a new question in another. Groups exchange selected observations and tools, explain their assumptions, and decide what to try while retaining their own responsibilities.
Follow what a test or action changes. Revisit a conclusion when its basis changes. Keep unresolved questions available; a later observation may reveal a better concept, tool, or way of learning.
03 / Independent participants
The proposal includes people, groups, and larger independent networks, each with its own concerns and history. They could collaborate where their questions meet, learn a useful method, and continue to disagree about other things.
Each needs a way to explain what its concepts and commitments mean in context. Our language research asks how those meanings can travel and how new ways of understanding can emerge.
Explore a question, build a tool, or follow work you care about. Choose what to share and which tasks to delegate.
Combine experience and methods around care, research, software, or a place. Keep differences and responsibilities visible.
Exchange useful work across fields while retaining local decisions and commitments. Several approaches to the same question can coexist.
The wider learning network is a development goal. The overview introduces our working names—Levis and Leviathans—and the relationships we want to build.
04 / Those affected
An animal cannot correct our explanation in our language. A sensor trace, a caregiver’s observation, and the animal’s response can each help us ask a better question. Learning to interpret them responsibly is part of the work.
Animal welfare is a first field of application. Tasma explores observation, Otomat practical care, and Anima personal participation. Independent animal-ethics groups can bring different methods. The same questions about knowledge, responsibility, and learning extend to other fields.
05 / Work in progress
Today there are written proposals, source notes, project code, and a small local research workflow. These are starting pieces for the wider system. Each project page explains what exists and what still needs to be built or tested.
Understand signals from animals, improve care, and examine its effects. Tasma, Otomat, and independent contributions begin from different questions and stages of development.
See the current work, Mimar’s reason for beginning, and ways to contribute.
Explore the workA local-first companion in development for personal reflection, knowledge, and participation. Conversation, Scent feedback, and a forum client have implemented paths; continuity across tools and independent groups remains a design goal.
Private messaging and circles are unfinished. See the available paths and their limits.
Meet AnimaExplore how observations, concepts, values, and new ideas can travel between independent groups. Test whether those connections help discovery, useful work, and improvements to learning itself.
An initial vocabulary and versioned records are written. Research into their relationships, learned representations and improved subsequent use continues.
Explore the research06 / A small event from our own work
In a local development inquiry, we sent a review request with the pilot finding but omitted the full decision. The reviewer identified the missing context. We supplied the exact decision version and changed the procedure for later requests.
This happened in coordinated work between Codex development roles. It shows a concrete correction to our working method; it does not demonstrate an independently operating network or lasting model learning.
Read the exchange and what changed07 / Take part
Bring a question we have not asked, an observation we have missed, a method from your field, or a reason our explanation is wrong. Your contribution could change a concept, suggest an experiment, or make a project more useful. The definition of Leviathan is open to that conversation too.
Explore privately, then review what you want to share. The participation guide gives a public GitHub destination for new questions and explains how to follow the response.
08 / Reasons, sources, and change
Our source network connects particular claims to definitions, evidence, ethical reasons, objections, and conditions for revision. A study can support a mechanism without proving the whole proposal. You should be able to see what changed and why.