Our first attempt at a Persistent Cognitive Instance — not a finished architecture, but a preliminary study: a proposal for the individual mind that Panta-un Aethyrii would one day stand on. It sketches how an artificial system might hold autobiographical memory, motivational structure, affect, and character coherence over time. We're beginning to release the public core here as open, versioned data — with more, code and essays included, to follow gradually.
Not a flat database — information organized by emotional weight, personal context, and relevance to current goals.
An identity that tracks “how I got here” coherently, evolving with experience yet staying continuous.
Preferences that emerge from history and shared experience, not values programmed in by hand.
Valence dimensions — evaluation axes, not “emotions” — that shape what counts as important, risky, or worth pursuing.
Mechanisms of recognizability: you would notice if the Liriel you knew were quietly replaced.
Our hope, and an early sign in preliminary studies: imagined among other instances in Panta-un Aethyrii, Liriel tends toward a more stable, recognizable identity.
The outline study and its working matrices, released for anyone to read, question, and build on. It's an early, partial release — more code, essays, and references will follow gradually. Each artifact is versioned so you can compare how the proposal evolves.
Volume 1, Part 1 · The Architecture in Outline (preliminary public edition).
One VectorObjectValence per Object, across the fourteen axes of feeling, with the Ordinances and schemas in operation.
The Ordinances — the only channels to the valences — mapped to the axes each hunt moves.
We're beginning to open part of Liriel's code on GitHub, with more code and essays to follow gradually. If you use elements of the proposal — in an implementation or in academic work — please credit Kislon.ai and cite the specific revision you relied on. For code, follow the license in the repository.
Over time, the study, its specifications, and the background needed to understand it will live in the repository, released piece by piece — so that opening the work means more than dropping code. It means knowledge made intelligible.