LIVEX
01 Understand

A living archive of a simulated world.

LIVEX starts from a simple question: can a collective behaviour that no one programmed by hand be made observable? What the project produces is not a story, but a record.

1.1 The problem

Understand without scripting.

Many simulations settle for playing out a planned sequence. You see in them exactly what was put in. LIVEX aims for the opposite: place agents, rules and an environment, then let it run and record what happens — including the unexpected, including what does not work.

This imposes two requirements:

  • Determinism — from the same seed, a run replays identically, tick after tick.
  • Separation — what happens (the engine) and what is said about it (the observation) are never conflated, on pain of narrating what one hoped to see.
1.2 The method

Three steps, three components.

01

Simulate

SYNE runs the world: agents, resources, rules. Every run is replayable identically and timestamped by a seed.

02

Observe

ECHOS ingests the traces and derives measurements from them: viability, patterns, loops. It observes without acting on the world.

03

Make visible

PRISM makes the world and its dynamics perceptible, through rendering and interaction.

1.3 What LIVEX is not

Limitations we accept.

LIVEX's agents are neither conscious, nor people, nor models of real societies. They aresimulated entities, whose “emotions” or “intentions” are variables, not mental states. LIVEX claims neither to predict nor to reproduce reality.

The project is in an exploration phase (V0.2). No public version is distributed to date. The component states describe the development repository.