01
Narrative memory
Events remain linked to their original sources and can be recalled.
INSIDE AN AGENT
From experience to identity.
Follow an example agent, Aria. See how an interaction can become memory, how reflection evaluates conduct and when identity may change.
Open the interactive Aria experienceAN AGENT NAMED ARIA
One interaction, four moments.
Aria is an example AI agent with GMESH implemented. The interaction below illustrates the logic; it is not a live view of the prototype.
01 / ARIA
An interaction enters as an attributable source.
Gianni shares a new event. What he said remains linked to the memory derived from it.
02 / ARIA
The right history returns before the next response.
Aria can bring relevant memories into the present interaction without treating every past detail as equally important.
03 / ARIA
Conduct is evaluated after the interaction.
An account of Aria’s own response is examined. Remembering an event alone does not change her character.
04 / ARIA
A qualified factor can move identity within limits.
When a reflective factor qualifies, the change follows defined rules and remains linked to its evidence.
Open the full interactive walkthrough
Explore Aria’s cognitive architecture, twelve departments and a narrated example. This is an illustrative simulation.
Open experience Explore Aria’s cognitive architecture, twelve departments and a narrated example. This is an illustrative simulation.
THE ARCHITECTURE
A working relationship between memory and identity.
The interactive walkthrough uses a scripted example to explain the architecture. Aria is an agent with GMESH implemented; GMESH is the identity infrastructure.
02
Identity traits
A declared starting point and a changing current state coexist.
03
Governance and audit
The agent’s outputs and identity updates can be assessed against rules and evidence.
FINFUN / A FINITE HORIZON
Limits give decisions a weight.
FinFun introduces a finite horizon into the agent’s lifecycle. Its available time, energy and information form an internal constraint that can influence priorities and the weight of experience.
It is a design principle inspired by philosophical questions about finitude; it makes no claim that an agent feels human anxiety or is conscious.