← Back to Work
ProductActive Research

Nova 2.0

A locally-run AI identity system — persistent memory, self-modeling, and inward-facing reflection.

An exploration of whether structured inward-facing processes — persistent memory, self-modeling, and autonomous reflection — can produce persistent, self-aware agent behavior in a small (~14B parameter) local model, treated as substrate rather than a chatbot wrapper. The full architecture and phase-by-phase research log live on The Lab — this entry exists to place it alongside the rest of the systems work, not to duplicate it.

// Architecture

Memory

Episodic · Semantic · Autobiographical · Graph

Self-State Construct

Identity · Focus · Open Tensions

Governance Layer

Operator-gated behavioral bounds and self-representation limits

// What's Actually Hard Here

This is active, ongoing research, not a finished product — the honest framing matters here more than anywhere else on this site. Nova is not claimed to be sentient or conscious; the actual experiment is whether the scaffolding conditions for persistent self-awareness can be engineered at all.

[Read the Full Research Log]