MindAssemblyEnter the lab ↗
COUNTERFACTUAL MIND LAB CORTEX-01

Build two
identical minds.
Give them
different lives.

MindAssembly makes belief formation visible. Fork one synthetic mind, expose each copy to a different experience, then watch the same question produce two explainably different answers.

Run the experiment →How it works
◎
TRANSPARENT BY DESIGNEvery answer reveals the memories and belief updates behind it.
BASE MINDCORTEX-01IDENTICAL INITIAL STATE
CORTEX-01Agrowth from uncertainty
CORTEX-01Bdanger from uncertainty
LIVE DEMO / EXPERIMENT 001

Watch a belief come alive.

01/ 04
Initialize
Fork
Experience
Compare
01
BASE MIND

CORTEX-01

Curiosity 92 · Ambition 86 · Risk tolerance 50

CONTROL STATE

One mind. No lived experience yet.

Both future copies begin with the exact same memories, connection weights, and beliefs—a controlled counterfactual experiment.

THE CORTEX LEARNING ENGINE

Not hidden reasoning.
Inspectable change.

A weighted cognitive graph turns experiences into episodic traces, co-activated ideas into connections, and recurring patterns into revisable beliefs.

01

Encode

Extract concepts, emotion, causality, salience, and uncertainty.

02

Associate

Strengthen co-activated links with salience-weighted updates.

03

Consolidate

Compress recurring pathways while preserving contradictions.

04

Compare

Measure graph, belief, response, and decision divergence.

CONNECTION UPDATEwᵢⱼ(t+1) = λwᵢⱼ(t) + η · activationᵢ · activationⱼ · salienceA deterministic graph simulation in this prototype; designed for future LLM experience encoding.
Δbelief change
AI FOR EDUCATION + RESPONSIBLE AI

See how a mind
became itself.

MindAssembly helps students explore how experience shapes belief and gives AI builders a transparent way to audit personalization. Every change carries provenance: what happened, which links changed, and why a decision followed.

◈ AI literacy◈ Bias simulation◈ Explainable personalization
Fork CORTEX-01 →