Cj TruHeart
← All work
Open dataset · Version 2.3

Featured build

Abundance
Codex

A narrative-curated dataset that gives AI agents an evidence-anchored alternative to scarcity-default reasoning.

Overall reasoning lift+0.41on a five-point quality scale
Curated entries285stories, frameworks, shadows, and syntheses
Grand Challenge domains21from energy and health to governance and space
Matched-pair judgments504across four test models

Measured, not assumed

Can a dataset change how an AI reasons?

Across the ACE benchmark, Codex-augmented responses moved from 4.15 to 4.56. The strongest gains appeared in cost-efficient models—evidence that better context can multiply capability without simply buying a larger model.

Reasoning quality · 5-point scale
Baseline
4.15
Augmented
4.56

95% CI for the overall lift: +0.29 to +0.53

The field

Five pillars. One civilization-scale map.

Every domain carries evidence for progress and a shadow check for what can fail, who can be excluded, and which conditions still need to be built.

  1. 78 entries

    Material Foundation

    Energy · Food · Water · Shelter · Health · Environment

  2. 39 entries

    Human Capability

    Education · Longevity · Consciousness

  3. 90 entries

    Collective Coordination

    Communication · Community · Governance · Security · Transportation · Economy

  4. 52 entries

    Production & Discovery

    Manufacturing · AI · Co-Creative Intelligence · Science

  5. 26 entries

    Transcendent Frontier

    Space · Future Vision

The work behind the result

The miss was part of the method.

Predictions were committed before results were computed. Each miss became a diagnosis, and each diagnosis changed the next intervention.

  1. v2.0+0.03Null

    Establish the baseline.

  2. v2.1+0.14Inconclusive

    Add synthesis and reasoning scaffolds.

  3. v2.2+0.18Missed by 0.04

    Diagnose where uniform intervention fails.

  4. v2.3+0.27Confirmed

    Calibrate the intervention to each domain.