To save as PDF: File → Print → Save as PDF (Cmd+P)
| Live site (Vercel) | https://cerebro-deploy.vercel.app/ |
|---|---|
| Codebase | cerebro/ |
CEREBRO models four societal pendulums (clocks) and a Composite Pressure Index to quantify system state. Users ask plain-English predictive queries via the Oracle and read long-form dispatches in the Pendulum Gazette.
| Concept | Description |
|---|---|
| Four Clocks | Harm Tolerance, Class Permeability, Sexual Pendulum, Good vs. Evil |
| Composite Pressure Index | Weighted aggregate (−10 to +10); e.g. +0.8 = "SYSTEM SWINGING" |
| Saddle | Inflection point: low velocity, acceleration opposes direction; high policy-change risk |
| Oracle | Plain-English predictive query engine (GSS + Pew Ring B loaded) |
| Pendulum Gazette | Long-form dispatches for each clock |
Maximum Harm ←→ Maximum Protection
Constraint ←→ Autonomy
Extraction ←→ Redistribution
Cruelty ←→ Accountability
| Ring | Purpose | Data Type | Confidence |
|---|---|---|---|
| Ring A | Structural (behavioral, institutional) | FRED, BJS, FBI, CDC, World Bank | 90%+ |
| Ring B | Normative (attitudes, beliefs) | GSS, Pew, ISSP | 94% (GSS) |
| Ring C | External pressure (geopolitical, environmental) | UCDP, ACLED, Freedom House, UNHCR | 88–92% |
| L1 | Leading indicators (search, attention) | Google Trends | 85% |
| L2 | Narrative intelligence (text, migration) | GBCD Corpus | 82% |
detect_saddle_canonical: |v| < 0.15 AND sign(a) opposes sign(v)
v = velocity (rate of change). a = acceleration. Opposes = (v>0 and a<0) or (v<0 and a>0). Calibration expands candidates for statistical power; production uses core saddles only.
Harm Clock: clock_position_10pt, velocity, acceleration, saddle_score (0–3), ring_B_pct, analogues, release_risk, energy_score.
Composite Pressure Index: Weighted average of four clock positions. Phases: SYSTEM SWINGING, LOADING, RECOVERY.
Country Risk: (Gini × 0.3) + (income polarization × 0.25) + (redistribution demand × 0.2) + (migration narrative × 0.15) + (UCDP trend × 0.1)
Scripts: run_all.sh, cerebro_export_ui_data.py, cerebro_oracle_router.py, cerebro_pipeline.py, cerebro_calibration.py, cerebro_risk_engine.py, cerebro_core/__init__.py
Data: cerebro_harm_clock_data.csv, public/cerebro_data.json
Docs: DATA_SOURCES_REGISTRY.md, DATA_COLLECTION_STEPS.md, DEPLOY.md
Vercel: Output directory public/. Build: python cerebro_export_ui_data.py.
Refresh: ./run_all.sh then python cerebro_export_ui_data.py, then push.
| Intent | Example |
|---|---|
| METHOD | "What is the exact mathematical rule for saddle detection?" |
| era | "What historical era does 2022 most resemble?" |
| FORECAST | "When does the wealth redistribution crisis peak?" |
| country_unequal | "Top 10 most unequal countries" |
| country_risk | "Country risk of class-driven political violence before 2030" |
| apogee_multi | "Show all four apogees" |
| loaded | "How loaded is the system right now?" |
| trends | "Cultural velocity for 'defund the police' vs 'back the blue'" |
| punitive | "Are we heading into a punitive crime era?" |