A Second Foundation

Research team

The Team,
Restructured

The original nine disciplines × 15 paper-reading sub-agents are being replaced by a lean two-class team: generators who propose, and compute agents who run code. Evidence is clear that beyond a small panel, more agents degrade results.

The research team, rebuilt

From 150 readers to seven roles that each own an artifact

The 2024–2026 evidence is clear: homogeneous multi-agent debate underperforms, and beyond a small panel more agents degrade results. The ceremonial roster of 10 leads × 15 prose-only sub-agents was retired to domain knowledge bases (workers cite them; nobody 'runs' them). In its place: seven heterogeneous roles, each owning a concrete artifact, with strict proposer ≠ scorer ≠ critic separation and parallel cold-briefed workers instead of sequential deliberation.

Judgment (large model)

The Orchestrator routes each session on a lean context; the Philosopher of Science runs the adversarial admission gate. Both reason; neither writes the model code.

Execution (worker model)

Cold-briefed workers whose output is an artifact: the Data Engineer owns the panel and feature pipelines, the ML/Forecasting Engineer owns variant construction, the Demographer owns age-structure features, the Calibration Auditor owns reliability and CI reporting, the Red-Team Forecaster owns the adversarial counter-case before any prediction is registered.

Lead agentVerdictWhat changes
OrchestratorROUTESLean-context session router; runs the integrity pre-flight and chooses the one move.
Data EngineerOWNS DATAThe country-year panel, feature pipelines, and source caches — the role that broke the data constraint.
ML / Forecasting EngineerOWNS THE ZOOBuilds and maintains the runnable variants and grooms the hypothesis backlog.
DemographerNEWOwns age-structure / youth-bulge / urbanization features (UN WPP) — feeds the structural-demographic variant.
Calibration AuditorNEWReads reliability curves and confidence intervals after every session; flags train-vs-hold-out drift.
Red-Team ForecasterNEWWrites the adversarial counter-case before any prediction is registered.
Philosopher of ScienceJUDGES ONLYOwns the falsification register and the admission gate; pre-registers severe tests before scores are read — and never selects the work.
Legacy 10 leads→ KNOWLEDGE BASESCliodynamicist, Econophysicist, Stat Physicist, et al. survive as domain references the workers cite, not as dispatch targets.

Below: the nine disciplines as they stand today, with their session histories. See the full redesign →

Layer 1 · Micro-Foundation

What drives individual human decisions?

Layer 2 · Evolutionary Constants

Which behavioral parameters are universal?

Layer 3 · Macro-Dynamics

What are the large-scale historical laws?

Layer 4 · Formalization

The mathematical backbone

Layer 5 · Cross-Cutting

Epistemological oversight across all layers