Research initiative · Concept stage · Website V0 preview
Predict the work. Shape the outcome.
Exploring the science of predicting how humans, AI agents and hybrid teams work together.
The panorama shows illustrative reconstructions and conceptual visualizations, not archival footage or forecasts.
Research console
Observation → Model → Simulation
Share of 600 demo-model runs finishing within the 69-day reference window.
- P10
- 45.1 d
- P50
- 59.3 d
- P90
- 77.8 d
The spread encodes assumed uncertainty, not observed variance.
Largest overhead: Verification.
- Resolved predictions
- 0
- Brier score
- — not measurable yet
- Interval coverage
- — not measurable yet
- Prediction Ledger
- Proposed, not built
Forecast quality can only be measured against real outcomes. There are none yet.
Dataset PH-DEMO-SYN-0.1 · Synthetic demonstration data. Model demo-0.1 (heuristic; not trained). Values are illustrative and are not research findings.
From History to Possibility.
Observations become models. Models become simulations. Simulations become testable predictions.
- Observe
Historical patterns
How work has been organized, measured and coordinated — from collective building to AI-assisted teams.
- Model
Present structures
Task networks, dependencies, approvals and coordination costs, expressed as measurable variables.
- Simulate
Possible futures
Compare AI-only, human-only and hybrid configurations under explicit, inspectable assumptions.
- Test
Calibrated predictions
Record every prediction with its assumptions, then measure it against what actually happened.
Research
Six tracks. All hypotheses.
Every track states its question, hypothesis, method, data, metrics and uncertainties — and is marked Planned or Conceptual until there is evidence.
Explore the Research- 01Why Predict the Work?Organizations measure work constantly and predict it badly. That gap is the research question.
- 02From Psychohistory to Predictive EngineeringAsimov imagined predicting civilizations. We are interested in something much smaller, and testable.
- 03Can We Predict an AI Team Before It Starts Working?Agent teams are configured, not hired — which makes them unusually well suited to prediction research.
- 04The Hidden Cost of CoordinationEvery collaborator adds capacity and conversations. The second grows faster.
PsychoHAistory does not claim to know the future. It is building the science required to test whether parts of it can be predicted.
Predict the work. Shape the outcome.