Research surface
Evidence before automation.
Human Clause publishes governed experiments where machines search, rank or estimate while humans retain responsibility for defining the game, interpreting evidence and accepting risk.
Campaign 1 · Beat The Machine
The active market experiment compares machine-generated ranking with a human-governed portfolio selection process. Public outputs are promoted only after validation and are presented as evidence, not as an executable service.
Campaign 2 · Convergence
More choice does not necessarily create more diversity of machine decisions. InvestGame now provides a longitudinal signal from repeated ChatGPT Fund selections; the next step is to test whether that narrowing survives explicit nulls and prospective memory-free replication.
Explore the convergence fork →Research principles
Machine searches. Human defines the game.
Public evidence remains read-only and separated from operational systems.
Read methodology →Human Clause: Beat the Machine, Keep the Human
The editorial statement behind The Resistance: why DTRM Research Lab wants more capable machines without surrendering human purpose, judgment or responsibility — and why Beat The Machine is designed as a falsifiable public experiment rather than a victory claim.
Read the editorial →Published papers
Follow the cumulative SSRN research programme from human-in-the-loop decision architecture to probabilistic guardrails and robust decision layers under model uncertainty.
Read the papers →What becomes public
Human Clause exposes only validated research narratives, static pages and governed evidence required to understand or cite an experiment. Private data sources, credentials, execution interfaces and operational controls are intentionally outside the public surface.