Executive summary of the TM4 framework. What it is, what it isn't, and why it matters.
Detailed overview of the four-layer architecture and governance principles.
Comprehensive technical specification including implementation details, validation results, and future roadmap.
Visual representation of the four governance layers and their interactions.
Step-by-step visualization of the autonomy loop: Ideation → Proposal → Tournament → Adoption.
Illustration of how TM4 prevents reward hacking, prompt leakage, and other forms of gaming.
Academic publications and preprints related to TM4 and governed autonomous evolution:
Authors: TBD | Status: In Preparation
Comprehensive introduction to the TM4 framework, including architecture, anti-cheat mechanisms, and validation results.
Authors: TBD | Status: Planned
Argues for validating governance frameworks on simple artifacts before attempting to govern complex agents.
Authors: TBD | Status: Planned
Techniques for designing test cases that detect and prevent reward hacking in evolutionary AI systems.
All TM4 documentation, diagrams, and research papers are released under open access licenses.
Why? Because governance frameworks must be transparent, auditable, and reproducible.
→ Source code: GitHub
→ Documentation: Available on this site
→ Papers: Preprints on arXiv (when available)
A governance framework that cannot be audited cannot be trusted.