I just published a research paper on Zenodo laying out the origin story and full rationale behind Project-AI — a multi-layered constitutional governance framework for AI systems.
This isn't just another alignment paper. It argues that governance needs to be a structural property of AI architecture — not an external constraint bolted on after the fact.
Core components covered:
- AGI Charter (identity + continuity as protected surfaces)
- Thirsty's Symbolic Compression Grammar (TSCG / TSCG-B)
- STATE_REGISTER (operational continuity)
- OctoReflex (syscall-level containment via control theory)
DOI: https://doi.org/10.5281/zenodo.19592336
Full paper (open access): https://zenodo.org/records/19592336
Feedback welcome. This is solo independent research — built from lived experience and technical investigation into what real enforceable AI governance looks like.
[link] [comments]