PolyAgora v1.0 – A Natural-Language Multi-Axis Cognitive OS (Whitepaper, DOI) (zenodo.org)

PolyAgora v1.0 is a newly published whitepaper (DOI attached in the paper) introducing a “Natural‑Language Cognitive Collective Engine”: a multi‑axis operating system built and operated entirely through natural language prompts and interactions rather than conventional code. The system defines a structured cognitive field using three orthogonal reasoning vectors — Arc, Ann, and Saku — and runs coordinated multi‑agent reasoning cycles that include dynamic opposition, reference‑switching and layered A→B→C processing. The paper details the architecture, operating rules, emergent behaviors and concrete use cases for executing complex cognitive workflows purely via verbalized instructions. For the AI/ML community this matters because PolyAgora reframes an OS-level control plane as a natural‑language meta‑reasoner: a multi‑agent OS model with built‑in cognitive stability mechanisms that aim to preserve coherence across simultaneous, orthogonal reasoning streams. Key technical implications include new primitives for agent interaction (dynamic opposition and reference switching), a layered processing pipeline for idea expansion and refinement, and demonstrated emergent behaviors relevant to meta‑cognition and exploratory research. If robust, this approach could accelerate human–AI collaboration, rapid hypothesis exploration and automated reasoning orchestration — while raising questions about evaluation metrics, scalability, and how to rigorously verify correctness and safety in language‑mediated system control.