Original research on agentic failure modes and the mathematics of governance — the foundation the score and SREL are built on, published openly so the field can scrutinize it.
Timothy Poschel — Unisapience Labs. Filed as US Provisional Patent Applications 64/066,231 (SSPLX-001-PROV), 64/075,009 (SSPLX-002-PROV), and 64/077,286 (SSPLX-003-PROV). Discloses the n-simplex risk decomposition, the shadow-simplex construction, the seven-factor SSS, and a structural extension: multiplicative product-form aggregation, modal register stratification, capability normalizer C(κ), compositional periodicity with explicit alignment-edge exclusion, and a transcendental meta-condition veto layer.
A conceptual tool for analyzing pathological attractors in self-evolving AI systems. Five fundamental pathologies, ten pairwise couplings, and ten higher-order emergent dysfunctions mapped onto a 4-simplex topology, with testable hypotheses and experimental protocols.
Five base aspects — Model, Data, Harm, Emergence, Purpose — project onto the rectified pentachoron. Ten pairwise edges enumerate the operative risk dyads. The A–Ω alignment edge is explicitly excluded as structurally irreducible: alignment is the conjunction the framework addresses, not a peer-level item within it.
The framework is public because the field benefits from replication and critique. The durable value lives in calibration, not secrecy.
Explore the SREL standard