Research
Papers on how long evidence stays true, and what happens when systems act faster than they can check it.
The papers name the failure classes; the software carries the response. The Δt series is the scaffolding for the check-versus-action logic in standing and the admission checks in Constellation. Some of this work is theoretical only: not everything here has an implementation, and the papers are not evidence that anything is deployed.
Selected claims from the Δt series are machine-checked in Lean 4: unpingable/lean (proof reader’s portal). The theorems prove the boundaries of refusal classes, not that any deployed system is safe. A small separate family, design constraints, holds general results used to bound design claims — what a quantized score, a perturbation distance, a spend limit or a provenance chain can and cannot license. These are formal results used to constrain design claims; they are not a formal verification of Constellation or ABSD as whole systems.
Engineering and reliability work first, then the broader work. Every paper is listed. Sources and drafts: github.com/unpingable/papers · ORCID
Timing, verification and reliability
How long a check stays true, what happens when systems act faster than they can verify, and where that turns into faults.
- Δt-Constrained Inference: A General Model of Temporal Coherence in Hierarchical Systems
- Cybernetic Fault Domains: When Commitment Outruns Verification
- Detecting Temporal Debt in Language Models and Software Systems
- The Temporal Attack Surface: A Δt Framework for Asynchronous Security Systems
- Control Laws for Hierarchical Kinetics: Design Principles for Multi-Timescale Systems
- The Gain Geometry of Temporal Mismatch: Shear, Leverage, and Capture in Multi-Timescale Systems
AI systems and governed reasoning
The same timing and feedback questions applied to language models, reward, alignment and reasoning under explicit authority.
- Scalar Reward Collapse: A General Theory of Eigenstructure Evaporation in Closed-Loop Systems
- Temporal Closure Requirements for Synthetic Coherence
- You Need More Than Just Attention: Invariant Requirements for Temporal Coherence in AI Systems
- Representational Invariance and the Observer Problem in Language Model Alignment
- Bounded Lattice Inference: A Governed Reasoning Substrate with Persistent State and Non-Linguistic Authority
Institutions, media and society
Broader work using the same models on organisations, censorship, social media and propaganda.
- The Coherence Criterion: A Unified Framework for Stability in Hierarchical Systems
- The Second Law of Organizations: How Temporal Lag Drives Irreversible Institutional Decay
- Capacity-Constrained Stability: A Control-Theoretic Framework for Institutional Resilience
- Temporal Asymmetry in Censorship Systems
- Eigenstructure Collapse in Social Media Platforms
- Receipt the Compiler: Propaganda as Hidden Epistemic Policy and the Architecture of Legible Memory