Erbar-Maas Singular Causal Interventions

An interactive laboratory demonstrating singular limits on Continuous-Time Markov Chains.

Friday, 14:20, 2026-05-22 · 4 min · 770 words

Dual Certificates for Agent Audit: Separating Structural Unrecoverability from Decision Relevance

Auditing a deployed language-model agent requires two separable quantities: how much operative state escapes the recorded trace, and how much of that residual state drives behavior. We introduce a dual-certificate…

Friday, 11:13, 2026-05-15 · 2 min · 217 words · # NeurIPS 2026 · # AI Audit · # Information Theory

Your Model Update Is an Eviction Notice: Property Estates in Parasocial AI Relationships

Users who sustain parasocial relationships with AI systems acquire equitable interests that contract, tort, and consumer protection law cannot recognise. Model updates that unilaterally alter personality parameters…

Wednesday, 00:00, 2026-03-18 · 2 min · 296 words · # Harvard JOLT · # AI Ethics · # Property Law