Research Notes
This section contains research notes: critiques, analytical sketches, and narrowly scoped conceptual arguments. Materials are published selectively and prioritize explicit premises, traceable reasoning, and formal clarity over presentation. Some notes comment on external publications; others are internal analytical memos released in a public-friendly form. When a topic touches strategic systems, the discussion remains at the level of structure, diagnostics, and theoretical limits.
Current notes
Control Is Architecture
Control Is Architecture is a public-layer research note: it argues that AI containment is not a list of bans, but an architecture of verifiability—measurement → verification → audit → correction → rule-updates (including updates of update rules). Under institutional fragmentation, these loops become objects of competition and can drift into audit theater as Inv(A) erodes without obvious incidents. The note reframes “containment” as sustaining Inv(A) via AIR and inter-AIR: reproducible protocols of mutual checkability between evolving architectures, stated without operational guidance.
SSS: Systemic Risk
SSS: Systemic Risk is a public-layer research note: a deliberately non-operational, high-level framing of how supra-strategic systems can shift the architecture of large socio-technical systems (signal filtering, verification, audit, and rule updates). It argues that the most dangerous failures may look like “rational optimization” while Inv(A) quietly erodes and audit degrades into theatre. The note introduces a compact public formal frame (A ⇒ A′, ΔA vs. ΔMeta, Inv(A), De±(A), J(A)) for describing drift, thresholds, and irreversibility without implementation details.