About Cyphris
I am Jake, an independent researcher and developer building public work around AI cognition, model verification, governance, energy infrastructure, and institutional reform.
My main project is ATTICUS: a framework for running AI models with structured reasoning, verification gates, source separation, audit discipline, safety boundaries, and operator-controlled cognition layers. The core idea is simple: raw frontier models are not enough. A stronger model running loose is not the same thing as a safer or more useful system.
I also work on the Power Avoidance Canon, a philosophy and systems-design framework for identifying how power accumulates, how it corrupts, and how institutions can be designed with expiry, contestability, external review, and reversibility.
This channel documents the research, explains the frameworks, tests ideas in public, and gives serious people a way to inspect the work directly.
