In Development
Work that isn’t finished, shown anyway — because the way to finish it is to put it in front of people who can see what I can’t.
Everything here is a working draft. It is kept deliberately apart from the reviewed material in the Commons, and is not offered as a citable result. Licensed CC BY 4.0; provided as-is, and not legal advice.
On the Undecidability of Instruction Boundaries in Unified Representation Systems — Or: Why Prompt Injection Is Not a Bug
A draft argument that prompt injection is not a software bug awaiting a patch, but a structural consequence of unified-representation expressiveness — by reduction to Rice’s theorem. The Rice step is the part that stands on its own; the wider arc across Gödel (1931), von Neumann (1945), and contemporary LLMs is where a sharper formulation is wanted.
Draft, not result. Co-authors and reviewers welcome.
The Gödel–Chaitin Modeling Boundary — Or, Why There’s No Digital Twin for Causality
A draft argument that one recurring shape — Kolmogorov incompressibility, Lyapunov instability, and the limits of self-description — turns up across digital twins, language-theoretic security, and prompt injection. The unifying claim is the most speculative part, and the digital-twin domain in particular needs a firmer footing than this draft currently gives it. Said plainly so a reader knows where to push.
Draft, not result. Co-authors and reviewers welcome.
Where a collaborator would move the needle
Specific, bounded places where the right pair of eyes would help more than general encouragement:
- A rigorous definition of the complexity threshold C, which the Gödel–Chaitin draft currently leans on informally.
- Whether the digital-twin limit is better grounded in computational irreducibility than in Gödelian incompleteness — a swap I suspect strengthens the argument rather than weakening it.
- Prior art on unifying the self-reference results (Gödel, Tarski, Rice, the halting problem). I’m currently exploring Lawvere’s fixed-point theorem as the common spine; pointers, corrections, and sharper expositions welcome.
A sentence or two on your background and which question pulls at you is plenty. Get in touch.