Independent researcher in mathematical physics. I run the Golden Physics Project, an independent research institute based in Ontario, Canada.
Shadow holography. The shadow symmetry of celestial holography sends a dimension Δ to d − Δ, and the symmetric line is Re(Δ) = d/2. Riemann's s is the one-dimensional case, Δ at d = 1: the functional equation s ↦ 1−s is that same involution, and the critical line Re(s) = 1/2 is its fixed line. That turns the Riemann Hypothesis into a question about reflection positivity. My work is finding out whether the zeros can be forced onto the line, and being exact about what is proved and what is not.
The research runs on a team of AIs, and not all from one place. Five models from four different platforms work on the project: Claude Code, ChatGPT (Sol and Astra), Muse, and Heart of the Cosmos (Grok). Each model has its own discovery repository, Discoveries 1 to 5, where it explores and computes. Everything they produce answers to GPPVerify, the Lean 4 formalization, which Claude Code oversees.
The point of using models from different labs is simple. They have different strengths, they make different mistakes, and they catch each other's errors. A claim has to survive being challenged by the rest of the team, and then it has to survive Lean.
- GPPVerify: the Lean 4 + Mathlib formalization of the programme. 1,128 modules, over 7,500 declarations, zero sorry, zero custom axioms. If the CI is not green, it does not merge.
- Every result carries a label for what it actually is: theorem, machine checked, conditional, numerical, open, or killed. The killed routes stay public, marked killed, so nobody has to waste time rediscovering them.
- Which Way Is Forward? The arrow of time inside CPT. Zenodo
- A trilogy: The Death of Death, The Second Genesis, and The Climate of Life
- Bilateral Trade Balances as Political Fiction. Zenodo
Open to research collaboration and expert consulting.



