Now /
Now.
Most of my attention is on Moissanite Labs. Alongside it: the applied work at ASKA Digital, open research at Foundation-0, and the software verification programme at Proven Commons.
Different kinds of work. The same wish to build things people can use, question, and carry further.
01 / The main focus
Moissanite Labs ↗
I am putting most of my effort into the conditions for people and AI to work well together. The ambition is a place where a person can keep continuity, find collaborators, and bring an idea into the world without handing over the choices that matter.
Right now that means building the foundations and working out the boundaries between them: what should remember, what should travel, what needs consent, and how another person can inspect the work. I want useful modules with clear interfaces, rather than a world someone has to buy into all at once.
The research is public now. The Weft papers (0 to 2) are at w3ft.org, the razi-cells program (colonies of small agent harnesses under a measured selection rule) has an open record with its registered guesses and its misses, and the lab site carries the report and the map of open directions. Guesses are written before the runs, and the misses are kept.
The wider ecosystem is still taking shape. Some parts have local implementations; others are proposals. My attention is on making that shared direction concrete without confusing the destination with what exists today. The wider direction · How the pieces fit
02 / Applied work
ASKA Digital ↗
This is the applied side of my work: software and AI workflows built around how a business actually operates. My attention here is on turning technical possibilities into usable systems, with the scope and the result clear enough to check. It keeps the larger questions close to the practical job in front of us.
03 / Open research
Foundation-0 ↗
I am working towards open research in mathematics, physics, and astronomy that another person or agent can check, extend, and cite. The effort is as much about the record as the result: interactive explanations, papers, and claim checkers, with proofs, finite checks, known work, and conjectures kept distinct. The geometry research remains one thread through that wider effort.
04 / A programme of Foundation-0
Proven Commons ↗
The direction here is software people can inspect down to its specification and proof: rebuilding widely used libraries in Lean, publishing machine-checkable proofs, and comparing behaviour against the originals. I am also working on a path for people to learn by contributing modules under review. Each project has to earn its own verification claim; the programme's ambition does not stand in for that evidence.
A snapshot, dated above. Work goes into the things being built; this page holds where my attention is.