degenie

An AI R&D lab building verifiable autonomous systems.

Autonomous software is starting to act in the world — moving capital, signing decisions, running unattended. The common safety story is testing. Ours is proof: systems whose behavior is specified, machine-checked, and attested at runtime, so a claim about what the system did is something you can verify rather than something you have to trust.

The model

The lab researches and publishes — papers, open source, skills, tools, courses. Products carry that work into production. The lab does not sell; it establishes what is true. The products are built on it.

Flagship

DEED — autonomous capital operations with a verifiable record. Every action journaled, every claim checkable. Built on the lab's research into machine-checked laws of motion for autonomous capital.

Current research

Coalgebraic self-custody: a machine-checked law of motion for autonomous capital — two safety theorems proven in Cubical Agda, with a byte-reproducible type-check and a fail-closed runtime attestation gate.