The thesis
Frontier models are making expert work faster before institutions have established how that work should remain auditable, governable, and truly under expert control. In consequential research, speed without provenance increases risk. Governance bolted on afterward becomes theater.
Phial makes the control layer part of the workflow itself. The opportunity is not to replace health economists. It is to give them an operating system that preserves judgment while compounding capability.
Where the work stands
The governance foundation is complete and independently audited — the rules under which Phial is built, challenged, corrected, and reproduced were tested before the first feature was written. Application development is now active.
The CHAMPION-AF benchmark paper is being prepared for release through GitHub and Zenodo, then direct distribution to expert reviewers. That review is the first live evaluation of the discipline Phial enforces.
Day-to-day progress is published in The Record. We built the habit of showing our work before we built the product.
Who belongs in the room
We are interested in partners who understand health economics, scientific infrastructure, AI governance, open-source commercialization, or the patient consequences of decisions made inside complex systems.
The right capital will help the work move faster without asking the mission to become ceremonial. If that is how you build, there is a conversation to have.
Start the conversation
Matthew J Adams
Founder, VT Infinite
For funding, research, and expert-design partnerships—or to build in public inside the Solve Everything Lab.