A useful health module inside fudge would need to be private by architecture, explainable in its reasoning, and built to support clinicians, patients, and operators rather than replace them.
Why this matters
Medical information is fragmented across systems that do not remember together.
Many AI products in health produce answers without adequate provenance or durable context.
Patients are often reduced to isolated events instead of long, evolving stories.
How we approach it
fudge provides the local-first shell, the shared object model, and the ability to keep sensitive workflows closer to the user.
samar provides the long-memory and evidence-sensitive reasoning layer instead of a generic chatbot veneer.
As Wiyc’s data and people branches mature, they can support domain-specific curation and training for health-facing workflows inside the same platform.
Where things stand
This is a future fudge module, not an announced standalone product.
The software and intelligence primitives are the active work today.
Branches in play
Software. Enables private, durable tools for clinicians and operators.
Data. Supports domain-specific annotation and curation under strict controls.
Intelligence. Adds memory, evidence tracking, and careful inference.
People. Keeps human expertise as the final authority.
If Wiyc approaches health well, it will be because fudge becomes a strong enough platform to host serious domain work without losing restraint.