Care
We build hospitals and clinics, and we operate multiple of them. Each is a Clinical Deployment Environment [CDE].
Our clinicsAbout
The frontier research lab for AI-native healthcare systems.
Why we are here
There are too few doctors and nurses for the care people need. Our mission is to make care more accessible and efficient through AI-native hospitals and clinics: facilities designed with AI as part of how care is delivered.
We bring clinical work and research together. Our experiments are designed to test how models reason through care over time, where an answer cannot always be checked at the moment it is given.
What we do
Research
We build clinical benchmarks and publish the Clinical Benchmarks Index, where model results link to their sources.
Our researchSystems
Our method, in development, is designed to capture a visit as an episode: what was expected, what was found and what happened next.
Data and methodsHow we work
In the clinic
Care runs as a relay, from administration and triage through examination, laboratory, pharmacy and follow-up. Our research method is designed to record what was known at each decision, with test results kept sealed until expectations are recorded.
In the lab
We build benchmarks and examine the evidence behind model results. Clinical Reasoning Bench is in development: clinicians and models work the same cases, step by step. Follow-up evaluations are designed to compare a model's prediction with the later outcome.
Our office
250 Montgomery Street, Suite 450San Francisco, CA 94104