03 / Redesign what comes next
Compute-Powered Specialty Labs
Practitioners, builders and operators take one real healthcare system or specialty and redesign it around compute.
The question
If abundant compute, AI and continuous data existed from the beginning, how would we design this system differently?
Most healthcare systems were designed around constraints that are quietly disappearing: information was expensive to move, measurement was episodic, and expert attention was the only way to interpret anything. A specialty lab starts from how the work is actually done today, then asks what the same specialty would look like if it were designed now.
These are small working groups, not conferences. Everyone in the room is expected to contribute something they know first-hand.
What a lab maps
Worked through in sequence. Each step constrains the next.
- 01Current systemHow the specialty actually runs today.
- 02BottlenecksWhere time, information and attention are lost.
- 03DecisionsWhat gets decided, by whom, on what evidence.
- 04Information flowsWhat is known, when, and by which party.
- 05Human workThe judgement and relationships that must stay human.
- 06AI-suitable workWhat compute could genuinely absorb.
- 07Future workflowThe system as it would be designed now.
- 08Experiments worth runningThe smallest tests of that design.
Example
Compute-Powered Gastroenterology
How gastroenterology might be redesigned around AI, continuous data and scarce clinician attention — from referral and triage through diagnosis, procedure scheduling and long-term follow-up.
What a lab produces
Paper Clubs, Field Sessions and Specialty Labs are scheduled on Luma.
Sessions are small and run regularly. Subscribe once and you will hear about each one as it is announced.