Notes from inside (and outside) the system, on the care model we could build instead.
Healthcare pushes its coordination work onto patients and families because everything must pass through one narrow opening. Why that happens, what it costs, and what becomes possible when software can absorb the work instead of streamlining it.
People live their health over a lifetime. Healthcare should be there for the long haul. The care model that follows from taking that seriously.
A personal health companion is not a better portal. It is a different surface, with a specific architecture — and the constraint has shifted from what we can build to whether patients and health systems would use it.
Demand rising, the clinical workforce retiring, and the family caregivers who absorb the overflow thinning out — all arriving at once. This is not a forecast. The people are already alive and their ages are known.
Nearly everything in healthcare passes through one narrow opening. Whatever cannot fit is refused, and refused demand lands on families.
People live their health on three timescales at once. The system is built for the shortest of them.
Every prior wave of health technology made the existing channel faster. Absorbing the work is a different thing entirely.
The same reading means different things on different trajectories. Alert on departure from the expected path, not a population cutoff.
The connective layer between capable models and useful care. Almost everything below it already exists; almost none of it does.
Diffusion is not paced by model capability. Three gates meter it, and all three open on the harness rather than the model.
Three supply-and-demand curves closing at once, on a schedule already set by demography.
What the medical record becomes when it stops being static. Not a digital twin, not a risk score.
What is open right now, who owns it, and what is overdue. The faster record healthcare does not keep.
One essay every two or three weeks. No cadence promises beyond that, and nothing else.
Subscribe