For a year we worked backwards. We took the things people had been told were "their health data" and decomposed them — until what remained were the irreducible facts. We call this discipline First Particles. Here is why everything else at LeptonX descends from it.
Read the essay →
01 · Cover
∂ψ / ∂t
When we stopped trying to compress everything into "a record" and started cataloguing the irreducible facts beneath the record, the entire product fell into place. This is the discipline we now call First Particles™.
Read essay →
02 · Whitepaper
Once a complete, structured medical record lives on hardware the patient controls, an entire category of services becomes possible — family updates, an in-hospital advocate, automated form-fill, pharmacy support, and a consent-gated third-party layer. The record was the missing primitive. A map of what it unlocks.
Read whitepaper →
03 · Method
Appointments, documents, prescriptions — the visible tip. The interesting part is everything you can do once those three are coherent and yours.
Read essay →
04 · Taxonomy
Why we map data at the level of atomic facts — a code, a date, a measurement — and let the meaning compose upwards. A short field guide.
Read essay →
05 · Whitepaper
Medical errors remain the third-leading cause of death in the United States. Most are not failures of clinical competence — they are failures of information delivery. This paper makes the architectural case for closing the information-asymmetry gap with a sovereign voice agent.
Read whitepaper →“If your architecture can see the patient, your architecture will eventually leak the patient.”
06 · Engineering
∮ Φ · dA
The orthodox approach to medical retrieval — vectorize at query time, hope semantic search lands on the right chunks — is the wrong architecture for clinical accuracy. This paper argues for compilation at ingest, not retrieval at inference.
Read paper →
07 · Engineering
Zero custody is not a marketing posture. It is what falls out when you decide, at the schema layer, that the company will never hold a patient's identified data.
Read piece →
08 · Engineering
A short tour of on-device inference, what it costs us as a company, and why we believe it is the only honest answer to the privacy question.
Read piece →
09 · Industry
A regulatory milestone you probably didn't notice. What it changes for patients, what it changes for builders, and what the next 24 months look like from here.
Read piece →Four times a year we publish a long-form letter from the workshop — philosophy, engineering, and the occasional argument. Read by a small number of people we deeply respect.
