Institute for Differential Medicine

Everyone is racing to test the patient. Almost nobody is training the doctor.

You can buy 160 blood markers for the price of a phone. The report lands on a desk, and no general practitioner was taught to read all of it as one connected picture. The tests sprinted ahead of the people meant to use them.

That gap is the opportunity. Train the doctors who can act on this data and you sit between the testing boom and the patient. Both sides pay you, over years, not once.

• Demand showed up. People now pay out of pocket for their own health data. Function Health and Superpower built real businesses on consumer biomarker panels. The buyers exist and they are not waiting for permission.

• The tool became buildable. Reading genomic, protein, and metabolic data as one picture is now tractable with modern AI. Five years ago no single doctor could hold that much in their head. Now software can. (At the Institute, that software is still in development.)

• The supply cracked. Doctors are tired of react-and-prescribe medicine and drifting toward systems and functional practice. The old credentials are decades old. Room for a rigorous, AI-native one is wide open.

Two products, one loop. They arrive at different times. Be honest about which is which.

The course is live. Twenty-four modules, about 250 hours, from differential endocrinology to whole-body integration. It teaches a doctor to connect biochemistry to treatment instead of fixing organs one at a time. This is the credential, and it is the part that exists.

The software is on the roadmap. It walks a patient through four steps: deep phenotyping, omics analysis, pattern synthesis, then a bio-individualized fix. AI as the amplifier, never the decider. It is not shipped. Treat the recurring revenue as a plan, not a number.

Tuition gets paid once. The software, when it lands, bills every month.

Start with the course. Practitioner training in functional and precision medicine is, generously, a low-single-digit-billion market worldwide. Small. But it is the way in, and it builds the one asset that matters: a base of certified doctors who all practice the same way.

Then the model grows. Every certified doctor is a future software seat. Bill per seat, per month, and what one doctor pays over a few years dwarfs a one-time tuition. As the base grows, software revenue rides along, and the prize stretches from the course into the much larger layer beneath it.

Two more doors open. The certification becomes what clinics and employers hire against. And treatment data from thousands of patients makes the software harder to catch. (All figures are order-of-magnitude estimates, not company numbers.)

Education is slow to build and slower to dislodge. Once the Institute's certification is the one serious systems-medicine doctors hold, it becomes the default. Defaults are murder to unseat. The incumbent credential held its spot for thirty years.

The software, once it ships, adds a second lock. Every patient run through the four steps teaches it. More doctors, more patients, sharper patterns, software that is more useful to the next doctor. The course and the data feed each other. A diagnostics company cannot copy the trained doctors. An education company cannot copy the data. You need both. That is the point.

Pauline Jumeau is a PharmD with a doctorate in physiology, working in exactly the right fields: biomarkers, micronutrition, nutrient-gene interaction, metabolic and mitochondrial health. She is also a real voice in health and longevity, with an audience that exists before the product does. For an education company, that solves the hardest problem there is: finding the first students.

My job is to build the company around her. I run the first raise as fundraiser-in-chief. I build the hiring machine for the founding clinicians the curriculum needs and the engineering lead the software needs. I open the network and bring the operating muscle. The medicine is hers. The company-building is mine.

• Trust takes time. A new credential has to earn authority the incumbents spent decades building. Without recognized accreditation, a certification is a course with a nice name. This is the slow, unglamorous part.

• The software does not exist yet. The AI layer is still in development and the original timeline may have slipped. Until it ships and works on real patients, the recurring-revenue half of the story is a promise.

• Regulation. Treatment guided by genomic and metabolic data sits close to regulated medical practice. The line between teaching, helping a doctor decide, and diagnosing has to be drawn carefully, in the US, by people who know the rules cold.

Medicine is shifting from treating chronic disease after it shows up toward reading the whole body before it breaks. That shift needs a new kind of doctor, and almost nobody is training them.

My bet: whoever trains those doctors, and later sells them the software they treat with, holds the strongest seat in the next era of medicine. Course first. Software underneath, once it is real. A founder who can reach the first thousand doctors before anyone else gets close.