This gathering marks the first of many we anticipate holding in collaboration with the DOC community. The DOC community — a network of physicians, scientists, entrepreneurs, financiers, and artists — organized itself around a persistent belief: that sound science, developed alongside capable individuals, can propel meaningful medical advancement.
We met in the Living Room Lab tradition, under the Chatham House Rule, which is why the voices reflected in these pages carry no names.
One host described the method with an image that stayed with the room. Each of us was asked to throw our flashlights toward the center and light up what we saw others might be missing. Investors sat beside inventors, founders beside scientists. Most were strangers when the hour began, and the collisions were the point.
THE GIST OF THE EVENING
For the entire hour, people treated artificial intelligence as an engine that had arrived before its roads. The optimism was real, and so was the discipline beneath it. Every claim about what these models can do met a question about what they are for, whom they serve, and what they might quietly cost. Again and again, the room returned to the living human at the center of medicine — the one the machine is meant to serve.
THREE MAJOR THEMES
I. The Population and the Person
Medicine bases its approach on averages, while the patient in the chair stands as a unique point within the distribution. One participant described years of suffering resolved within minutes once they treated themselves as a population of one, against the standard of care. The promise of these models is to fold the individual back into the evidence. The discipline that keeps it honest, as one physician put it, is knowing how the person in front of you differs from the average person in the trial, and carrying the uncertainty that follows. A counter-voice warned that a sample of one hides the harm only populations reveal. The reconciliation may be technical: methods first built by astronomers to read a single strange star already let group and single-subject studies run inside one trial. The instruments exist; the will to use them lags by decades.
II. The Rocket Engine and the Dirt Roads
The evening’s most repeated image was a rocket engine dropped into a world of horses, buggies, and dirt roads. We have built something powerful and set it down where there is no car to hold it, no highway to carry it, and no rules of the road. The real labor is unglamorous: rebuilding workflows that live only in a few overworked heads, redesigning incentives that punish the efficiency the technology delivers, and reinventing the business models and training pipelines the old order relied on . One founder named the operator’s dilemma, the reluctance of capable people to unmake what they have mastered. Another, who watched a core system rebuilt in seventy-six days rather than a year, drew the lesson: the gains came from redesigning the entire system, and no team that bolted the new engine onto the old frame could repeat them.
III. The Data Question
Beneath both threads sat the question of data: who owns it, where it lives, and what we cannot collect. One attendee asked whether anyone in the room could say where their own medical records are and hand them to a doctor tomorrow. Few could. Models trained on tidy, curated benchmarks fail the moment they meet a messy hospital. The starvation is uneven, with women’s health and neuroscience running on thin data while well-fed fields race ahead. The richest signals may be the ones we do not yet gather: the trust and repair that pass between two people, the voltage across living cells, the signals a clinician reads without naming. A model trained only on the doctor’s notes inherits a thin shadow of what the doctor knew.
TWO META-THEMES
The Words Become the House
Running under every theme was a quieter one about language. It surfaced first as pragmatics, the meaning a sentence carries beyond its plain words, and the room kept returning there. One operator called the people in her studies participants rather than subjects. A physician proposed retiring “adherence,” with its faint scolding, for “participation,” which invites a person into their own health. The line that gathered it all came as a paraphrase of Hafiz: the words you use become the house you live in. How we frame the technology, and how we frame a patient, shapes what we build and how it feels to live inside it.
Psychology Drives Biology, and the Human Asks to Be Seen
The evening’s deepest current was not technical; it was human. Speaker after speaker described people who had optimized every marker and grown more anxious, sometimes less well, for it. One named the engine under that striving as fear. A physician of thirty years offered the correction: we are bags of variables driven by a psychological process, so reading biology while ignoring the mind goes wildly wrong. The same conviction shaped the room’s caution about machine companionship, since a model that mimics the surface of care misses what a human absorbs across a hundred and fifty faces in a clinic. More than one person named the unease worth holding: that as this leverage grows, our organizations, and perhaps we ourselves, grow a little less human. The work ahead is to let the machine carry what it carries well, so the people inside the system have more room to be present.
The October gathering in Sonoma will carry these questions forward. Our thanks to everyone who threw a flashlight toward the center. The light was bright, the company was kind, and this is only the beginning.
VOICES FROM THE ROOM
The Lines Worth Carrying Forward
Selected quotes, lightly edited for clarity. Per the Chatham House rule, none are attributed.
On the population and the person
“The valid sample set for an individual is one.”
“We are just one dot in a Gaussian distribution, and the hope is that [artificial intelligence] can help hyper-personalize us into the study.”
“If all you do is quote a clinical trial and decide on that, doctors will be replaced immediately. Your job is to figure out how this person differs from the average person in the trial.”
On the rocket engine and the dirt roads
“Imagine a world of horses and buggies and dirt roads. And then, all of a sudden, someone invents a rocket engine. That is where we are.”
“How do you build the roads around the car? How do you build the rules for how to drive? These are the things you have to figure out to deliver abundant healthcare.”
“Redesigning from the ground up is the bare minimum.”
On the data question
“Do you know where your medical records are? Could you give them to a doctor if you needed to? I do not know if anyone can answer that in the positive.”
“Precision medicine is focused on the individual. The next era is about what happens between people, not just within them.”
On the words we use
“I say participant, not patient. I say human help, not drug development.”
“Stop using the word adherence and start using the word participation. If people participate in their health, you are good on the outcome.”
“The words you use become the house you live in.” (offered as a paraphrase of Hafiz)
On psychology, biology, and the human
“They are some of the most optimized people I know, and some of the most unhappy and unhealthy.”
“My desire to optimize my health comes from fear. If I am not optimized, I will not be okay.”
“We have this leverage in front of us, and somehow we have lost touch with our humanity. That bothers me.”
And a line for the road
“Rooms like this are what it is going to take to get to a paradigm change, instead of an incremental one.”
This story was developed with AI-assisted editorial support from notes and themes gathered during the evening. The final interpretation, structure, and language were reviewed and approved by Jordan Shlain 🙂