The Walking Encyclopaedia Is Obsolete. We Are Still Training For It.
For a century we selected doctors for what they could memorise, because the physician was the storage medium. That problem is solved. The skill that remains scarce is the one we barely test for: making sense of a patient when the data disagrees with itself.
Dr. Sven Jungmann
CEO
On a morning ward round, the registrar recites the patient's potassium, the creatinine, the last three blood pressures, the antibiotic and its dose, without once glancing at a note. It is an impressive performance, and for most of the last century it was the performance we trained for. The physician was the storage medium. When the nearest reference was a library down the road, the doctor who carried the most facts in her head was, quite reasonably, the safest doctor on the floor.
We built an entire profession around that premise. Medical school was, in large part, a feat of endurance memorisation: the Krebs cycle, the brachial plexus, the long tail of rare conditions you would meet once in a career. The implicit contract was simple. Hold enough in your head, and you are competent.
That contract has quietly expired. Recall is no longer scarce. A machine now knows the dose of vancomycin, adjusts it for renal function, and checks it against the current guideline faster than the registrar can clear her throat. The skill we spent a century selecting for has become the cheapest input in the room.
“The dangerous doctor of the future is not the one who has forgotten a fact. It is the one who accepts the machine's fact without the judgement to test it against the patient in front of them.”
Storage was never the hard part
There is a temptation to read this as a story of decline, of doctors deskilled by their tools. It is the opposite. Memorisation was always the lowest-value thing a trained mind did; it simply happened to be the thing we could measure, so we measured it and called it talent. What we could never reduce to an examination was the harder work that good clinicians were doing all along, often without naming it: holding two incompatible signals at once and deciding which to believe.
Conan Doyle understood this. Sherlock Holmes is supposed to have neither known nor cared whether the Earth went round the Sun; he refused to clutter his mind with facts he could look up, on the grounds that the lumber crowded out the work that mattered. His genius was never the data he held. It was the connections he drew between things that did not obviously belong together. That, not retention, is the part of medicine no machine has yet taken.
The grey zone is the job
Routine cases were always going to be automatable, and they are. The protocol exists, the answer is in the guideline, and a well-built system will reach it more reliably than a tired human at three in the morning. That is a gain, not a loss.
What is left for the physician is the grey zone, and it is most of what makes the work difficult. The lab value that says the patient is fine when the patient plainly is not. The three guidelines that point in three directions for a person who fits none of their cohorts. The social history that quietly explains the readmission no algorithm flagged. The question is no longer “what is the protocol for this?” but “why does the protocol not fit this particular patient?” — and the second question is the entire job.
What an interview should test
If that is true, then much of how we still assess young doctors is a measurement of the wrong thing. An interview question that a search engine can answer in a second tells you almost nothing about whether the person can practise medicine. Asking a candidate to recite the criteria of the Wells score sorts them by a skill the institution no longer needs.
The more revealing question is harder to mark. Put three conflicting recommendations in front of a candidate, describe a patient who fits none of them cleanly, and ask them to build a defensible case for the path they would take and the rule they would break. You learn quickly whether they can hold the contradiction without flinching, or whether they reach for the nearest protocol to make the discomfort go away. The first is the doctor you want when the machine is confidently wrong.
None of this asks less of physicians. It asks more, and of a kind that is harder to teach and harder to fake. The best doctor is no longer the one with the most answers. By now the answers are ambient, almost free. The scarce thing, the thing worth selecting and training for, is the judgement to know which answer the patient in the bed is the exception to.
