The moment
I was having a conversation with an undergraduate student recently — not a clinical one, just a discussion about AI and what it means for people who have spent years building expertise the hard way.
I mentioned my PhD. Ten years. A literature review of just under eighty thousand words. They asked how long that had taken to write.
About seven years, I said.
Seven years of searching for papers, tracking them down, reading them, digesting them, writing paragraphs that took half an hour each and came out at maybe twenty words.
"So basically you were just stripping stuff off the internet and putting it on a thing? We reckon we could have done that in forty-eight hours?"
In conversationAnd the thing is — they are right.
The insight
The instinct is to be offended by that. And I was not. Not really.
Because what they were describing — finding information, collating it, structuring it, producing coherent written output — AI can now do in a fraction of the time it took me. That is patently true.
But here is what did not get produced in those forty-eight hours.
Twenty years of clinical practice that taught me which papers mattered and which were noise. The ability to read a methodology and sense that something was off — not because I had consciously analysed it, but because I had absorbed enough context to feel the wrongness before I could name it.
The AI produces the material. It cannot provide the context.
That word — context — is the whole argument. Every time AI produces a summary or a suggestion, I am running it through twenty years of analog experience. I am the validator. Without the validator, the tool is not just less useful, it is potentially dangerous.
A young clinician starting today will move faster. But they will not yet have the clinical mileage. And clinical mileage is not a function of information. It is a function of time in the room with people, being wrong and learning from it in ways that leave a mark.
That mark is the thing. And no algorithm generates it.
The bigger picture
The machine can now do the things that used to consume you. What does that free you up to do?
For me, the answer has been: be present. Stop typing during sessions. Use that time to think, to write, to share what years in the clinic has actually taught me.
The forty-eight hour PhD did not take forty-eight hours. It took ten years plus forty-eight hours. The forty-eight hours is what the machine does. The ten years is what makes the forty-eight hours mean something.
That is the deal. And for anyone who has done the work — in any field — it is, on balance, a very good one.
Next: When AI Chooses Your Doctor.