Teaching an AI to explain ENT like a surgeon
I added a small tool to this site: you type an ENT topic, and it writes you an overview. My first version was… fine. It was also completely replaceable by Google. This is how I made it into something only I could offer — and the lesson that came with it.
The problem with "fine"
The first version produced competent, accurate, readable summaries. It was also indistinguishable from what you'd get from any search engine or chatbot. And that's the trap: if a tool can be replaced by Google, there's no reason for anyone to use it. That bothered me more than I expected it to.
What do I actually have that Google doesn't?
The honest answer to "why would anyone use yours?" isn't better information. Information is a commodity now. What isn't a commodity is perspective. I spend my days operating and being taught by people who think at the table — surgeons who care about the anatomy you protect, the moments where a case goes sideways, the judgment that doesn't fit in a textbook definition.
The model is a commodity. The way a surgeon thinks is not.
So I retuned the tool to answer the way I'm taught: the anatomy that actually matters in the OR, the danger zones, when to operate and when not to, the pitfalls, the pearls. Not a definition — a way of thinking about the problem.
The difference
Ask the new version about a parotidectomy and it doesn't just describe the gland. It talks about protecting the facial nerve, the danger-zone triangle where the trunk hides before it branches, the judgment calls a surgeon actually makes. That framing isn't something a generic model reaches for on its own. It's the lens I gave it.
The honest caveat
It writes original overviews for students and peers. It is not medical advice, and it's no substitute for textbooks, guidelines, or real supervised training. I'm a resident — I'm still learning a lot of this myself, and the tool reflects that, not some finished expertise.
The takeaway
The interesting use of AI in medicine, to me, isn't replacing the expert. It's amplifying a specific expert's way of seeing a problem. The engine is off-the-shelf. The judgment you point it with is the part that's yours — and that's the whole game.