Post

The thing that irks me is that AI sounds exactly like a medical student: fluent, certain, and not yet wrong enough times to know better. That's fine in a med student, because med students have residents, fellows, and attendings to correct them. AI doesn’t have any of those.

Everyone finds AI's confidence annoying. But I think physicians find it dangerously familiar. We were trained to treat that voice as a hazard, we learned why on real patients, and now we're the ones expected to supervise it.

Sometimes I’ll hear a confident medical student presentation, and my initial reaction is “Damn, they’re really cocky.” But then I remind myself… they just haven’t been burned by biology yet.

No shade to med students — I know that voice of confidence, because it was mine, too. We’ve all been there.

Then residency happens. You see weird presentations of common diseases. You watch real, terrible things happen to normal people — things no textbook warned you were even possible, let alone probable — and whatever innocence you had left goes with them.

Those phrases sound weak to people outside medicine, but I know them as dead giveaways of how experts talk. The experienced physician isn't just accumulating facts, they're accumulating reasons not to trust themselves too much. (Yes, some attendings never got the memo. Hi, ego.)

AI has the fluency but none of the battle wounds.

Us humans? We’re leaking uncertainty 𝘤𝘰𝘯𝘴𝘵𝘢𝘯𝘵𝘭𝘺: we pause, slow down, hedge, stop mid-sentence while our brain works something out. AI flattens all of it. A well-supported answer and a wild guess arrive in the same streaming cadence and grammar.