Hacker News Reader

Comment by paulryanrogers | Hacker News Reader

> ...only if you feed it perfect information, and the actual patient has a classic presentation, which is rarely the case. (And you don’t mind a 50% success rate that often misses life-threatening conditions.)

This sounds like a step backwards?

View on HN

It isn't a doctor, it's trained on Reddit threads and other random shit off the internet not medical journals.

Given that "The Pile" includes PubMed and NIH as data sources it would be unlikely to have GPT4 not use them at all. Even GPT3 uses Wikipedia which does have (mostly) factual data with cited sources.

Problem is half of the publications or even more is pure garbage that cannot be reproduced so there is that.

> Even GPT3 uses Wikipedia which does have (mostly) factual data with cited sources.

There's a LOT of stuff on Wikipedia where the source is a link to some random, long article and it's unclear where exactly the referred to information is coming from. Gets significantly worse for any "hot" topic.

It's lossy though, it tries to remember everything and relate that to everything else. Performance would increase significantly if it were tuned for that use case.

You should try asking it detailed questions about medicine. It’s fairly deep and broad in its medical “knowledge.” But it’s not good at deductive or inductive reasoning and has no agency, so is entirely unsurprising it’s not good at differential diagnosis. We actually have good differential diagnosis expert systems the challenge is getting the providers to input queries properly. I can imagine GPT4 acting as an intermediary between human natural language and expert systems to great effect. It seems rather rash to be judging these technologies based off a crappy web UI thrown on a chatbot a few months ago.

Providers don't input queries "properly" because differential diagnosis software is complicated, clunky, and not easy to use for the overworked clinician.

Then you haven’t had much experience with doctors… you go in and you’re most likely getting either a cast, a broad spectrum antibiotic, or an antidepressant. The only things that get specific diagnoses reliably is when somebody finds a tumor.

In the last year I had to get a full antiparasite treatment course because a doctor misdiagnosed my partners psoriasis… that was fun.

True, 90% of doctor diagnosis are garbage and maybe even worse than ChatGPT. The 10% is for when you have a visible problem that can be cured with a surgery, they are pretty good at that. A doctor with GPT-4 could be way better than one without.

They already have diagnosis flowcharts where it’s a game it follow the symptom. It would have to be considerably better than the tools they already have to be appropriate.

Reply on news.ycombinator.com