Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

I’m a doctor: Here’s what I found when I asked ChatGPT to diagnose my patients

inflecthealth.medium.com

Post reply on HN
Print view View original

121–130 of 368 posts

Re: I’m a doctor: Here’s what I found when I asked ChatGPT to diagnose my patients

#121

Post by matt_heimer »

I argued with someone (online) about a ChatGPT diagnosis recently. They have back pain and they've had an MRI but the scans (shared online) don't show any significate disk bulging and no herniations. They put their symptoms into ChatGPT and got a possible diagnosis of a disk herniation along with a treatment of a microdiscectomy.

Despite the fact that a couple doctors have told them they don't need surgery they are convinced that they do. I understand that they are desperate for a solution to their pain but they are now doctor shopping until they can find someone willing to perform a procedure that ChatGPT suggested.

People are already being misled by these systems.

Re: I’m a doctor: Here’s what I found when I asked ChatGPT to diagnose my patients

#122

Post by QuantumGood »

When GPT starts to auto-incorporate the best yet known prompts, then we'll have a better idea of its potential.

You must, must use the best prompts, of which many are not widely known, and some have not (of course) been discovered ... yet. Even with human experts, you must provide sufficient detail, and the expert must ask clarifying questions for differential diagnosis.

Re: I’m a doctor: Here’s what I found when I asked ChatGPT to diagnose my patients

#123

Post by lanstin »

Atsuii wrote: post #104
tejohnso wrote: post #102

> about 8% of pregnancies discovered in the ER are of women who report that they’re not sexually active. This is the most surprising thing I read in the article.

As woman this doesn't surprise me at all. There is a lot of circumstances where a woman would not want to admit to an ER that they are sexually active; Religion, the sex was non consensual, they are with someone who doesn't know they are sexually active.

I think we will learn more about the biased sample that transcribed speech has vs private speech. I have just started thinking about this but there are huge area of speech that I use with my family and wife, and coworkers around the coffee room, that I would never consider putting into writing or a YouTube or whatever. And more types of things that I might say when with some younger folks while, trying to debug s production emergency but would not make it into the post mortem. The training I guess is not going to be able to benefit from these more private speeches, or maybe we will have to have people become convinced they are sentient and fall in love with them and share their whole humanity with the LLM.

Re: I’m a doctor: Here’s what I found when I asked ChatGPT to diagnose my patients

#124

Post by egl2021 »

petilon wrote: post #107

On the other hand, I have multiple minor issues where doctors have not been able to offer a diagnosis (they just say "I don't know") and ChatGPT has been able to offer multiple possible diagnoses.

"I don't know" is exactly what I want my docs to say when they don't know.

Re: I’m a doctor: Here’s what I found when I asked ChatGPT to diagnose my patients

#125

Post by Waterluvian »

matt_heimer wrote: post #121

I argued with someone (online) about a ChatGPT diagnosis recently. They have back pain and they've had an MRI but the scans (shared online) don't show any significate disk bulging and no herniations. They put their symptoms into ChatGPT and got a possible diagnosis of a disk herniation along with a treatment of a microdiscectomy. Despite the fact that a couple doctors have told them they don't need surgery they are c…

You said it: they’re desperate. People do considerably less rational things in a desperate attempt to make pain go away.

Re: I’m a doctor: Here’s what I found when I asked ChatGPT to diagnose my patients

#127

Post by paraxion »

I wonder if, instead of asking ChatGPT for a diagnosis, he could've got it to prompt for further questions he could ask?

My thinking is that given the nature of LLMs of connecting related information, it might be a good way to figure out the gaps in the diagnostic process, rather than actually provide one.

Re: I’m a doctor: Here’s what I found when I asked ChatGPT to diagnose my patients

#128

Post by jacobsenscott »

QuantumGood wrote: post #122

When GPT starts to auto-incorporate the best yet known prompts, then we'll have a better idea of its potential. You must, must use the best prompts, of which many are not widely known, and some have not (of course) been discovered ... yet. Even with human experts, you must provide sufficient detail, and the expert must ask clarifying questions for differential diagnosis.

This is nonsense - a LLMs output is a pseudo random mishmash of its training data. How can there be a "best prompt" when the same prompt gives you different output each time, but there is only one correct output?

Re: I’m a doctor: Here’s what I found when I asked ChatGPT to diagnose my patients

#129

Post by recursive_loops »

matt_heimer wrote: post #121

I argued with someone (online) about a ChatGPT diagnosis recently. They have back pain and they've had an MRI but the scans (shared online) don't show any significate disk bulging and no herniations. They put their symptoms into ChatGPT and got a possible diagnosis of a disk herniation along with a treatment of a microdiscectomy. Despite the fact that a couple doctors have told them they don't need surgery they are c…

I've recently been using ChatGPT to learn about Linux networking stuff, including most recently how to setup a bridge connection for KVM that behaves the same as virtualbox's.

Given how catastrophically wrong (confidently wrong at that) ChatGPT has been, I cannot even imagine the frustration that will be for doctors from people who don't understand how LLMs work and think they are "thinking."

Re: I’m a doctor: Here’s what I found when I asked ChatGPT to diagnose my patients

#130

Post by Xcelerate »

“I’m employed as an [X]. Here is why machine learning won’t ever replace my job. Even once the point has been reached where all overwhelming evidence points to it being better than me in every regard, I’ll still summon an emotional appeal to that ‘magical’ human quality that machines will just never replace.”

To be honest, I think I’d rather be friends with ChatGPT than most humans as it continues developing over the next decade.

Post reply on HN