Anonymous at Sat, 29 Mar 2025 08:21:20 UTC No. 16631570

both regurgitate whatever propagandas they're fed on so it's a toss up

Anonymous at Sat, 29 Mar 2025 08:27:26 UTC No. 16631572

>>16631560
Yes goy, the chatgpt will do the surgery on you. Chop chop

Anonymous at Sat, 29 Mar 2025 09:06:52 UTC No. 16631590

>>16631572
And the grok will guide you in nutrition! Eliminating the need for surgery in the first place.
Prevention over cure! - The benefits of objective advice!

Anonymous at Sat, 29 Mar 2025 09:12:11 UTC No. 16631593

>>16631590
>advice
Slightly misworded.. information*
The current chatbot setup encourages endogenous reasoning, greater informed decision making possible with this route as opposed to the current “trust me” strategy

Anonymous at Sat, 29 Mar 2025 10:59:09 UTC No. 16631687

https://www.youtube.com/watch?v=4ntfQv1OqoI

Anonymous at Sun, 30 Mar 2025 20:56:22 UTC No. 16633124

>>16631560
At raw diagnostics when presented with all the data? Yes, but that's about it. It's not particularly great at actually gathering the data, and it obviously can't do procedures.

Anonymous at Sun, 30 Mar 2025 21:19:39 UTC No. 16633138

>>16631560
Yes, but they're both pretty retarded. Google Scholar works a lot better.

Anonymous at Sun, 30 Mar 2025 21:21:34 UTC No. 16633141

>>16631572
>>16631590
>>16631593
>>>/pol/
>>16631560
ChatGPT shouldn't be a licensed doctor, but it's capable as a tool for doctors. Hope that helps answer your question

Anonymous at Sun, 30 Mar 2025 21:48:48 UTC No. 16633166

>>16631560
I showed ChatGPT my penis and it's said it's lupus

Anonymous at Sun, 30 Mar 2025 21:56:00 UTC No. 16633169

>>16633166
maybe dont fuck period hoes with open wounds

Anonymous at Mon, 31 Mar 2025 00:24:01 UTC No. 16633250

>>16631560
Bro if there were a way to get prescription drugs without a doctor, but still safely, nobody would ever go to a doctor again unless they needed surgery.

Anonymous at Mon, 31 Mar 2025 00:48:38 UTC No. 16633270

>>16633250
When can ChatGPT start prescribing me oxycodone for my chronic pain?

Anonymous at Mon, 31 Mar 2025 01:20:04 UTC No. 16633299

>>16633270
Don't worry anon, shitting and breathing are primitive needs of the past now that we're in the age of AI

Anonymous at Mon, 31 Mar 2025 01:24:44 UTC No. 16633304

>>16631560
The thing is that Chatgpt is drawing on a large body of work, but our understanding of medicine is still so limited that the current literature is still largely deferring to clinical judgement on the differential and treatment options. At best, LLMs have a role in medical education on the wards and image recognition software can help with menial tasks that are currently delayed by hours in even the best hospitals in America (EKG/EEG interpretation, plain film reads, path specimen analysis)

The issue is that LLMs suck at gathering the information, so a doctor still has to know everything about their specialty to conduct a proper H&P on a new problem.

Anonymous at Mon, 31 Mar 2025 01:51:45 UTC No. 16633337

>>16633141
Quit with the gatekeeping of medical knowledge.

Anonymous at Mon, 31 Mar 2025 04:05:51 UTC No. 16633404

>>16633337
reasoning LLMs are free you nigger im not gatekeeping anything

Anonymous at Mon, 31 Mar 2025 04:07:26 UTC No. 16633407

>>16633304
>The issue is that LLMs suck at gathering the information
They are starting to suck a little less though, one of the services out there does research strictly from medical sources and cites them

Anonymous at Mon, 31 Mar 2025 05:31:02 UTC No. 16633456

>>16633407
Yeah but most of the guidelines we have written down are still super vague - medicine and surgery in 2025 is still using the apprenticeship system because there is no good way to actually teach the perfect decision for every patient, we just don't have enough info. If you ask a clinical question of any significant complexity to OpenEvidence, it just sputters something about clinical judgment back out

Anonymous at Mon, 31 Mar 2025 06:14:27 UTC No. 16633482

>>16631560
I am in med school and ChatGPT can answer everything single one of our practice questions perfectly. I can bet by the end of this century every single non hands-on doctor will be out of work, if not sooner. Many physicians delegate all their dirty work to nursoids anyway (preliminary bloodwork, biopsy collection, x-rays) and do nothing but read charts and diagnose and prescribe, and no doubt hospital admins will start drooling at the idea of a machine that doesn’t need to be paid 200k (or get sued) to do the exact same shit (Looking at you radiology, pathology, neurology, psych).

Anonymous at Mon, 31 Mar 2025 06:17:00 UTC No. 16633485

>>16633482
Thinking that radiology, pathology, or ESPECIALLY neurology and psych will be replaced by a talking monkey like an LLM makes it clear you haven't even gotten through first year yet, let alone the wards. Psych will never get replaced by a machine, period

Anonymous at Mon, 31 Mar 2025 06:25:25 UTC No. 16633487

ai cant even consistently draw 4 fingers and a thumb on each hand.

Anonymous at Mon, 31 Mar 2025 20:07:21 UTC No. 16634162

>>16633485
t. mad rad

Anonymous at Mon, 31 Mar 2025 20:22:25 UTC No. 16634177

>>16633166
Same here but it said I had elephantiasis.
>>16633407
>and cites them
Which is what it should do but I'm concerned doctors will get lazy and just accept whatever the LLM puts out uncritically. Whatever the LLM's accuracy rate, it's still wrong 100 - that rate percent of the time. Since LLMs are quite often in different ways than humans tend to be wrong, having both the LLM and the human doctor make judgements in theory should lead to better outcomes but we know human doctors, because they're human, will get lazy or simply start to defer to what it sees as a medical authority in the LLM.

Anonymous at Mon, 31 Mar 2025 20:34:29 UTC No. 16634185

>>16633485
Ten years ago they said artists would never be replaced by machines, too. Don't be so sure you know what the future will bring.
>>16634177
Considering the exploding numbers of useless pajeet doctors in Western medical settings, I think uncritically accepting LLM output might lead to better results overall.

Anonymous at Mon, 31 Mar 2025 20:46:03 UTC No. 16634197

>>16633482
I'm not in medical school but if you give me a list of all your questions, along with the answers, I'll be able to give you the answers to your question.

Anonymous at Mon, 31 Mar 2025 22:18:39 UTC No. 16634252

>>16634162
Actually going into surgery, just being realistic about AI

Anonymous at Tue, 1 Apr 2025 01:18:23 UTC No. 16634426

>>16634177
If a doctor has good judgement then they will only leverage an LLM's generally high accuracy in situations that the doctor suspects his own judgement can't resolve a rare problem. For rarer diseases it still leverages high accuracy for those. Plenty of students in medical school talked about this.
>>16634185
>Ten years ago they said artists would never be replaced by machines
We're never going to be at a point where artistic creation is replaced until it's killed in full. Artists and machines still produce things in parallel; at most, machines are currently being used as exoskeletons of creativity, but that doesn't fundamentally "replace" the human involvement of creation. It requires a full eradication of the human race for true replacement to be possible.

Anonymous at Wed, 2 Apr 2025 11:23:51 UTC No. 16634921

No its, hallucinating on every answer. It cannot think. Watch how chat gpt thinks in the reason mode. It goes into an introspection. Like a word based ego. Then attempts to answer the question. This won’t reflect reality. The thing is chatter. Wikipedia is closer to fact because it’s curated by experts. It doesn’t understand what its end result is or means. It relies on words and serves you unlimited language goop. It can’t understand what it’s saying but emulates how an expert would talk. Kind of like a grifter. It cannot be used except to do an exploration. It can’t tell me what’s fact.

Anonymous at Wed, 2 Apr 2025 11:26:39 UTC No. 16634923

>>16634921
kek funny

Anonymous at Wed, 2 Apr 2025 11:32:49 UTC No. 16634929

Yes, doctors have become that bad.

Anonymous at Wed, 2 Apr 2025 11:35:39 UTC No. 16634932

>>16634923
You can laugh but it can’t strictly work without interference. That’s why people have these science modules programmed for it. It strictly produces piles of (racist) crap by itself. That’s why meta dumped it.

Anonymous at Fri, 4 Apr 2025 06:48:30 UTC No. 16637023

>>16634426
and humans can still do math on paper in parallel with calculators, yet one of these we utilize more. You can mail a letter instead of sending an email, it's still possible to write out audio transcriptions by hand in lieu of speech-to-text programs. It's not a matter of capability but one of saving time and effort and especially money as long as humans remain evolutionarily naturally cruel and uncaring (newsflash, we are)
>>16633485
never say never, nigga. surgery may be safe for now, you should really worry more over nursoid creep and H1b degree-mills flooding in from the 3rd world to stagnate pay with an imminent great depression approaching