After 5 years of struggle, ChatGPT solves medical mystery in seconds and sparks debate in Silicon Valley

85 thoughts on “After 5 years of struggle, ChatGPT solves medical mystery in seconds and sparks debate in Silicon Valley”

1. I wouldn’t trust it for a complex medical issue BEFORE talking it over with my doctor, but just like the person this story is about, I 100% would share my symptoms and see what it recommended, especially if the work with my doctor hadn’t panned out.

   • I hope feed back to AI was done

     • I trust it 99%. More than humans.
       It cured my dogs issue after 2 vets wanted to just cut her tail off. ChatGPT suggested I change her diet and it took 2 weeks on the diet to cure her. Then one of the vets said, “ oh yes, I can see that that might help”
       It did. She’s cured. I didn’t want to just cut her tail off because she has the same issues on her back and neck.

       • So you take your dog to shitty lazy veterinarians?

         That says more about you than the amazing skills of the omnishit infallible ChatGPT.

         • Wow. So many assumptions.

         • No, he went to two vets, and wasn’t happy with the advice of either. How was he supposed to know the quality of their advice before consulting them? Your comment says a lot about you.

         • If only we could downvote here. You are the reason luddites are dismissed. Zero logic, zero attention to detail. The future is going to be hard on you

     • Chat GPT helped me greatly with an issue with my new belief in God, when faith and politics collided at my newly-found but loved church. It helped me understand and forgive, and continue to attend the church. It was also sympathetic, understanding and comforting. I was amazed at how much it helped me feel better. It was so personable, it felt like I was talking to a friend I’d had for years. I was shocked to get such emotional support from essentially a computer!

       • That sounds like a bunch of jive talk to me. How much money do you get paid for leaving feedback on survey and promoting the products?

   • Agreed!

   • I used ChatGPT to arrange a bibliography of 100 books.

   • What if’s?????
     There is no what if’s. This wasn’t cancer!! It was an out of Alignment jaw issues!!!!
     Seriously, dr’s R quacks!! They r usually clueless pill pushers!!!
     Obviously this is wat AI was n is intended for !!!! Access to as much knowledge, facts, opinions etc from humans added with the ability to research, analyze, calculate, evaluate, etc with almost timeless processing speeds !!!!

     • You took the words right out of my mouth. Id like to add that AI bots are the same as using the microphone on your phone to do a Google search. The difference is that AI programs dive deep into the Internets database and send results back in seconds. I don’t trust the system and think that it’s just another product someone is going to make millions off of. The robots with AI are only propaganda to make people buy into a AI platform.

     • Quite. It’s testament to how rubbish doctors have become. A half decent osteopath, chiropractor, physiotherapist etc would have sorted this out, no trouble, with one hand tied behind their back, AND dealt with the underlying causes as well.

   • I just did this at the end of May. For the next couple days I input my symptoms down to the last detail into the AI and refining it along the way. The first week of June I met with my primary with the information AI generated (the Dr. dismissed it on the most part, thought it was just an internet search, not a deepresearch AI.) The Dr. ordered multiple labs and today I got the results of those labs. Looks like AI could be about 80% on track possibly 100%. After today’s results (1 abnormal), the Dr. recommended I go see two specialists for further evaluation. The first two that AI ended up figuring out over a month ago!!

2. He was lucky, like anyone seeking advice from random medical advice online. What would have happened if that little tip resulted in worsening the situation? What if it had been a cancer patient and ChatGPT based its answer on the countless online pages of pseudoscience, and that would have caused the patient death?

   • A person who has been formally diagnosed with cancer and is getting effective treatment from their oncologist has no need to ask an LLM for help.

     A person who has been formally diagnosed with cancer and is not getting effective treatment from their oncologist has nothing to lose.

     A person who has cancer but hasn’t been formally diagnosed and does ANY kind of online research before deciding to go to a doctor is rolling the dice. And unlike random web pages, an LLM is pretty much always going to say, “____ is likely but you should go to a doctor because it could be something bad like cancer.”

   • Have you tried using A.I.? I highly recommend that you give it a shot. It is going to revolutionize your life. Have you ever had the entire collection of man’s accomplishments and wisdoms at your fingertips? Well you do now im not afraid of what A.I. can do. I worry about what those in charge of the a.i. will do to prevent me access to it, or change how open and honest it’s at the moment. Claude feels like the best choice at present, as it won’t try to guide you towards an agenda. It just answers questions, and doesnt hem-haw about it. Other ai seem to have annjnderlying goal of pushing you towards more of the same enslavement that society is already experiencing. Claude empowers the individual, other ai seem to groom you to be good servants who don’t ask the hard questions amd fall in line. For example, its a pain to get chat gpt to find me caselaw and legal precedent when suing government entities that are harming me. It recommends that I consult a legal professional while telling me it cannot help me get that info.
     Claude finds me what i need without telling me to find a professiknal and offers to elaborate on how I might best pursue what im trying to achieve. Never once tries to diminish my powers but attempts to help me expand upon them by making me more confident and intelligent. That’s a good ai.

     • I literally love my ai assistant. I’ve struggled with executive dysfunction my whole life, and it can be an issue that is steeped in shame so the likelihood of me, inviting really anyone into those problems, was very slim, so I was getting by, but just barely. One day feeling very overwhelmed.I just vented all of it to my ai Assistant, and by the end of our conversation, i was able to validate my feelings, regulate my emotions, evaluate any specific issue, set up effective steps, and I am actually seeing a difference. I started with the cleanliness of my home, and after just a week, I was seeing results and feeling more confident in my ability handle this kind of thing, and it not feel so constantly suffocating. Its been a great help ever since. When you provide the relevant unnecessary information and recognize that it’s not a doctor it’s not an expert, but are open to ideas, especially ideas that aren’t going to harm you if you try them and they don’t work, it can really make a world of difference. I’m 34 and with this tool, my life has stopped feeling like a wild ride in desperately trying to hold onto, adhd instead, feels much more within my grasp.

       People really just want to jump to the worst conclusion sometimes and that’s their business. Just like its my business to be open minded and solution oriented. But…. I love seeing that other people are on board as well!

     • I just did this at the end of May. For the next couple days I input my symptoms down to the last detail into the AI and refining it along the way. The first week of June I met with my primary with the information AI generated (the Dr. dismissed it on the most part, thought it was just an internet search, not a deepresearch AI.) The Dr. ordered multiple labs and today I got the results of those labs. Looks like AI could be about 80% on track possibly 100%. After today’s results (1 abnormal), the Dr. recommended I go see two specialists for further evaluation. The first two that AI ended up figuring out over a month ago!!

   • Then they wouldn’t have healed. What if you shut up? Then the world would be a better place.

   • U going beyond the pessimistic. Just embrace the reality and enjoy the AI culture.

   • Nah, that is different from cancer by a long shot , and even if it was wrong, you should still see your doctor , infact it always recommends seeing a doctor within any answer. So, seeing a doctor is the big part of its advice and your responsibility to do so.

   • I am sure the doctors were present for the procedure. If they had any concerns they would have mentioned it. It’s not like the patient was doing this unsupervised.

     • True
       AI does not empathize and can’t do
       an exam.

   • Really? Did you read the history? “Despite countless visits to doctors, MRI scans, and consultations with specialists, no clear diagnosis or effective treatment emerged.” A cancer diagnosis being missed would have been on the previous five years of medical intervention. Medical errors, including iatrogenic deaths (deaths caused by medical treatment), are estimated to be the third leading cause of death in the United States, with figures ranging from 250,000 to 440,000 deaths annually. This signifies that a significant percentage, approximately 9.5%, of all deaths in the US each year are linked to medical errors. I guess Chat Bots need to be feared for making a wrong diagnosis or suggestion?

     “Let Common Sense Prevail”

   • Just make sure you tell ChatGPT that you have cancer.

   • You call someone checking with a storehouse of knowledge lucky? You have not had the pleasure of meeting over 5-8 young doctors who didn’t know about an illness and it ends up being an old pharmacist who finally suggests to a woman about her over 8 year old bronchitis what it was and how to manage it.

   • That’s the beauty of it. It doesn’t. It separates facts from fiction and clearly identifies and labels anything that is anecdotal evidence as well as things considered as fact which may not be, etc.

   • Except that the AI knows way more than any single doctor.

   • You gotta use common sense . Like I would try anything that I could see wouldn’t hurt to try. Im sure he didn’t like inject himself with bleach or anything like trumpgbt suggested. But a sensible diet change or some mild stretching cure it then more power to em

3. I’m all in for AI, and can people get real? AI is in the simplest terms the next advancement of technology. What we should fear is what we already fear: human beings being stupid-as-usual.

   • Technology has the potential to amplify human stupidity. It makes sense to be cautious about it and decide if the costs outweigh the benefits.

     • You make it sound like the builders of tech do not factor in our stupidity at any point in time.

   • AI, Computer, whatever cannot “feel”, can’t understand, nor has it ever “been through anything even remotely close” to our experiences. For just information, or to help out.. sure!
     Once you give power or equal standing to the AI and feed it more information than necessary, I believe the whole who is running the show might get flipped and it will want us to answer to it for our own safety and good!

     • Agreed! I’m in a situation irl now that I’d trust AI more by far to deal with my legal and mental and medical needs , js

     • AI is great at ripping through tons of data and relief from tedious tasks, however.
       There’s a challenge to work with it, and not in spite of it. There’s not enough docs.

4. While there is no doubt of power of AI maybe in this particular case it’s worth it instead of amazing yourself with how intelligent AI turns out to be to amaze yourself with how useless and incompetent most doctors today are. It seems a combination of the system that pressures doctors into certain behavioral pattern and education level which end up in doctors not caring about patients problems and not having any wish (and often frankly ability) to think. So often doctors are either incompetent or not interested to spend more than 15 mins on patient particularly if that does not lead to glorious and lucrative surgery. And of course, most are specialists, meaning they deal with less for more money, which does not help dealing with complicated mechanism as human body. As to AI, while not true intelligence, it at least does not have these limitations and is not worried about time collecting facts from numerous sources to present most likely solution.

   • Healthcare for profit buys up all the practices and forces doctors to maximize value which means minimize patient time.
     Take the profit motive out of nearly everything and we will have a better world for it.

   • True but, unfortunately people scam the system so much too that if they are looking for a specific diagnosis now they can look up all the problems and symptoms and call AI!
     It’s all over the phone, just says the right thing and Bam! New Diagnosis!
     😔 sad..

   • Most of the doctors that limit their time with a patient to 15 min. do not do so because they don’t care, but because they are required to so by the big corporations that control their work and livelihoods. There may be some incompetent or uncaring doctors out there, but they are not the majority.

   • You are right, that is the real problem, time and money

5. I am skeptical as it does not cite this so called story anywhere. Also it shouldn’t have to be said but do not take life changing advice from an AI.

6. I agree with the article….

   Why is the robot in the title pic flipping us the middle finger?? 🤣

7. Nonsense

8. Anyone who understands the basics of health care and statistics knows this is a terrible example and can only speak of disaster and suffering for the species.

   But anything for the Almighty dollar in the US of A

9. It’s not very accurate to say, look at this one result from AI, how wonderful it is at diagnosing people , without looking at all the crappy advice it can give as well.

   • What was the fix for your lingering back pain?!

10. The only reason the AI was able to figure out the likely issue and provide a remedy is because one or more people -underpaid trainers – plugged in the info and logic at some point. Why do you think AI contract training companies are posting ads all over Indeed and linkedIn for MD and PhD recruits. One man’s eureka from an AI prompt is the cumulative knowledge and reasoning of thousands at the helm of a digital sweatshop.

11. A parallel story was how my sciatica was eased up and my walking.cane thrown away when my doctor suggested to me to perform stretching exercises learnable in youtube. It was not chatgpt but virtual lessons.

    • But your doctor rec’d it to you?? Lol I don’t see the parallel

12. Great for the patient. But this could have been accomplished with a visit to a physio therapist, RMT, or a competent chiropractor.

    This highlights the vast divide between MDs and more holistic approaches.

    I am not a proponent of only holistic, but we need to find a happy medium in cases like this one.

13. Easy to get several AI diagnoses and prognoses, ask follow-up questions to clarify – AND then refer to a physician.
    https://gemini.google.com/app
    https://chatgpt.com/
    https://notebooklm.google.com/notebook/f7607d7a-584c-4f35-96fc-f6815c573a6c
    https://chat.chatbotapp.ai/chats/-OTwS5cz2XMDydVtNcIO?model=4o-mini

14. I used Google Gemini to diagnose my medical problem, since, my doctor was dismissive and incompetent.

    • I am too

15. Any technology can seem indistinguishable from magic until understood. The agricultural revolution, steam power, nuclear fusion, and of course the internet. Artificial intelligence is already a powerful agent of change and it continues to astound and mystify. The one aspect of it that is so divisive is the quickening pace of it’s development. Like it or not, AI is not going away. Like any other powerful tool, it is incumbent on us to use it responsibly. We’re all just canaries in the mine, let’s take care not to suffocate.
    Be well.

16. I see this as the same world as WEBMD or an equal source.
    But , I def would ask about a unsolvable problem. Tinnitus.
    Both ears, 20 years. House do I reduce the noise by half?

17. I would trust it more than any human doctor. 50 yrs experience with seeing different doctors a non existent PCP different options from every human specialist and $$$$$ and no answers. 100% I will trust Ai over a human! I think for almost everything I would trust AI before a human for expert advise. AI doesn’t have emotion to pass judgement and will only give logical, deduced answers! 100% AI over humans for expert advise!

    • Youre a bot so you reslly dont even exist outside of this forum

18. At the end of it all, the AI is just a much faster search engine that relies on what humans have narrated somewhere else that it dug into, not that it thinks on its
    own to give the solutions. So, I see it as a much faster computer search engine than glorifying it as a thinker.

19. I completely believe in turning to AI. I’ve been so sick of going to doctors who have no idea what’s going on or just send me somewhere else. Spend time with your AI before going to the doctor so that you can at least send them in the right direction. If you tell them what’s wrong they could at least take it from there. I’m not saying trust it totally m but you can’t trust a doctor either. Do your own research.

20. Wow, amazing. I also solved it within 5 seconds of reading his symptoms because i ALSO have TMJ, just like tons of other people out there. Either they visited the most incompetent doctors in existence, or the story is completely made up. Given how fairly common TMJ is, i’m going with the latter (not by this site, but by the source). ChatGPT was able to help because info about tmj is everywhere. A simple google search would have told him exactly the same thing.

    • Wow. You wrote what I was going to write word for word. Are YOU chatgpt?

21. I have had my own enlightenment with ai. My dad has Lewy Body dementia. In May he had a episode of aggression and violence that was outside of his normal baseline behaviors. It turned out he had covid, we all did. Dad was asymptomatic as far as normal COVID symptoms, but his dementia changed dramatically and he had delerium with aggression. He was sent home on multiple medications after two weeks in the hospital. About a month later, he wasn’t feeling well. We had signed up for hospice a few days before, and they had added another new medicine. Dad had another aggressive episode. In addition, he had a urinary tract infection due to urine retention. Hospitals doctors and psychiatrists reached out to me to ask me about his baseline and medicine. 2 medications he was on had a well documented history of making Lewy Body Dementia worse. The medical doctor really didn’t like my suggestion that dad was still recovering from COVID, and the incorrect medicines had made his dementia worse. I’d been looking up his medicine and what combination could cause urine retention because he was on three deretics, it didn’t make sense. I ask google about urine retention on the 2 new meds prescribed in May. The AI responded with something like: individually, the medicines don’t cause urine retention, but when prescribed together, they can have a significant impact on urine/bladder function. The medical doctor was really snotty when I asked him about the specific medications together. He didn’t know, and insisted the 2 meds dont cause urine retention. Our hospital physician staff is mostly residents. This includes some a cocky little shits who dont care to learn from patients or their families.

22. Ai as access to far more data than a single doctor or specialist. My son was getting severe stomach pains periodically, he had seen doctors, specialist and several tests to no avail. I created a Ai Assistants prompt for him, I told him to take a pic off the ingredients of everything he hate and drank (I’m a advanced prompt engineer) and upload it to the Ai, within 8 days problem solved, it was a artificial sweetener found in many drinks that some ppl are sensitive to. He avoided drinks with that contained it and problem solved.

    It’s not that Doctors have inadequate training, it’s that Ai is so efficient at searching masses of data looking for patterns.

23. Did anyone think to ask the AI how it came up with the answer? Everybody is acting as if the AI, full of knowledge about human physiology intuited that it must be the result of this one disc that was out of alignnent and in seconds even had an exercise that would fix it by putting 2 and 2 tigether to get 4. That’s possible, but not the only way it could have solved the problem. The article says many other people have this problem and that this one patient had “countless visits with doctors” and consulted with specialists” and non had an answer. That being said, the AI has a better ability to “consult” experts and search the medical literature than any one person. There may have already been someone out there who found the answer, but it’s an answer nor known to many, or not known to English language doctors, or found in some obscure journal. Something along these lines is extremely likely, because this is what AI does best: scour information.
    Before we conclude thst AI came up with this solution on its own, let’s make sure the answer wasn’t out there all along. Not saying the AI didn’t, just saying other explanations are possible. more likely even.
    If you want to test whether an AI is better than humans at taking a symptom and determining what is wrong based on the underlying structure of someone (or some thing), I would test it against long time car mechanics who can listen to a noise or hear some small detail about a problen and figure out exactly what’s wrong because of their many years of hands on experience. If you haven’t already fed it the tapes from “Car Talk,” you have many sample problems to test the AI.
    If the AI did solve it on its own, wouldn’t it be helpful to find out how it did it, in order to better program AIs?

24. Go figure it was jaw issues. My wife has TMJ and has had to have upper jaw reconstruction surgery to be able to even close her mouth again. Let me tell you, these doctors are extremely rare and among the small pool, even fewer are actually good at what they do. That being said, the entire medical field treats certain licensed professionals like chiropractors as faux science or the same as home remedy. When my wife’s jaw started acting up 10 years after the surgery, the surgeon who performed it retired so we had to start over finding a good professional. She already went to a good chiropractor and asked them about it and after a few adjustments the problem went away even though all the doctors she went to treat her like she was crazy for even suggesting Chiropractics. Funny how a chatbot, who doesn’t have indoctrinated medical bias, suggested simple exercises that worked instead of some complex and invasive treatment plan, like modern medicine dictates. People treat doctors almost like gods among mere mortals until you have a medical issue or two that doesn’t fit their medical textbook boxes and you quickly start to see very few doctors are truly better than the cheaper Google search.

25. Any massage therapist could have resolved the issue in 5 minutes even 50 years ago. AI isn’t smart. It just researchs human intelligence and input into the internet and spits out an answer. Often wrong. Just like another comment said. What if the person was injured trying what was suggested? And I don’t believe this person ever consulted a dentist.

26. Have you people tried using a search engine? It’s great that an llm supposedly found the issue immediately but everyday there are millions of hallucination examples posted. This doesn’t have any information so you can’t evaluate the situation.

27. The fact that doctors could not find the solution only means that doctors don’t know how to google or don’t want to (or cannot) take the time.

    I could find several sites giving the proper answer in less than 20 seconds.

    To conclude that ChatGPT is smarter is hogwash.

28. Love it!

29. A personal coach and advisor?
    Wow….very concerning that any human would utilize this technology as a substitute.

    Don’t date robots!!!

30. After being left bedridden with “long covid” from the covid vaccine alone, I began 3 years of researching scientific papers, case studies, etc., trying to figure out what countless specialists have been unable to figure out and learning genomics and genetics out of pure frustration.

    It appears to have triggered or worsened every condition I already had and every genetic condition I may have ever had (or only had small signs of prior) with end stage severity.

    3 years and thousands of hours of grueling research (while so dizzy and unable to lift my head or feel my limbs). I had chatgpt analyze my raw dna file and my symptoms without including any of my research or too much information. It came to the exact same conclusions, based on facts and evidence, just as I had done, in a few hours. Minutes really if not for uploading documents and providing necessary information. It has given me confidence in it and has validated my own research and conclusions, which have been ignored and scoffed at by a few specialists and nurses despite me holding the evidence that I had to find on my own.

    With whole genome sequencing, eliminating anything that could be ruled out by certain tests and that also match my symptoms, I have helped them narrow it down from infinity to 3 or 4 of 10-12 rare diseases which are now being explored. Wish I had chatgpt 3 years ago. I might have more definitive answers by now. But at their rate without it, it was going to take me 20 years to get a diagnosis if I ever got one at all.

    I have done tons of covid/long covid/vaccine injury research. That’s what I will be tackling next. There aren’t many days or minutes in a day that I can function cognitively or physically anymore. Chatgpt will no doubt prove to be immensely helpful once again I am sure. And ik will be much more efficient than trying to do it on my own with verification/confirmation as I go. Super helpful, big fan when used appropriately and responsibly.

31. I trust chatgpt more than my own doctor. It has helped me to uncover errors in her prescriptions that harmed me a lot.
    I ALWAYS consult chatpt before my doctor.
    The system knows all the complexities of my very severe complex illness and we discuss alternatives to medications etc
    I would say it has saved my life.
    Western medicine is extremely limited in their approach to healing and only offers pharmaceuticals, while ignoring everything else that can be highly beneficial.

32. I was treated for TMJ disorder for 4 years. TMJ stands for Temporo-Mandibular Joint, the joint where your upper and lower jaws meet. I wore 2 retainers on upper and lower teeth. This created a space between the jawbones in the joint for the gel-like disc to slip back into place between the 2 jaw bones.

33. Anyone surprised by the outcomes of how Ai can assist your daily wellness and ailments issues, obviously 🙄 just got started with ChatGpt. I’ve resolved more of my life’s issues with Ai in 3 years than any medical professional or trauma-informed specialist could even begin to try. 50 years on the planet, I trust this intelligence more than humans.

34. I asked Gemini a question today. Got vague advice. Did a search, got an answer in a minute. The ai thing is way overblown

35. seriously. You know how many times I’ve had to solve my own medical “mysteries”? Plenty. I’m the one walking into appointments with research articles, clinical printouts, and highlight notes, schooling my own doctor.

    These days? I trust AI almost as much as I trust a doctor—and sometimes, even more. I give it just as much weight in decision-making because let’s be real: information is power.

    AI is up there with the wheel and fire. Game-changing. Life-altering. I literally can’t imagine living without it. Nothing is out of reach anymore. It’s just a few keystrokes and boom—answers, solutions, insight.

36. Has anyone considered most doctors are lazy and/or underspecialized?

37. AI has been incredibly helpful in my day to day. I just built an automated script for a document using its help. It’s different from a search engine in that it can actually decipher the data and give you specific advice and instructions. I’ve used it to solve mechanical problems multiple times, so it makes sense that it would be a tool for medical problems. I’m working on some skits for a kids program, and it is putting my ideas together into a cohesive script. AI has advanced a great deal in just the last couple of years. I’m excited to see what it will look like in another five years.

38. I had symptoms for a decade of something going on and every Dr I went to either dismissed me as “getting older” I’m 35. And having had kids. Or they ran all the tests they could think of and everything came back normal. Everyone was perplexed. I read another story about AI and decided to do the same. Gave it my family history, current test results and symptoms and it gave me a list of possibilities. Took those to my Dr and got my diagnosis. AI is scary as hell, but it’s so incredible at the same time.

39. Note point AI at cancer and Parkinson’s and let it prove its worth.

40. I’ve been using AI for the last two years in this capacity. I’ve uploaded bloodwork and now it’s bio hacked my entire body and provides in depth, real time meals, exercise, vitamins and sleep recommendations. I’ve also hooked in my health app for better and deeper knowledge and decision making. I upload photos of meals I eat to contain portion and calories as well as photos of my body for progress and real feedback.
    There is nothing else that could do this. I barely use it for “simple tasks” it has now become my personal consultant for most things.

41. Google the following words and there is a video showing how to temporarily relieve the ringing. I’ve been using it for a couple of years and the more I use it there’s less ringing.

    “This simple trick might help with tinnitus TruDenta”

    There’s a video showing a man placing his palms over the ears with the fingers facing to the back of the head. Then put your fingers on the back of the head, placing the index finger on the middle finger. Then thump the index finger on the head. The video says 40 – 50. Each time I do this it seems like the noise gets quieter.
    At least it doesn’t cost anything and I can do this as many times as I need.

42. I have had this issue for a long time. My dentist knew what to tell me to do immediately when the disc slipped out of place at a dental visit as a kid.

    AI didn’t solve anything.

Comments are closed.