Hacker News Reader

The only reason this worked is because your situation was exceedingly simple.

The trouble is you are not educated enough to tell what is simple and what isn't. A cough could be a cough or it could be something more serious, only a "real" examination will reveal that. And sometimes even that's not enough, you need an examination by a specialists.

I'll tell you a story. Once upon a time I got pain in my balls. I went to a doctor and he felt around and he said he didn't feel anything. I went to another doctor and he felt something, but he had no idea what it was. He said could be a cyst, could be a swollen vein, could be an infection - he didn't even know if it was on the testicle or on the tube thingy.

Then I went to a Urologist. You can tell this man has felt up a lot of balls. He felt me up and said, "yup, that's a tumor" almost immediately. He was right, of course, and he ended up being the one to remove it too. Since I caught the cancer pretty early the chemotherapy wasn't too intense.

Point is, expertise matters when things aren't straight forward. Then, experience and perspective gets to shine.

I don't think this is the slam-dunk you think it is. You had to go to three doctors before someone recognized something was wrong. ChatGPT is the substitute for the first two doctors, not the third.

And even there, I bet ChatGPT would have told you to go see a doctor, since it can't feel your balls. And after your first appointment, if you had told it that you still thought something was wrong, it would probably have told you to go see a urologist.

I fall slightly on the “doomer” side of the safety spectrum, but if I’m being honest I’ll be first in line - rain, sleet, or snow - to buy the chatgpt that can!

>And even there, I bet ChatGPT would have told you to go see a doctor

So what is the value add of ChatGPT in this scenario? Either it says you don't need to see the doctor which could very well kill you or it tells you to visit the doctor and it doesn't actually help you in any way.

The value-add would be getting my insurance to approve me to go see a "specialist" (the urologist) immediately rather than forcing me to schedule an appointment with my primary-care doctor first, paying $200 for the office visit, and delaying care by 2 weeks.

So, a technical solution to a bureaucratic problem.

Also, even when I'm on a PPO plan instead of HMO...I know when I have a yeast infection. I don't want to pay $100 to tell a doctor that I have a yeast infection. They don't even look at my genitals - they just take my word for it. But I can't buy fluconazole without a prescription. Such a waste of everyone's time and my money.

So again, a technical solution to a bureaucratic problem.

Get a PPO plan, then? I can go directly to any specialist I want, although I find it’s better to always talk to my PCP first, who then makes a referral and often recommends the specialist she prefers.

(If your HMO charges $200 for a PCP visit - wow, I’d look for a different plan.)

My second example was specifically called out as an example of an issue even with PPO's. Also, not everyone (including myself) can easily change jobs or get insurance that simply isn't even offered by the employer. Worth noting that in my state for example, none of the public ACA health insurance options are a PPO.

Yeah, it's true that a lot of ACA plans are quite poor right now. The subsidy often isn't really worth it compared to a non-ACA plan.

That isn’t value created by the AI, it is valued bestowed on it by your insurance company. They could bestow that same value on a Google search or a call center worker in Southeast Asia and it would provide you the same value.

The AI is just making an already convoluted system more convoluted.

You're missing the case of it telling you not to see a doctor and it's right, in which you save a lot of time (and potentially money).

Mistakes will happen, that's the nature of statistics of this nature. But don't forget that saving 1,000s of people time is worth something, even if you never directly see it like you do the one case that it didn't catch.

> But don't forget that saving 1,000s of people time is worth something

How many unnecessary deaths does that buy you?

I don't know. But pretending there are no tradeoffs is stupid, and also disconnected from reality.

>I don't know.

Which invalidates the whole point. There are countless ways that the healthcare industry could save time if it was willing to sacrifice an unknown number of lives.

How is this scenario any different from some other arbitrary decision made for the sake of expediency such as doctors deciding to no longer collect medical histories from new patients? Think of all the time that could be saved if you never had to fill out those forms the first time you saw a doctor. Who cares if that leads to more people dying, it saves some time!

We’re discussing this on a thread about an attempt to answer that question!

When I say "I don't know", it means I genuinely don't know. Obviously, since this is a new technology, it's possible that it is better than doing arbitrary things like saying that people shouldn't fill out forms anymore.

You're saying "some change is bad therefore all changes are bad", which is obviously fallacious.

No, I'm saying change that even its supporters concede will kill people is bad if we can't actually quantify either its benefits or its cost in human life. You talked about a "tradeoff" while having no idea what is being traded for what. That is advocating for change just for the sake of change. Opposing that is not advocating against all change, it is saying "move fast and break things" is an immoral approach when dealing with human lives.

Your original comment argued there is no value add. I gave an example of the kind of value add that you can get (saving money/time on following up something wrong).

I did not advocate for this specific release being better, and I certainly did not say we should "move fast and break things". I don't know who you think you're arguing against but it's not me.

I only pushed back on the specific idea that this there is no possible value-add by this kind of software.

When you can't actually quantify the savings in time or the cost in life, you have no idea if this is adding value or not. I could just as easily flip a coin to decide whether to see a doctor, do you think that automatically "adds value" because the coin occasionally gives me the right answer regardless of how often it gives me the right answer?

Right, so as you've pointed out, chatgpt wouldn't have helped me in any way.

And, for the record, I was referred to a Urologist. The doctor was able to say "something isn't right but I don't know what" and he referred me to a Urologist and got my ultrasounds.

No offense to uh... AI... but it's not very good at being humble. It seems hubris wasn't in the training data. It's quite difficult to get an LLM to say "I don't know". Usually you have to prime it or ask leading questions, which at that point you already know that it doesn't know so therefore trying to elicit that response is worthless.

> I bet ChatGPT would have told you to go see a doctor, since it can't feel your balls.

Not _yet_!

> ChatGPT is the substitute for the first two doctors, not the third.

The third doctor used to behave more like the first two. In a world where ChatGPT replaced the rookies, how do we get good doctors?

As a competing anecdote, last summer my toddler woke up from a nap with bruises on his legs and swelling around his joints. We initially thought maybe he was tangled up in his crib, but later when changing his diaper found his testicles were swollen and rushed him to the emergency room. Over the next 6-7 days we saw no less that 5 doctors including his pediatrician, orthopedic specialists, and doctors at the ER and urgent care. None of whom were able to give us any answers about this weird bruising and swelling that randomly appeared on his legs. After all of this, during one of his flair ups, I took a picture of his legs and searched it with Google lens. The results set included pictures and symptoms of Henoch-Schonlein purpura [0], that were identical to what he was experiencing. We confirmed this with his pediatrician and decided on a treatment plan, but I was floored at how many doctors we had to see (and how much money we had to spend), only for me to diagnose this on my own with Google lens.

[0] https://www.mayoclinic.org/diseases-conditions/henoch-schonl...

> Why do I have to pay $93 (on top of the cough syrup that costed ~$44) just so that the doctor can see me on Zoom for less than 5 mins and submit an order for the med?

Because you're paying for the expertise of someone who studied for more than a decade which you won't get from a random web search.

An AI system with today's technology should be less trustworthy for medical diagnosis than a web search. At least with a web search you might stumble upon a site with content from experts, assuming you trust yourself to be able to discern expert advice from bot-generated and spam content. Even if a doctor is doing the searching instead of me, I would pay them only for their knowledge to make that discernment for me. Why you think an AI could do better than a human at that is beyond me.

Your question reminds me of that famous Henry Ford GE invoice story:

> Making chalk mark on generator: $1.

> Knowing where to make mark: $9,999.

AI systems have been improving. O3 now has the capability to decide to search multiple times as part of its response.

Even o4-mini (free) uses web searches and runs Python scripts very eagerly. I'm not sure how long they'll be able to afford giving all of that away.

You answered why someone would want to pay $93, but not why they have to pay $93.

There's a limit though right? How about, if you can accurately predict the doctor's diagnosis it's free? If not you pay. This person needed a doctor's approval presumably for a prescription they couldn't get without authorization which leads to this gatekeeping. Not to mention also contributes to the insane medical costs in the US. $93 for 5 minutes is over 1k _an hour_ for what amounts to a rubber stamp (because how much can you really dig into a person's medical history and condition in 5 mins).

I mean, it really depends on what you're trying to diagnose/do. Yeah, a math PhD might have spent a decade+ studying the subject. Doesn't mean I need to pay for their skills instead of using a calculator/program to solve my math problem.

> Why you think an AI could do better than a human at that is beyond me.

Why do you think an AI couldn't do better than a human, when we have ample evidence of computers/AI exceeding humans in many areas?

> Why do you think an AI couldn't do better than a human, when we have ample evidence of computers/AI exceeding humans in many areas?

I was specifically referring to the ability of discerning between accurate content and nonsense. SOTA LLMs today produce nonsensical output themselves, partly due to their training data being from poor quality sources. Cleaning up and validating training data for accuracy is an unsolved and perhaps unsolvable problem. We can't expect AI to do this for us, since this requires judgment from expert humans. And for specific applications such as healthcare, accuracy is not something you can ignore by placing a disclaimer.

Many human 'experts' produce nonsensical data. Verification of data by humans is also mostly based on 'prior' data. We've had many popular medical practices over the years developed and supported by medical experts which turned out to be completely wrong.

The main thing missing right now, imo, is the ability for LLMs to verify data via experimentation, but this is completely solvable.

An AI system with today's technology should be less trustworthy for medical diagnosis than a web search.

This is the problem with reasoning from first principles. This statement is easily proven false by giving it a try, whether it "should" be true or not.

I dislike that quote. It's not knowing where to make the mark. It's that there's no one else who knows where to make the mark charging less that $9,999. Knowing where to make the mark implies that payment is proportional to ability/productivity, but it is proportional chiefly to relative positions of power. Learning where to make the mark won't earn you $9,999, it will instead put you in competition with the other $9,999 markers in a downward race to the "true value" of those marks.

> Why you think an AI could do better than a human at that is beyond me.

You strike me as someone who thinks like me before I started using LLMs. Now when I read things like this rather than thinking in terms of black and white like I might have before, I instead think about percent likelihoods, and once you let in the grey it's no longer a matter of can/cannot but how many resources will be invested into making it so.

I think you hold doctors in pretty high esteem and I'm not sure how many other people hold doctors in as high esteem as you do. I've certainly ended up worse for having visited one. How seriously do you think most doctors take continuing education, particularly if they are overworked? An AI can consume the latest research/guidelines/text books immediately. Doctors can often act on 20 year old medical dogma. AI can get immediate social media feedback weighted in a way professionals deem appropriate.

I think diagnostic work is one of the things most likely to be taken over by AI. Even when asking specific technical questions in my own field, AI is capable of hitting some key points, and with extra prodding even getting to some very very specific niche points without too much coddling.

> You strike me as someone who thinks like me before I started using LLMs.

Cool, except I've been using LLMs since their inception, and they're now part of my daily workflow, for better or worse. My opinion is based purely on my experience as a user over the last few years.

Besides, my point in that paragraph was specifically about the ability to judge accuracy in information and to produce accurate information, not whether AI can outperform humans in other tasks. Modern AI systems fail miserably at this, which might be acceptable for casual use, but when human lives depend on that accuracy, such as in the healthcare industry, the process of integrating them should be gradual, with a lot of oversight and regulation. We've done that in the automotive industry, for obvious reasons, and the same should be done in others.

> I think you hold doctors in pretty high esteem and I'm not sure how many other people hold doctors in as high esteem as you do.

We hold the medical profession, not individual doctors, in high esteem for good reasons. It is critical to society, and requires intense education for many years. At the end of the day, doctors are as fallible as any human, but they're required to be licensed, and we have regulations in place for medical malpractice. The idea that a computer system based on probabilities and randomness, that hallucinates most of the time, will be more reliable than the average human doctor is ludicrous, regardless of what some benchmark says.

I'm not against using AI in healthcare. Pattern matching to detect early stages of cancer is a very good use case that's been helping radiologists for many years now. But that's the key word: _helping_. My concern is that when these systems are used for medical diagnosis and advice, they will instead be relied on much more than they should be.

>Why you think an AI could do better than a human at that is beyond me.

AI can take your entire medical history and the entire history of physics, chemistry, biology, and medicine up to May 2025...

A human was sleep deprived in Med school 15 years ago.

Where are you that you need a prescription to get cough medicine? The only ones I know of that require prescription are the ones with controlled substances.

> Of course, some of the medications that include opioids will only be sold to me with the doctor's prescription

Yes, and the commenter was not complaining about such cases!

Ain't nothing wrong with a little Wockhardt to get you through the day

Benzonatate is a commonly prescribed somewhat stronger cough medicine which requires a prescription. I've known a number of people who've had to take it for a persistent or stubborn cough.

Wikipedia says "Overdose of only a small number of capsules can be fatal" but the study they quote shows that you'd need to take 10x-20x the usual dose for the smallest amount that's been shown to kill anyone -- which is admittedly more narrow of a safety window than over-the-counter Tylenol/Acetaminophen/Paracetamol but still, few people take 10-20 pills in one day of something that says "take 1 per day".

There are simply not enough doctors for doctors to be doing this work.

We have a massive, massive shortage of doctors.

The industry is doing everything they can to make it worse by the day, so I won't hold my breath that we'll get the slightest bit of respite.

It'd obviously be ideal if everyone could see a doctor for an affordable price any time they wanted.

We don't live in the ideal world.

This would be a HUGE win for most people.

This is why most routine primary care can be delivered by Nurse Practitioners and Physician Assistants working under a doctor's supervision.

NPs and community paramedics are definitely the future for a huge chunk of the current GP workload, but PAs are a failed experiment which are being rolled back pretty much everywhere.

Are you sure about that? The ranks of PAs have been expanding rapidly. I know there has been some political resistance over expanding their scope of practice but I'm not aware of it being significantly rolled back.

https://www.bls.gov/ooh/healthcare/physician-assistants.htm

The US is the global outlier here and it's in the face of the evidence. The number of PAs globally is plummeting as they're abolished by national authorities.

Nah. Total PA numbers are increasing in many other developed countries as well, including Canada and much of Europe. We can have a discussion about whether this is a good idea, but first you need to stop lying and making things up.

https://capa-acam.ca/pa-profession/pa-facts

https://www.srh-university.de/de/folder/news/2025/04-25/erst...

https://www.bigregister.nl/over-het-big-register/cijfers/ver...

The only country in Europe where PAs made any significant inroads was the UK, where their most significant role has been abolished.

In Germany and all other countries, PAs are unable to treat patients without direct oversight and they are single digit percentages compared to NPs.

It's clear you have no experience in this area, so I wonder why the need to comment at all?

PAs have to practice under some level of physician oversight pretty much everywhere. That part was never in dispute. But they can offload physicians for some routine primary care cases where a real doctor isn't needed.

The data clearly shows that PA numbers continue increasing in many countries, so obviously they don't consider it a failed experiment and you're just lying to push some kind of personal agenda. It's clear you have no experience in this area, so I wonder why the need to comment at all?

We are the global outlier because it costs an arm and a leg to go to college and medical school. In the rest of the world the education is free/ still affordable. And thats why all the young doctors are also imports.

No, it's because the role played by PAs is performed by the much more capable and clinically proven NPs and community paramedics.

Yes and being able to buy and take whatever you want has led to rampant antibiotics abuse.

Unfortunately medicine at the high level always has to cater to the lowest common denominator.

Personally, while I agree AI can probably take up some of the mundane triage I have some reservations about AI in its current LLM form being a good judge as patients are unreliable narrators and doctors rely on many senses and heuristics to make their recommendations.

An AI fine tuned to be a doctor with perfect health data and patient history is an interesting prospect, especially given the loss of the "family doctor" but that kind of data collection is a way away and has other Pandora's box considerations

I mean, if such is your typical experience with healthcare, what you need is a big shake up, don't count on AI to help you there! (but it might inflate the profit margins for the few running the show at your expense).

You think you won’t have to pay the AI? lol.

IMHO it is much easier to increase the supply of AI doctor hours than the supply of human doctor hours, and supply constraints are the main thing maintaining doctor hours at the high price it currently sells for.

My $700 laptop can run an 8B model in 2023.