Doctors Told Him He Was Going to Die. Then A.I. Saved His Life | Hacker News Reader

Doctors Told Him He Was Going to Die. Then A.I. Saved His Life

nytimes.com

33 pointsasnyder14 comments

Back when I used to visit doctors that was their biggest concern they brought up with me knowing I was in tech. My opinion for them was that AI was not the threat but rather not staying current in medicine, biology, chemistry is what would get them. It's one of their first lessons they are taught in medical school, that once they graduate more than half of what they learned will be no longer applicable or incorrect. I've only met a few doctors that took that lesson to heart and did not incorrectly interpret it as a disclaimer. There is also a decent doctor on Youtube that embraced this concept Dr. Roger Seheult, MD on MedCram. He was way ahead of and on top of Covid long before any of the other doctors as can be seen in their channels history. Such doctors are quiet rare.

Tell me about it.

Due to some health issues, I've connected with physicians/chemists/researchers. Reading case studies and medical research became a bit of a hobby. As time passed by I learned more medical terms and about health/medicine in general.

I've met many physicians for various issues 40+, and been to Mayo and other top clinics. These were often doctors interested in their niche and would reference information from very recent papers.

On the other hand, most physicians are quite bad, and appear uninterested in furthering their knowledge. Either through burn out or laziness? I don't know. I have met doctors who are clearly WRONG, and they were wrong because they accepted a book from 1998(example) as truth, they would be right years ago.

Medicine is an ever evolving science. I don't want to be treated like I am living in 1998. I don't want to be dismissed because I look "normal".

The amount of pushing to get tests done I've had to do is absolutely crazy. Maybe this warrants writing a book.

I even had a sports injury many years ago, that doctors referred to as "chronic pain", because they couldn't see anything wrong on tests and it lasted months. I've researched it, I've found a radiologist who created a specific protocol on how to set up an MRI machine to pick up the injury. I went to an MRI place with a protocol, gave it to them, they took it but ignored it and performed a regular MRI. No injury picked up. I then met a surgeon 2000 miles away who was passionate about medicine and operated on people with similar presentation. He inspected my abdomen and I was on a surgery table 3 hours later. I had images of massive muscle tears I've had in my abdomen.

If I had accepted opinion of like 6 doctors I saw, I would be living in pain today.

If I had accepted opinion of like 6 doctors I saw, I would be living in pain today.

I ran into a similar issue with 4 ulcers. A doctor rubbed my belly and said it was a virus. I already knew what it was. I wanted a camera and optimally a laser to clean things up so I could finish repairing them with a few protease, nutrients, mastic gum in a series of manual procedures but the wounds were uneven and not healing correctly. I went to 5 different facilities including one that misrepresented themselves as GI doctors and then turned out to be for gunshot wounds and still happily took my money and then acted like I was the idiot for going there. A GI doctor was only able to find 2 of them and said they were "unremarkable" which was a load of crap. In hind sight my mistake was taking a time slot on a late Friday afternoon.

I ended up just taking the time for the cellular turnover rate of the gut in combination with my hackish techniques to heal them. I was able to heal 3 out of 4 of them. One is in a tricky spot and is only about 85% healed. The last one only hurts if I use digestive enzymes. Similar to you I would still be in great pain today if I had not performed my own procedures.

It wasn't AI. The doctor who actually found the cure is basically a House M.D. of rare diseases...His A.I. just searches through known treatments and suggests treatments from similar conditions.

Or in other words, what any medical expert in the field should have done...

https://www.aamc.org/news/doctor-saved-his-own-life-now-he-s...

https://archive.today/55keI

It's very frustrating how the one detail, what AI models these docs are using, is unmentioned, yet every other detail about patients, medications, diseases, and rare disorders are. I understand they're under development. But why can't you tell me what they're calling them??

Because AI cannot invent cures or therapies this just means that AI searched medical records and/or documents and likely found a treatment a real doctor innovated. This is more analagous to a search engine than "True AI" saving hsi life. This is like using Google.

Doctors don't make treatments out of thin air but by testing drugs/procedures and seeing what works. Besides that models can be used to create new drugs, utilizing a drug's side effects to treat diseases that weren't the drug's main focus is quite similar to what doctors do. As for the "true AI" part, AI is a fancy term meaning multiple things nowadays. There's no such thing as "true AI".

> Doctors don't make treatments out of thin air but by testing drugs/procedures and seeing what works.

Its not doctors who do testing.

I was several times in situation, when doctor with early ChatGPT confidence prescribed some treatment, and I googled actual NIH study at home and seen very different conclusions to what doctor told me.

Doctors can confidently hallucinate whatever they want without being held accountable.

Doctors (in the US) tend to pay a whole lot for insurance because they can be sued for huge amounts of money when they screw up. Being confidently wrong and deviating from the standard of care by dismissing things or mistreating them is exactly the sort of thing that loses them malpractice cases.

Of course an after the fact payout may not be able to make you whole.

> Being confidently wrong and deviating from the standard of care by dismissing things or mistreating them is exactly the sort of thing that loses them malpractice cases.

its theoretical assumption, but looking at my personal experience, it doesn't look like doctors feel much pressure to do some deep research, and they just give some semi-generic advises which could contradict actual studies.

The standard of care is more conservative than doing whatever you can find a study saying might help, and being a bit behind of the latest research is very different from "hallucinate whatever they want without being held accountable".

> The standard of care is more conservative than doing whatever you can find a study saying might help

or may be not aggressive enough to prevent deadly decease. The point is that you can't know if doctor did some real high quality research, or just made some superficial decision, and there is no clear way to held him accountable for being lazy or uneducated.

nyt should be ashamed. AI did not save this poor fellas life. If anything it shows how stupid his doctors were to tell him he was going to die when it was not true. This is hype and they must be getting paid to write this, right??? It does tactically say "search", but then at the same time says "not thought of by a doctor" which is pure BS.

For real the enhanced autocorrect just cured a rare blood disorder wow! I bet father clippy is proud!!! The idea that the AI is connecting dots that previously did not exist is pure manipulation.

Reply on news.ycombinator.com