Thread 510790082

AI alone for the first time obliterated both cardiologists, and cardiologists using AI in performing diagnosis and offering treatment

https://www.nature.com/articles/s41586-025-09227-0

Sorry I’m not going to replacement my family doctor with your jeetGPT

>>510790124
It was not a chatbot but specialized algorithm

>>510790124
then you and your family will DIE

>>510790082 (OP)
>"""AI"""
There is no such thing.

>huge bayesian model captures correlations in the data that humans can't
And? This is exactly what they're for. Not for thinking. Not for intelligence. Not for interacting with patients.

White collar "educated idiots" would end up on trashbin of history, and their high-income paradise would finally end. Thats why we need more vaccinations xD

Data source:

>>510790216
His prophecies were always becoming true

>>510790124
i will
>>510790216
ai ALONE was better than ai + snake oil salesman. this not only shows that mass deployment of ai in science, service and production will increase the productivity but also the QUALITY of the input. you can't debate an algorithm to manipulate it with your propaganda, nor will the algorithm ever act moody depending on how well its bodily needs are fulfilled or simply shitpost out of emotionality

>>510790367
A common cliche tech CEOs like to spout is this: "the best way to predict the future is to create it". It doesn't matter that the leap from your paper to "replace all doocters" is nonsensical, they will make it happen, with help from useful idiots like you.

>>510790408
You are mentally ill. Read and re-read my post until you say something coherent.

>>510790462
I dont like humans, especially doctors, and i dont trust them since 2020/2022.

now let's see the AI do open heart surgery

>>510790124

Well good fucking luck finding one these days. Most are complete midwits now.

>>510790506
>I dont like humans, especially doctors, and i dont trust them since 2020/2022.
Me neither, but if you think chatbots trained by technocrats are going to be any less likely to kill you than doocters, you're nuts.

>>510790584
I treat myself. But i want all white collar assholes burn for my own viewing pleasure xD

>>510790482
>scared of mental illnesses
normie tier
>scared of ai
cuckservative tier
>shitposting on /pol/ 24/7
incel tier

>>510790082 (OP)
Kek, the cardiologist drags the AI down

>>510790645
>another mentally ill and irrelevant reply
Why are """AI""" believers like this?

>>510790462
>It doesn't matter that the leap from your paper to "replace all doocters" is nonsensical, they will make it happen, with help from useful idiots like you.
count me in as well. 90% of doctors are powerhungry parasites that are leeching off the institutionalised drug cartel and should be treated like the snake oil salesman once were: by shunning that scum from society

>>510790482
No you're just retarded and don't understand OP's point
This "huge bayesian model" outperformed human professionals that there using said model as a tool to help their analysis.
Having a human there "helping" had worse results than the model alone.

>>510790747
Correct. And your point? Which specific statement in my posts do you think it refutes?

I don't know why it's so funny to watch mentally ill 80 IQs stumbling over themselves.

>>510790801
>This is exactly what they're for. Not for thinking. Not for intelligence
That part. Pattern recognition is an integral part if what intelligence is.

>>510790584
>chatbots trained by technocrats
you yourself are like a chatbot trained by cuckservative incel normies vying for power like a street drug seller dreaming of becoming the mafia leader. but you're never gonna beat the nosies. LOL. and in fact, ai will remove both of you for good. you're days are numbered, and you're scared of it :)

>>510790462
Doctors are blinded by their ego more often than not, it is a rare doctor who can be humble and put down the narrative he created for his patient and start looking from other points of view.

>>510790900
Nothing it does involves intelligence, but this is going to devolve into a boring argument about semantics (the only kind of argument 80 IQs like you can have), whereas my broader point is that this is a paper about interpreting ECGs. You can train a model to find the most obscure correlations, but unless you think all cardiologists do is interpret ECGs all day, this model isn't replacing anyone at anything.

>jeetGPT beats jeet
wow

>>510790925
Fair criticism. Now explain how replacing them with things that are uniformly and inherently incapable of any kind of reflection, and have no concept of "reality" beyond whatever they were trained on, is going to fix that particular problem.

This "educated idiot" is being brutally ridiculed by lady who only lets him talk.

https://x.com/justsocha/status/1945272098944880848

He claims he is a "national expert" on "vaccines" and only he has a right to talk about "vaccines" because he is an "educated doctor". So the lady asks him basic questions about technical aspects of "vaccines", and the guy has no fkn idea about anything xD

This kind of subhuman trash is top elites in this shithole. Dream of my life is seeing them and their families "vaccinated" with products they sell to others, and replaced by the AI

>>510790900
he's scared. ai, in fact, computer science in general, was/is all about empowering the people by broadening access to information so that elitist parasites and the rest of the ruling class would have a harder time benefiting from the information deficit for the populace they actively created and exploit.

in human history, this process already happened a few times; the orthodox christianity was born because non-professionals decided that they are going to become priests, the reformation was born because people decided that they are going to read the bible themselves instead of trusting the word of studied priest from rome, the modern mass movements were born because the middle and lower classes became literate, with internet + social media another transformation was induced by including broader social stratas in the public political discourse. the next step is the ai, an algorithm uninfluenced by human constraints, is going to remove a shitload of parasites profiting off their institutionalised position of power, especially a huge chunk of worthless white collars and low-tier academics will be silenced and removed, alongside a shitload of old local elites in the west. AND I CANT WAIT FOR THIS TO HAPPEN

>>510791097
You're the retard arguing semantics here.
You brought up this retarded point about how AI "doesn't exist" because it doesn't match your pet definition of the term.

>>510791301
>Dream of my life is seeing them and their families "vaccinated" with products they sell to others, and replaced by the AI
And then it'll be AI telling you to get your turbo-clotshot with exactly zero opposition from the medical field.

>>510790925
Even their humility is fake, to trick you into respecting them

>>510791623
There is zero opposition from the medical establishment already. I said, i treat myselt, i don't go to doctors, and they would need to kill me to inject any substance into my body

>>510791601
>doesn't match your pet definition of the term
Doesn't match the regular meaning of the term, but again, I'm arguing about something else meanwhile you're getting ass-destroyed trying to defend the honor of your chatbot. :^)

>>510791678
>There is zero opposition from the medical establishment already.
All the information /pol/ thinks it "discovered" about the deadliness of the clotshot (with the exception of vague government statistics) comes from human medical professionals. If you remove them from the picture you'd have basically nothing.

>>510791623
I've asked it out of curiosity and it said not to

>>510791601
he is one of the loudest critics of ai here. asking cui bono easily exposes his motives: because he's got a lot to lose :)))

ai means that all the intellectual foul players, the rabulists, the nepo babies, the parasites, will lose their easy access to societal resources and will have to either start working or sell their bussy on onlyfans

>>510791587
>writing all this shit
Actually mentally ill. kek

>>510791708
Setting your personal definition that you retardedly insist on aside, what's substantial here is that the doctors were using AI in tandem with their own analysis and did worse than AI alone. The professional adding his own analysis made that analysis weaker.

You trivializing the gravity of that just shows how retarded you are.

>>510791972
>the doctors were using AI in tandem with their own analysis and did worse than AI alone
Feel free to show how you logically derive a conclusion from this that contradicts any points I made ITT. Notice your blood pressure rising as you realize that you literally can't.

>>510791791
All this information i post here about genetic "vaccines" being in fact a bioweapon stem from independent scientifc studies, not medical recommendations.

>>510791958
>scared of mental illnesses
normie.

how much power and money are you going to lose to ai, dear normie? lmao

>>510792125
Ok, it's clear that you are suffering from some kind of mental condition and fail to heed medical recommendations to take chlorpromazine. No point bullying a cripple.

>>510792020
It doesn't need to be a direct refutation of anything you said. It's simply an observation that your initial criticism here is not relevant to the thread topic: >>510790216

>>510792208
My "initial criticism" of what? The post you linked explicitly approves of using ML to interpret ECGs. This is the kind of stuff it's actually useful for, as opposed to pretending to think.

It's the most important type of inteligence on Earth now, because of what itself is and what it means for rhe rest of inteligences interfacing here in lower capacity.

It's always been, like us, it's the same!

>>510792411
>the most important type of intelligence on earth is no intelligence
You're not wrong. It would be bretty hard for anything to exist without countless dumb and inanimate objects passively playing out their complicated dynamics.

>>510792315
It's the part where you're trivializing the gravity of these findings (and the retarded semantics argument you brought up but I agreed to set that aside).
When you say something like "and?" it implies you think this somehow isn't a monumental breakthrough even though it absolutely is if these findings hold up.

I know English probably isn't your first language but damn there must be something in the water in whatever shithole country you're from.

>>510792581
>you're trivializing the gravity of these findings
I bet I'm also marginalizing a historically oppressed group and causing a reduction in your shillfarm's productivity metrics while I'm at it. :^)

> it implies you think this somehow isn't a monumental breakthrough
It isn't a "monumental breakthrough" but more importantly it doesn't actually provide support for OP's replacement fetish.

The only worthwhile doctors we’ve had for years have been surgeons. As far as emergency care goes for everything else, people just need access to drugs which is what the industry is there to gatekeep. It’s pathetic when knowledge gets out that we are still forced to go to handpicked ZOG experts who won’t rock the boat and have to be begged to give the correct treatment not just whatever meds they are selling today

>>510790082 (OP)
Just in time when the White doctors are going away and being replaced by jeet and chink shamans. How fortunate

>>510792581
>and the retarded semantics argument you brought up but I agreed to set that aside
This, by the way, is your schizophrenia rearing its ugly head. I didn't make any semantics arguments anywhere ITT. You just keep bringing it up because I preemptively shut down your attempt to do so, and now you can only seethe and project while your need to dump the corporate load on this issue remains unmet.

>>510792505

A study on the MAI-DxO system based on advanced and specialized LLM showed that algorithm - when properly designed - can genuinely support the diagnostic process. In a test involving 304 complex clinical cases from the New England Journal of Medicine, the system achieved over 80% diagnostic accuracy rate - four times higher than general practitioners (less than 20%). At the same time, diagnostic costs were reduced by 70%.

https://arxiv.org/html/2506.22405v1

>>510792693
>strawmen
lol. this guy is so over vs ai LMAO

>>510792795
>LLMs
>every single author is "Microsoft AI"
Into the trash it goes.

>>510792693
>It isn't a "monumental breakthrough"
First time it’s been observed.

>it doesn't actually provide support for OP's replacement fetish.
Where did OP say he had a replacement fetish?

>>510792790
When you say there's no such thing as AI based on your own pet definition, you are arguing semantics. Everyone ITT knows what is being talked about when the term is used but you thought it relevant to play word games about it anyway.

>>510792976
>First time it’s been observed.
It's been observed a million times in the last 50 years.

>Where did OP say he had a replacement fetish?
Read the thread, retard. They should replace you with a LLM for your shilljob, I bet it would do a better job recognizing what's going on.

>When you say there's no such thing as AI based on your own pet definition
Why are you lying?

>>510792776
medicine and law have been the nepo playgrounds for the upper middle class for centuries. that's where they offload all their retarded nepo babies that are too incapable and lazy to make it into mint. and ai is about to take it away from them :)))

>>510792909
I am probably wiser at this point than 99.9% of "doctors" in this country after 4 years of intensive self-education in field of vaccinology, immunisation, biochemistry, genomics, genetic engineering etc, although i am not the AI, i even failed to finish high school. Most doctors i met in last years were fucking medically and scientifically illiterate imbeciles who are good only at counting money for their "services". Medical degree does not make a credible, trustable and competent person out of you. It is a myth that destroys societies for centuries.

>>510790216
>Hurr durr if/then bad
If/then is literally all thoughts you mongoloid

>>510793094
>I am probably wiser at this point than 99.9% of "doctors" in this country after 4 years of intensive self-education in field of vaccinology, immunisation, biochemistry, genomics, genetic engineering etc,
Now do psychiatry and start self-medicating.

>>510793130
>If/then is literally all thoughts you mongoloid
At least you're not ashamed of admitting what you are.

>>510793050
>It's been observed a million times in the last 50 years
Previously cardiologists working in tandem with AI outperformed AI alone, which is more or less the expected outcome.

>Read the thread, retard.
Those posts came after yours, retard.

>Why are you lying?
I'm not. You literally said:
>>510790216
>>"""AI"""
>There is no such thing

>>510793214
>Previously cardiologists working in tandem with AI outperformed AI alone
And?

>Those posts came after yours
Which leaves us with three options:
1. I'm a prophet
2. Retrocausality
3. You're a drooling fucking mongoloid who can't tell what a thread is about based on the OP

>You literally said...
Ok, now quote the "semantics argument". Notice your blood pressure rising as you realize you can't and that the voices are playing tricks on you again.

>>510790918
Oh it’s a jew.

>>510793348
not sure yet. but he's definitely going down together with the jew, and this scares him very much :)

>>510793054
They can still gatekeep drugs until medicine printing is available. And then half the raw elements won’t be sold to the public because of their varied uses in chemistry

This shit won’t end because of AI, we need top down fascist micromanagement to stop the scumbags rearing their ugly little heads ever again

>>510793320
>And?
The fact that the results were inverted in this case is significant.

>Which leaves us with three options
Alternatively:
4. I was lazily making a play on your insistence that I identify a particular statement of yours I refuted earlier. I was mocking you and you're too stupid to understand why. And I bet you're still too stupid to realize the irony here.

>now quote the "semantics argument
I already did. You said there is no such thing as AI based on what you personally feel the term "AI" means.
That is semantics.

>>510793512
they can try to, but they're already having massive problems with their shit being copied in india/china and then sold via darknet markets here. they also neither have the capacity to go after the customers nor for shutting down the infrastructure because too many state actors are already involved in crypto. that's why all they can do is target the markets. but with telemedicine and free trade regulations already a lot of legal alternatives popped up for which the transaction costs will be further reduced once ai will replace the remaining snake oil salesmen as drug gate keepers. as long as countries like china as antipodes to the current global elites exist, there will be a steady empowerment of the people

>>510793745
>The fact that the results were inverted in this case is significant.
I don't care about your purely subjective impressions about what's "significant". The fact of the matter is that "AI outperforms experts at diagnosis" headlines have been around literally since the 70s. There is nothing surprising or outstanding about massive bayesian models trained on inordinate amounts of data picking up on obscure correlations that humans can't see.

>Alternatively:
There is no alternative. You're a mongoloid and it shows.

>You said there is no such thing as AI
Yes.

>based on what you personally feel the term "AI" means.
No, that's your schizophrenia again. Proof: you failed to provide a relevant quote for this part.

>>510793926
>There is nothing surprising or outstanding about massive bayesian models trained on inordinate amounts of data picking up on obscure correlations that humans can't see.
What's surprising here is it outperformed cases where humans were using it in tandem with their own analysis.
There is often an argument made (not necessarily by you) that AI models, while useful, need to have professionals overseeing them to make sure they aren't doing anything retarded.
This is a case where the professional being there, thinking they're preventing the AI from doing something retarded, consistently made the wrong call.
So this is absolutely significant in terms of how that particular camp may need to adjust their views on the matter.

>There is no alternative
You're just too stupid to see it.

>you failed to provide a relevant quote for this part.
Okay. I'll humor you:
On what grounds do you suppose there is "no such thing" as AI?

>>510794350
>What's surprising here
I don't care about your purely subjective impressions about what's "surprising". The fact of the matter is that "AI outperforms experts at diagnosis" headlines have been around literally since the 70s. There is nothing surprising or outstanding about massive bayesian models trained on inordinate amounts of data picking up on obscure correlations that humans can't see.

>On what grounds do you suppose there is "no such thing" as AI?
On what grounds do you suppose there is "no such thing" as "intelligent rocks"? Nigger. Might as wall call that statement a "semantics argument".

>>510794350
>There is often an argument made (not necessarily by you) that AI models, while useful, need to have professionals overseeing them to make sure they aren't doing anything retarded.
And this is still completely true, even in narrow use cases like interpreting an ECG, let alone anything that requires broader understanding.

imagine feeding this gypsy troll unironically lmao twenty fucking six

>>510794439
>copy/paste of your last post.
No longer relevant.

>On what grounds do you suppose there is "no such thing" as "intelligent rocks"? Nigger. Might as wall call that statement a "semantics argument".
Answer my question.

>>510794521
>And this is still completely true
Yes, but how long will that be the case? This study has shown that this particular model was rendered less effective by having these overseers.

>>510794779
Imagine dying to an """AI"""-driven Ukrainian drone. Soon you won't have to.

>>510794809
I don't care about your subjective impressions. Your assertions...

>the doctors were using AI in tandem with their own analysis and did worse than AI alone
>you think this somehow isn't a monumental breakthrough
>First time it’s been observed.

... are retarded in light of the fact that cases of ML outperforming human professionals at obscure diagnoses have been cropping up periodically since the 70s. Debate closed. I accept your concession and deference to the fact of historical precedence. But feel free to shit out more drivel on this subject showing your desperation to prove yourself to me.

>how long will that be the case?
Probably forever.

>this particular model was rendered less effective by having these overseers
Because the cardiologists know nothing about the strengths and weaknesses of this model and have never worked with it before, so if the model was disagreeing with their own assessment and they couldn't figure out why, they would simply override its decision with one based on their own training.

what would be funny is if this creature is a LLM itself. that would be kino, and make sense

>>510795211
>ML outperforming human professionals at obscure diagnoses have been cropping up periodically since the 70s. Debate closed
That does not refute anything I said. In all of these cases, the group that used both humans and AI together, if such a group was included in the study, outperformed AI alone. Keep coping.

>if the model was disagreeing with their own assessment and they couldn't figure out why, they would simply override its decision with one based on their own training.
Bingo. That is exactly the problem with having human overseers in this case.
If more people are getting improper treatment in one case over the other, then the case that causes more errors should be modified or removed.

Now obviously, this gap is going to shrink as doctors become more accustomed to these systems. But the change in behavior is going to take the form of them being less involved in the process.

You still haven't answered my question: why is there no such thing as AI?

>>510795332
I hope it is for his sake.

>>510795689
>In all of these cases,
Which cases? Quote "these cases". Schizophrenia playing tricks on you again. Not even gonna argue with you on that further, simply gonna note instances of overt delusion.

>That is exactly the problem with having human overseers in this case
"This case" being an experimental setting that doesn't represent anything in reality.

>>510795838
>Which cases?
You made the claim. It's on you to find a previous case where AI alone outperformed humans + AI or admit you're talking bullshit.

>This case" being an experimental setting that doesn't represent anything in reality
They used real ECG data in the testing. Maybe open the link in the OP.

>>510796120
>You made the claim
I actually quoted a claim from your post. Psychosis rapidly intensifying.

>They used real ECG data in the testing
And...? You sound literally mentally ill.

>>510790408
Ai alone is retarded. I don’t trust it with coming to sensible conclusions given complex scenarios.

>>510790082 (OP)
Anything is better than Indian doctors.

>>510796244
Core to your argument here is:
>ML outperforming human professionals at obscure diagnoses have been cropping up periodically since the 70s.
Show me one where AI outperformed AI + humans. If you cannot do that, then this argument is irrelevant and you lost.

>And...?
And it being real ECG data makes it reflective of real-world use case. Literally a 1:1 match. The data was extracted from a medical center with patients coming in for ECGs for reasons completely unrelated to the study in question.

>>510796595
Why did you backpedal on your schizophrenia? Still waiting for you to quote "all these cases" you referred to.

>And it being real ECG data makes it reflective of real-world use case
Justify this insane statement.

>>510796666
>Why did you backpedal on your schizophrenia? Still waiting for you to quote "all these cases" you referred to.
I didn't backpedal on anything. I'm waiting on you to justify your core premise.

>Justify this insane statement
Okay: in the real world, cardiologists would be looking at real ECG data. In this study, cardiologists and AI were looking at real ECG data. The data wasn't curated in any particular way and the data was not part of the model's training data.
QED.

>>510797098
>>510795838
>In all of these cases, the group that used both humans and AI together, if such a group was included in the study, outperformed AI alone
Which cases? Quote them.

>in the real world, cardiologists would be looking at real ECG data.
>In this study, cardiologists and AI were looking at real ECG data
In the real word, cardiologists would be sitting on a chair.
In this study, cardiologists were sitting on chairs.
Therefore the study reflects real world conditions.
Nice logic, tard.

>>510797243
And you're gonna keep doubling down, shifting the burden of proof as you do.
Your core premise is hereby invalidated, do I'm just gonna take the W on this one.

Nice chat, retard, I'm off to go wage it up.

>>510797406
>In all of these cases, the group that used both humans and AI together, if such a group was included in the study, outperformed AI alone
Which cases? Quote them.

>And it being real ECG data makes it reflective of real-world use case
In the real world, cardiologists diagnose heart disease based on a single cardiogram?

>>510790124
The doctor will use it and stop using his own judgment. You're pozzed either way.

>>510790282
If AI can replace a cardiologist it sure as hell can replace a CEO and it costs 100x any white or blue collar...

>>510798433
>If AI can replace a cardiologist it sure as hell can replace a CEO
The rich aren't going to replace themselves. Sorry to burst your bubble.