Do not talk to Dr. ChatGPT. They're a quack. Seriously.

Welcome to Wondercafe2!

A community where we discuss, share, and have some fun together. Join today and become a part of it!

Mendalla

Happy headbanging ape!!
Pronouns
He/Him/His
Some researchers in Montreal tested how ChatGPT did on medical and health questions. It fared very, very badly. They asked for references and the danged thing actually made them up. ChatGPT can be hazardous to your health if you use it for health questions.


And, speaking from personal experience, even when using a conventional search engine, stick to hits from sites like Mayo, Harvard, WebMD, and Healthline. Even a regular search often gets dubious data where medicine and health are concerned.
 
Given that ChatGPT essentially crawls the web for its information and given the number of web-sources for questionable medical "information" that is not terribly surprising. I wonder if any studies have been done in other fields to test the accuracy of the information?
 
Given that ChatGPT essentially crawls the web for its information and given the number of web-sources for questionable medical "information" that is not terribly surprising. I wonder if any studies have been done in other fields to test the accuracy of the information?
I know there's been some recent general analyses showing that large-language model AI in general is actually getting worse. Apparently, it might be a feedback loop because a significant amount of the data they are drawing on is already AI-generated.
 
I know there's been some recent general analyses showing that large-language model AI in general is actually getting worse. Apparently, it might be a feedback loop because a significant amount of the data they are drawing on is already AI-generated.

Thus we could be directed towards the vector that AI is somewhat loopy and suffering Lupus (a medical malfunction causing ref faces)!

In addition there are legitimate sources like Johns-Hopkins, etc. as associated with reputable medical research schools. Note some care needs to be taken here too if you have ever read up on CIA corruption of psychological and neurological schools. Professionals are prone to corruption. Consider lawyers, and others in Black Robes for instance!

Some really weird science may be encountered while on the fly ... as a curios set ... ask questions!
 
I know there's been some recent general analyses showing that large-language model AI in general is actually getting worse. Apparently, it might be a feedback loop because a significant amount of the data they are drawing on is already AI-generated.
GIGO?
 
Agencylife Bingo GIF by MX Player
 
Some researchers in Montreal tested how ChatGPT did on medical and health questions. It fared very, very badly. They asked for references and the danged thing actually made them up. ChatGPT can be hazardous to your health if you use it for health questions.


And, speaking from personal experience, even when using a conventional search engine, stick to hits from sites like Mayo, Harvard, WebMD, and Healthline. Even a regular search often gets dubious data where medicine and health are concerned.
I'm amazed that this tech is being adopted by businesses

I have read that peeps have already lost jobs to this

I have also read a paper that predicts, due to the speed with which this tech can copy our works, eventually it will run out of our input and will rely on themselves..which will then result in more and more GIGO in a reflexive loop...
 
Given that ChatGPT essentially crawls the web for its information and given the number of web-sources for questionable medical "information" that is not terribly surprising. I wonder if any studies have been done in other fields to test the accuracy of the information?
Well I do know that various people, including prominent journalists, have tested it and seen that it can give false info
It "lies"
And can even try to blackmail
 
1000003960.jpg

Here is a nice flowchart I came across on a Statistician's post where he put chatgpt through the wringer with very abstruse stats queries
 
I also came across and article that said within a few months chatgpt became worse at coding....

which is one amazing thing i have seen people use it for
 
View attachment 8304

Here is a nice flowchart I came across on a Statistician's post where he put chatgpt through the wringer with very abstruse stats queries

It is said by some that this simulates a logical flow chart and you know how the fixated will go with that concept given their displeasure with logos ... and all that is obscure in word as they do not affiliate well with understanding and logos insight ... something needing looked intu ... like how word gets to be the way it is from some dark form long ago? That past ... was soon dropped and forgotten ... explaining the descending sensation some of us gather ... a sophisticated empath ... touché? It could be golden there in the dark ... with 'L OCHs ... ineffable ... regarding the Dan's in the dark theme ... another code!

How familiar are folk with it? They don't begin to know ... primal tomes say that knowledge is evil ... thus subdued in the pages and Oster's ... mythical vessels? Some defeated and disposed to be ejected ... thus logos flies in your face ... unseen ... yet a long story! It goes on ... continues!
 
There’s an “AI Therapist” that reacts music videos of a few musicians I like. It’s just for fun. The graphics are good - Freudian vibes - and it was programmed to be funny. I think these reaction videos are probably a human and AI collab (meaning I don’t think it’s generating them by itself but it’s hard to tell). There’s a disclaimer that it’s not a real therapist (I think - something like that is somewhere on the page). It also responds to some comments and is often hilariously witty with them. The laughs alone are good for mental health. I also wouldn’t mind if AIs were something like court jesters. But humans have to make them more reliable for us if they’re to give serious advice on anything or else we’re in trouble.
 
Last edited:
There’s an “AI Therapist” that reacts music videos of a few musicians I like. It’s just for fun. The graphics are good - Freudian vibes - and it was programmed to be funny. I think these reaction videos are probably a human and AI collab (meaning I don’t think it’s generating them by itself but it’s hard to tell). There’s a disclaimer that it’s not a real therapist (I think - something like that is somewhere on the page). It also responds to some comments and is often hilariously witty with them. The laughs alone are good for mental health. I also wouldn’t mind if AIs were something like court jesters. But humans have to make them more reliable for us if they’re to give serious advice on anything or else we’re in trouble.

If you look at life in a package, bag, box or carpet saac ... from outside the containment ... it is a sad representative of utopian vision! This happens with hated fringe folk as borderline personality disorders that appear chaotic ... all stuff of the psyche said to be mythical and thus beyond us yet! Leads to the question from the naivete: "are we there yet?" and the answer in undeterred ... negative giving a sense of down to understanding ... bu only if you look up at the fireworks required to get you out of what's up there ... DOX? (the latter defined as sarcastic documents on humans and their like)
 
@Kimmio Laughterlove might we describe these wayward things as lymph nodes? They serve complex and poorly understood functions like the pituitary axis ... and sometimes sources of pion particles ... objects of medical Mystery Disease related to forest origins. Here in NB the forest reports promised 7 years ago have been unfulfilled due to abuse of the dark mysterious place! RV'ng goes on there ... and that is sufficient cover for considerable misinformation! Naivete thus flourishes ... touché of the unknowing!
 
Back
Top