Hot Take Gemini AI feature: You can Augment your Family Doctor, save yourself

[Victor M]

The traditional Google Search allows you to find out more about the disease that you are diagnosed with or guess that you have.

Gemini AI can do it in reverse: you give it your symptoms, in your own terminology, and she will find a list of diseases that exhibit those symptoms.
It is akin to a diagnosis by a family doctor. However it is not a proper diagnosis because the doctor can order tests done to verify . And Gemini cannot. And your family doctor has access to your personal health history data. And Gemini does not. This usage of AI - to diagnose patients, is undergoing development.

If your family doctor decides a diagnosis relying on his experience without doing tests to verify all possibilites, there could be a chance of misdiagnosis.

Know also a family doctor is classified as a general practitioner and only has limited knowledge.

So this feature of Gemini AI allows you to give your family doctor a list of things that tells him of what to consider and test for. You are helping him. You can tell him that this list is obtained from AI. That it is drawn from Gemini's wide body of knowledge which includes medical textbooks and journals. And also tell your family doctor that AI is not just some program code but was designed to model the human brain with 20 yrs of research. Tell him these AI facts so that he will have some trust in what you are giving him. Of course he should also recognize that your list matches his training, at least in part.

You should be doing this to protect yourself, to increase your chances of a correct diagnosis.

Advances in technology is giving you an additional tool to safe guard yourself. Doctors rarely get challenged on their decisions. And Doctors are human too, they make mistakes. Some people choose to spend the time to go consult multiple doctors. That accomplishes the same thing. But now you can sit at home and ask Gemini , make a list, and save yourself the time.

Misdiagnosis can have very serious consequences. Something very dangerous may be overlooked and you can die. So just do this simple thing and minimize the risk. New AI technology now enables you to do so.

 

[cartaphilus]

According to the AI; as a midlife male I am currently suffering from menopause and an ectopic pregnancy.

Remedy: The AI recommends that I add glue to my pizza and gasoline to my pasta.

The future is soo bright I have to wear shades.

 

[CyberDevil]

I've been consulting with GPT for a while about vitamins, diet, medications, and possible symptoms, so I don't find anything surprising here. Human doctors often say stupid things and misdiagnose. It used to require a second doctor for verification, but now you can use AI.

For example, recently my family doctor told me that chicken and beef liver are not sources of iron, only red meat is. Yes, this person is a physician and needs to re-certify every couple of years.

 

[Victor M]

as a midlife male I am currently suffering from menopause and an ectopic pregnancy.

Well, to avoid that ridiculous possibility you would tell Gemini that you are male. Whether the AI can understand the concept of male-ness is questionable and you cannot correct her because you don't know the correct answer for the list of diseases. So anyways, give the list to the doctor, he will filter out the ones that don't apply. The entire list will not be totally wrong. This action does provide an extra safeguard to protect you.

Remedy: The AI recommends that I add glue to my pizza and gasoline to my pasta.

I don't think an AI will be that stupid to offer such a solution if you confine her to recommend medication.

I have more faith in the future development. This fault is known. Of course many researchers from both the academic world and commercial world will work to address it. And don't ignore the efforts by governments and their regulatory powers.

 

[blackice]

Really glad there’s enough power used to power my home for a day every time one of these AI gives dumbass answers.

 

[Victor M]

Really glad there’s enough power used to power my home for a day every time one of these AI gives dumbass answers.

Gemini, ChatGPT and Meta AI do not use an extra kwh of power than your normal computer usage. They use computer servers on the Cloud to do their calculations. They are just sending you the answer. Maybe with the new AI chips the will come built into MS branded laptops they will. I don't think that is available for sell yet. But you can continue to use the web sites.

I can give you a real example of how AI can save you money today. I asked Gemini to suggest me proven nutritional supplements that cost below $50 for 90 days. Gemini found scientifically backed products with plenty of tests over 20 years that cost around $20 per bottle. My first search on Google Web Search only found iffy products that costs more than $130. Google.com search is not capable of doing complex searches requiring an understanding of what scientifically backed means.

 

[blackice]

Gemini, ChatGPT and Meta AI do not use an extra kwh of power that your normal computer usage would use. They use Cloud Servers to do their calculations. They are just sending you the answer. Maybe with the new AI chips the will come bundled with MS branded laptops they will. But you can continue to use the web sites.

I can give you a real example of how AI can save you money today. I asked Gemini to suggest me nutritional supplements that cost below $50 for 90 days. Gemini found scientifically backed products that cost around $20 per bottle. My first search on Google Web Search only found products that costs more than $130 for 90 days.

This is misleading. Those cloud computing centers have seen an incredible increase in power consumption due to AI. The GPUs they use are power hungry. I was being sarcastic, but the AI garbage takes a lot more power than the google search you do yourself.

AI already uses as much energy as a small country. It’s only the beginning.

The energy needed to support data storage is expected to double by 2026. You can do something to stop it.

www.vox.com

 

[Victor M]

They do not use your GPU. They use the computers hardware in the cloud. A web page uses your browser only. It cannot instruct the browser to use GPU to do it's heavy calculations. The browser is designed only to display web pages.

I don't think it is misleading. My post addresses the fact that it does not affect you, your power usage cost. You have to take it as it reads literally.

You bring up an important fact. It does have a very high cost. But it brings important benefits that justify the cost. You have to spend something to get something.

 

[cartaphilus]

Yeah I can see printing the list and giving it to my doctor.

You know why doctors are the worst patients?? Because of crap like that. They think they know better. They want to tell the doctor how to do their job.

How would you feel if the client who asked you to pent test their network started to wrongfully analyze the results of your pent test while telling you that you must be mistaken because x y or z says so. What's worse. Now the patient will be non compliant with the treatment provided by the doctor because the patient will trust doctor Google.

Crap like that will just make Karen's even worse. I don't envy the future doctors.


Edit: fyi this post is not targeting anyone. It's just my observation. My ex was a physician.

 

[blackice]

They do not use your GPU. They use the computers hardware in the cloud. A web page uses your browser only. It cannot instruct the browser to use GPU to do it's heavy calculations. The browser is designed only to display web pages.

I’m not talking about my power usage. I’m saying every time you use AI their servers are using multiples more power for a garbage answer that I could have researched better myself in 30 seconds.

Their servers use massive power hungry GPU farms for AI.

 

[cartaphilus]

I’m not talking about my power usage. I’m saying every time you use AI their servers are using multiples more power for a garbage answer that I could have researched better myself in 30 seconds.

Their servers use massive power hungry GPU farms for AI.

Yeap and more farms will be built even though they might be 20% utilized since if you don't have the horsepower behind the engine then you will be left behind and other AI engines will dominate.

 

[blackice]

I believe that depends on your expertise vs what you are researching. When you are unfamiliar with a subject matter, your research time may not be 30 seconds. And AI can cut down that research time dramatically. You characterize AI answers are always garbage, they are not. They do have built in randomness, they do have occasional hallucinations, they do not do a perfect search. But they do shorten your research time depending on how stringent your research requirements are.

I will give AI credit the day I find it useful. This far it has not helped me in the slightest. I like researching and learning. I’d rather become an expert than have an AI give me half baked answers. I will give it credit the day it gives me a useful answer, which I’m still waiting for. I don’t do image work or content generation so none of that is useful for me. And it’s very problematic for the fields it’s affecting. So I’m pretty indifferent to AI and the massive amounts of resources being poured into them.

 

[Victor M]

will be non compliant with the treatment provided by the doctor because the patient will trust doctor Google.

It may very well be that ordinary people will trust Google more than their family doctor. But I explained why they shouldn't, that Gemini does not have access to patient data. That Gemini cannot order medical tests and perform medical tests. These 2 key things are components of making a medical diagnosis.

Gemini is very careful about these ethical matters. She appends warnings to every medical question I ask of her.

If people don't read, what can you do?

For the pen test scenario. So IF the client finds contradictory info from an AI. The client can refuse to pay the security professional. But will he be able to conduct a pen test himself ? No. He has to find another one. This AI technology will evolve the way we do things. The security professional will be required in the future to double check his pen test design and approach against the AI before approaching the client. The family doctor may choose to double check himself with an AI. The availablity of this AI does bring about improvement. It just changes the way we do things in unexpected ways.

 

[Victor M]

every time you use AI their servers are using multiples more power for a garbage answer that I could have researched better myself in 30 seconds.

I believe that depends on your expertise vs what you are researching. When you are unfamiliar with a subject matter, your research time may not be 30 seconds. And AI can cut down that research time dramatically. You characterize AI answers are always garbage, they are not. They do have built in randomness, they do have occasional hallucinations, they do not do a perfect search. But they do shorten your research time depending on how stringent your research requirements are.

 

[Victor M]

I neglected to tell everyone this: Gemini does not show all answers to your question. You have to ask the same question 2 or more times to receive a more complete list. This is due to her programming has a design goal to provide randomness, so that she does not appear to be producing canned answers. Meta AI has the same problem. So ask same question several times to produce a more complete list of diseases that exhibit those symptoms.