Context is key: Generative AI may not always be accurate, but that's because humans aren't always accurate.
To wit, I've received some patently incorrect information about diabetes from many healthcare professionals. I don't expect AI to be better, but it probably scales more efficiently.
The second generation of the app should have a feature where it alerts you "Go for a run, NOW!" and then if within 15 minutes it doesn't see that your phone is moving at more than 6 MPH it intentionally fries the circuits.
"the prevalence of chronic disease continues to rise"
Related: our country keeps getting older and disease screening is getting more accurate. Not saying that we couldn't be healthier, but if we want to reduce prevalence _rates_ we should probably boost immigration of young people. Did I YglesiAce the test?
"(A) busy professional with diabetes who is struggling to manage blood sugar levels."
Welp, this is me, although I'm not technically diabetic, at least not yet. And for me, there is no use case for this sort of thing. I know which foods are healthy - and I already eat them! I know that I should exercise - and I do! I wear a CGM (the Abbott one) so I know, literally minute-to-minute, what my glucose levels are, and if they get really out of whack in either direction, an alarm goes off. I certainly don't need push notifications to "remind" me to eat a healthy meal. I'm not an idiot!
What I *would* like is a better understanding of why I'm in this situation in the first place, being a healthy, active person with a BMI < 20 and very good eating habits. Yes apparently there is some family history, as I have belatedly learned. But why does half a sandwich send it higher than a little scoop of ice cream? Why does it crash so low when I'm sleeping? Why is my A1C high even though day-to-day the levels don't seem that high? If there were some kind of sophisticated analysis that could help me figure some of this out on offer, I would be interested.
Matthew writes, “ if you’re from a low-income county or have lower degrees of education, you are more likely to develop a chronic disease.”
Chronic disease is so high in America because people eat terribly and don’t exercise. That’s it. Sure people in poorer areas tend to have more of these bad habits. This is correlative because they poorness and bad eating/exercise derive from the same thing: a lack of self-discipline and sense.
There are lots of bad habits that poor people have more than rich people. Just guessing here, but I bet they watch more TV, read less, lay on the couch more, all kinds of terrible habits. The poor who don’t have these bad habits won’t help their statistics much, either, because they soon won’t be poor.
Everybody in my extended family came from roughly the same socioeconomic situation. The ones who ended up poorer had these very bad habits, and the ones who ended up richer generally didn’t. And reader, it’s the same in your family, except for the occasional bad luck story.
It's interesting to contrast this correlation (which I agree is strong and real_ with international and historical data points. People in the past in the US were less healthy but probably not because of sitting on the couch and eating junk food. Likewise, people in say, Peru have a lower life expectancy, but probably not because of poor individual choices.
Were people in the US in the past less healthy? I seriously doubt that. Ever since we got fat we are absurdly unhealthy. We are the worst in the developed world if I’m not mistaken. They have engineered food to be addictive and cheap. They did a good job.
Well, there's different ways to talk about health, but life expectancy would probably be the single stat I'd look to if I had to pick one, and that has steadily risen over the past 100 years before plateauing, mainly due to overdoses, around 2018 or so.
If you go back to just 1960 life expectancy was 10 years shorter. Sure people were skinnier, but a lot of them were suffering as adults from childhood diseases. My grandma had heart problems due to a scarlet fever (strep throat, essentially) infection as a child and her sister was deaf in one ear due to some other childhood disease. One of the reasons government got behind medicaid was that so many draftees were getting rejected due to treatable childhood health problems and malnourishment.
Also, when I've looked it up before, the data seems to show that being fat isn't healthy but as long as your not obese it only takes 2 or 3 years off of life expectancy.
Yeah, no doubt that modern medicine eliminated things like Polio and Scarlet fever, and that had a big effect on life expectancy. Also, medical interventions for heart attacks and trauma extend it.
It is a big “if” to except obese people to arrive at 2 or 3 years. In the US, 40% of adults over 20 are obese. Almost 10% are severely obese. Our dietary problems are probably 10 times more damaging than covid-19 ever was, and they are hitting children hard and ruining their lives. At least covid left the children largely alone.
Our food supply should be the single biggest issue in American politics, but the politicians are all bought off, and of the two guys who might be wild cards, Trump is a certified junk food junkie, and RFK has no chance.
The FDA, USDA, and the major medical schools are all in the pocket of the processed food industry and big pharma (which profits mightily on chronic disease).
Honestly, it really just sounds like the government should do some research into making cheap, healthy meals that you can microwave and then subsidize them, possibly through a SNAP rebate to the retailer. (I.e., when you pay with SNAP, you get a discount and the retailer gets a rebate for it.)
Here's some possibly relevant experience, I recently replaced my old analog wrist timepiece with an Apple watch, which has a whole series of prompts on it of this type. It monitors my exercise and urges me to surpass goals that it suggested I set. Frankly, only one of these elements has any effect on my behavior: it tells me to stand every hour...and I try to respond, and I think it's good for me that I do. All the rest....I like the record-keeping, but if I compare my behavior before and after purchase, it has made zero difference.
The other element that is key to the Apple watch's single success with me is that it has the ability to measure things about me...it's the fact that it's attached to my wrist and knows when I stand. The key to its success is not it's intelligence, it is its interface with me. So how will the Thrive AI App influence, say, my diabetes? Will it measure whether I have just consumed a sugary dessert? To my mind, more important than analyzing the medical literature is the interface that measures what I am actually doing, and what is actually happening in my body.
This actually sounds like a reasonable use for AI - use some personal information and a detailed body of knowledge to offer tips on living healthy. Far better than the sort of general AI of a ChatGPT. However....
1. What makes this application better than the health apps out there now? Why should my health care system or my insurance company be willing to purchase this product over the "nudges" they already use? Why should I?
2. If this is to be a personal nag, er, coach, why would I trust these people with personal health information? it's one thing for my insurance company or my health care system - they already know everything. Do I want that information placed in another set of (leaky) hands?
3. As Ben noted, a significant chunk of lifestyle-related disease happens to people who lack the means to use any of this knowledge. I don't know if that particular observation really affects the utility of an AI-powered health coaching app, but for anyone who cares about the metrics on this issue there's a chunk of people that AI can't help.
4. As I saw down the comments, there's another subset of potential lifestyle-related disease candidates who are already getting the needed health info, and ignoring it. They already know what to do, they just don't wanna. AI-driven shock collar anyone? Lest i sound smug, I AM one of these...
It's strange that Altman and Huffington aren't contemplating biometrics: AI has proven to be really good at identifying unexpected statistical relationships in dense, chaotic data streams.
Maybe there are technical hurdles (more likely, fear of liability), but that seems like the basis for significant public health gains. An app that could make lifestyle recommendations and back them up with detailed morbidity predictions tied to your personal data ought to be more convincing than generic nudges. (To say nothing of offering preliminary diagnostics, consolidated health scheduling, and other potential use cases.)
Here's the thing: this sounds a lot like what Noom already does. I tried Noom! I also quit. It was better than 90% of the diet apps I have already tried. But also, there is only so much that can be done through nagging. Eventually, people get tired of being nagged, they get tired of constantly tracking their food, they get tired of calorie restriction. AI nagging is not going to change that.
Now, could it be useful for people whose chronic conditions are more like "here's your reminder to take your medications"? Possibly, especially if we're talking about people who are old. I'd also be fine with an AI advice nurse type service if it could be trained to be as useful as the normal advice nurses. But the lure of tasty food is much stronger than the power of nagging ever will be.
Interesting speculative story on AI, specifically Google Mini / Amzaon Alexa. It's a story of two steps forward, 1 step back.
A year or two ago I noticed Google Mini got much worse, which I chalked up to perhaps Google abandoning the project because it wasn't making them money. But recently my dad was telling me about an article he read on how Amazon was behind on GenAI.
The article cited Alexa as an example, because it currently exists as a silo'd product, with different teams developing the "ask a question" / "play music" / "set the thermostat" aspects. They couldn't fit GenAI over it because Alexa would get confused about context. So for example you might be cold and ask it to "turn up the heat" and it would play "it's getting hot in here" by Nelly.
Anyways - it made me wonder if that's what killed my Google Mini. Google is "ahead" according to the article, but I'm guessing they are merely foolhardy, or at least breaking things in the name of progress. My guess is they applied their new GenAI to the Google Mini and broke up or reorg'd the silos that had previously had "ask a question" / "play music" / "set the thermostat" pretty much nailed down.
Anyways, that was just my speculation. But if anyone else has noticed their mini is stupid now, and has another explanation I'm all ears.
What if the health advice is wrong? What if the bot tells people to do something that turns out to be counterproductive? Let’s say it draws on what was sound science for decades and tells us avoid exposing infants and toddlers to potential allergens like peanuts? Or if it begins recommending that we personally follow strict masking guidelines to avoid respiratory diseases as was recommended during the pandemic?
The news is constantly full of “studies” that tell us one thing or another is going to kill us or cure us. Our socials and podcasts bombard us with supplements and diets. Much of this turns out not to work or to be outright fraudulent. Why should I trust this bot at all?
My BMI is 29 but I’m also in the best shape of my life. My blood work and physical this year were completely normal. The app my doctors makes me use to manage everything from appointments to test results to ordering prescription refills also tells me I need to lose weight and that I’m at risk for diabetes and cardiovascular disease. I showed my doctor and he said I should just ignore it because I’m doing great.
You want to fix the social determinants of health? Dump Ozempic into the water supply.
Ozempic is made of protein and therefore digested in the stomach before it can be absorbed. So you will need a government injection campaign, and people tend to get fussy about those.
Forcing people to change behaviors is HARD.
Context is key: Generative AI may not always be accurate, but that's because humans aren't always accurate.
To wit, I've received some patently incorrect information about diabetes from many healthcare professionals. I don't expect AI to be better, but it probably scales more efficiently.
You are invited to read about my medical system travails here: http://TypeADiabetic.com
Very demure, very mindful.
The second generation of the app should have a feature where it alerts you "Go for a run, NOW!" and then if within 15 minutes it doesn't see that your phone is moving at more than 6 MPH it intentionally fries the circuits.
I would like us to consider the possibility that both health status and income are endogenously determined by individual characteristics.
I am a shill for Noom, but they use psychology and gamification to get better health outcomes, and I think it's a much better strategy than AI.
"the prevalence of chronic disease continues to rise"
Related: our country keeps getting older and disease screening is getting more accurate. Not saying that we couldn't be healthier, but if we want to reduce prevalence _rates_ we should probably boost immigration of young people. Did I YglesiAce the test?
"(A) busy professional with diabetes who is struggling to manage blood sugar levels."
Welp, this is me, although I'm not technically diabetic, at least not yet. And for me, there is no use case for this sort of thing. I know which foods are healthy - and I already eat them! I know that I should exercise - and I do! I wear a CGM (the Abbott one) so I know, literally minute-to-minute, what my glucose levels are, and if they get really out of whack in either direction, an alarm goes off. I certainly don't need push notifications to "remind" me to eat a healthy meal. I'm not an idiot!
What I *would* like is a better understanding of why I'm in this situation in the first place, being a healthy, active person with a BMI < 20 and very good eating habits. Yes apparently there is some family history, as I have belatedly learned. But why does half a sandwich send it higher than a little scoop of ice cream? Why does it crash so low when I'm sleeping? Why is my A1C high even though day-to-day the levels don't seem that high? If there were some kind of sophisticated analysis that could help me figure some of this out on offer, I would be interested.
Yes, this is exactly why I have it. My doctor suggested it and I was kind of surprised it was covered. I pay about $45 with insurance.
Matthew writes, “ if you’re from a low-income county or have lower degrees of education, you are more likely to develop a chronic disease.”
Chronic disease is so high in America because people eat terribly and don’t exercise. That’s it. Sure people in poorer areas tend to have more of these bad habits. This is correlative because they poorness and bad eating/exercise derive from the same thing: a lack of self-discipline and sense.
There are lots of bad habits that poor people have more than rich people. Just guessing here, but I bet they watch more TV, read less, lay on the couch more, all kinds of terrible habits. The poor who don’t have these bad habits won’t help their statistics much, either, because they soon won’t be poor.
Everybody in my extended family came from roughly the same socioeconomic situation. The ones who ended up poorer had these very bad habits, and the ones who ended up richer generally didn’t. And reader, it’s the same in your family, except for the occasional bad luck story.
It's interesting to contrast this correlation (which I agree is strong and real_ with international and historical data points. People in the past in the US were less healthy but probably not because of sitting on the couch and eating junk food. Likewise, people in say, Peru have a lower life expectancy, but probably not because of poor individual choices.
Were people in the US in the past less healthy? I seriously doubt that. Ever since we got fat we are absurdly unhealthy. We are the worst in the developed world if I’m not mistaken. They have engineered food to be addictive and cheap. They did a good job.
Well, there's different ways to talk about health, but life expectancy would probably be the single stat I'd look to if I had to pick one, and that has steadily risen over the past 100 years before plateauing, mainly due to overdoses, around 2018 or so.
If you go back to just 1960 life expectancy was 10 years shorter. Sure people were skinnier, but a lot of them were suffering as adults from childhood diseases. My grandma had heart problems due to a scarlet fever (strep throat, essentially) infection as a child and her sister was deaf in one ear due to some other childhood disease. One of the reasons government got behind medicaid was that so many draftees were getting rejected due to treatable childhood health problems and malnourishment.
Also, when I've looked it up before, the data seems to show that being fat isn't healthy but as long as your not obese it only takes 2 or 3 years off of life expectancy.
Yeah, no doubt that modern medicine eliminated things like Polio and Scarlet fever, and that had a big effect on life expectancy. Also, medical interventions for heart attacks and trauma extend it.
It is a big “if” to except obese people to arrive at 2 or 3 years. In the US, 40% of adults over 20 are obese. Almost 10% are severely obese. Our dietary problems are probably 10 times more damaging than covid-19 ever was, and they are hitting children hard and ruining their lives. At least covid left the children largely alone.
Our food supply should be the single biggest issue in American politics, but the politicians are all bought off, and of the two guys who might be wild cards, Trump is a certified junk food junkie, and RFK has no chance.
The FDA, USDA, and the major medical schools are all in the pocket of the processed food industry and big pharma (which profits mightily on chronic disease).
Honestly, it really just sounds like the government should do some research into making cheap, healthy meals that you can microwave and then subsidize them, possibly through a SNAP rebate to the retailer. (I.e., when you pay with SNAP, you get a discount and the retailer gets a rebate for it.)
*Factor meals has entered the chat*
Here's some possibly relevant experience, I recently replaced my old analog wrist timepiece with an Apple watch, which has a whole series of prompts on it of this type. It monitors my exercise and urges me to surpass goals that it suggested I set. Frankly, only one of these elements has any effect on my behavior: it tells me to stand every hour...and I try to respond, and I think it's good for me that I do. All the rest....I like the record-keeping, but if I compare my behavior before and after purchase, it has made zero difference.
The other element that is key to the Apple watch's single success with me is that it has the ability to measure things about me...it's the fact that it's attached to my wrist and knows when I stand. The key to its success is not it's intelligence, it is its interface with me. So how will the Thrive AI App influence, say, my diabetes? Will it measure whether I have just consumed a sugary dessert? To my mind, more important than analyzing the medical literature is the interface that measures what I am actually doing, and what is actually happening in my body.
This actually sounds like a reasonable use for AI - use some personal information and a detailed body of knowledge to offer tips on living healthy. Far better than the sort of general AI of a ChatGPT. However....
1. What makes this application better than the health apps out there now? Why should my health care system or my insurance company be willing to purchase this product over the "nudges" they already use? Why should I?
2. If this is to be a personal nag, er, coach, why would I trust these people with personal health information? it's one thing for my insurance company or my health care system - they already know everything. Do I want that information placed in another set of (leaky) hands?
3. As Ben noted, a significant chunk of lifestyle-related disease happens to people who lack the means to use any of this knowledge. I don't know if that particular observation really affects the utility of an AI-powered health coaching app, but for anyone who cares about the metrics on this issue there's a chunk of people that AI can't help.
4. As I saw down the comments, there's another subset of potential lifestyle-related disease candidates who are already getting the needed health info, and ignoring it. They already know what to do, they just don't wanna. AI-driven shock collar anyone? Lest i sound smug, I AM one of these...
5. So....I'm thinking Ozempic all around....
It's strange that Altman and Huffington aren't contemplating biometrics: AI has proven to be really good at identifying unexpected statistical relationships in dense, chaotic data streams.
Maybe there are technical hurdles (more likely, fear of liability), but that seems like the basis for significant public health gains. An app that could make lifestyle recommendations and back them up with detailed morbidity predictions tied to your personal data ought to be more convincing than generic nudges. (To say nothing of offering preliminary diagnostics, consolidated health scheduling, and other potential use cases.)
Here's the thing: this sounds a lot like what Noom already does. I tried Noom! I also quit. It was better than 90% of the diet apps I have already tried. But also, there is only so much that can be done through nagging. Eventually, people get tired of being nagged, they get tired of constantly tracking their food, they get tired of calorie restriction. AI nagging is not going to change that.
Now, could it be useful for people whose chronic conditions are more like "here's your reminder to take your medications"? Possibly, especially if we're talking about people who are old. I'd also be fine with an AI advice nurse type service if it could be trained to be as useful as the normal advice nurses. But the lure of tasty food is much stronger than the power of nagging ever will be.
Interesting speculative story on AI, specifically Google Mini / Amzaon Alexa. It's a story of two steps forward, 1 step back.
A year or two ago I noticed Google Mini got much worse, which I chalked up to perhaps Google abandoning the project because it wasn't making them money. But recently my dad was telling me about an article he read on how Amazon was behind on GenAI.
The article cited Alexa as an example, because it currently exists as a silo'd product, with different teams developing the "ask a question" / "play music" / "set the thermostat" aspects. They couldn't fit GenAI over it because Alexa would get confused about context. So for example you might be cold and ask it to "turn up the heat" and it would play "it's getting hot in here" by Nelly.
Anyways - it made me wonder if that's what killed my Google Mini. Google is "ahead" according to the article, but I'm guessing they are merely foolhardy, or at least breaking things in the name of progress. My guess is they applied their new GenAI to the Google Mini and broke up or reorg'd the silos that had previously had "ask a question" / "play music" / "set the thermostat" pretty much nailed down.
Anyways, that was just my speculation. But if anyone else has noticed their mini is stupid now, and has another explanation I'm all ears.
What if the health advice is wrong? What if the bot tells people to do something that turns out to be counterproductive? Let’s say it draws on what was sound science for decades and tells us avoid exposing infants and toddlers to potential allergens like peanuts? Or if it begins recommending that we personally follow strict masking guidelines to avoid respiratory diseases as was recommended during the pandemic?
The news is constantly full of “studies” that tell us one thing or another is going to kill us or cure us. Our socials and podcasts bombard us with supplements and diets. Much of this turns out not to work or to be outright fraudulent. Why should I trust this bot at all?
My BMI is 29 but I’m also in the best shape of my life. My blood work and physical this year were completely normal. The app my doctors makes me use to manage everything from appointments to test results to ordering prescription refills also tells me I need to lose weight and that I’m at risk for diabetes and cardiovascular disease. I showed my doctor and he said I should just ignore it because I’m doing great.
You want to fix the social determinants of health? Dump Ozempic into the water supply.
Ozempic is made of protein and therefore digested in the stomach before it can be absorbed. So you will need a government injection campaign, and people tend to get fussy about those.