There is a lot of authoritative sounding text here but no real explanation of how this huge process solved the mystery fatigue. In fact she says in the middle that the specialist she saw was more valuable. Toward the end she mentions experiments that were run, but the heavy hitters appear to be from specialists adjusting the medication related to her tumor treatment and a real person reviewing her food diary and noticing she was in a caloric deficit.
For all of the writing about collecting data and using AI, she never explains what the AI did for her. Maybe it suggested trying iron supplements because her ferritin was slightly low on a blood test? I hope it explained that these blood tests can have transient highs and lows and that excess iron from supplements can have major negative health effects if not monitored.
For all of the lead up about using AI, there was never really a delivery where the AI was used to solve the mystery fatigue.
I’m glad this person is feeling better, but this post reminds me of those LinkedIn posts where someone is trying really hard to show their AI skills but used an example where the AI didn’t contribute to or accelerate the outcome.
My gut as a physician says AI will revolutionize primary care the most and the premise of AI having more specialized knowledge than a PCP holds water - I think the future of primary care is AI equipped midlevel providers - so I was very excited by the intro only to be let down. A lot of buzz about nothing sadly.
The only takeaway here is logging helps detect patterns, we already knew that.
For simple things, if an AI (combined with blood tests) can do what the doctor can do, then technically I don't need the doctor. I also don't need the friction that comes in working with a doctor, considering the AI is close to frictionless.
> ... "considering the AI is close to frictionless" ...
I feel like this bit is more of a problem than many folks realize. I know that people see "frictionless" as a desirable trait, but for many of the things folks are using AI for these days, you're gonna want "push-back" sometimes. Especially in cases of health care and computer programming, there are times when the user is straight-up wrong about what they're trying to do, say, or believe, and the last thing you wanna hear when your life or job or the lives or jobs of others are on the line and you're about to make a huge mistake is "You're absolutely right! Let me get right on doing that for you!"
You're absolutely right! What I do is have the LLM rank and score candidates for me given my situation. This helps me focus on what's more likely versus less likely. Blood tests are not free; they're in fact fairly expensive in aggregate, so this filtering has been essential.
I get it, and it draws parallels with debugging and incident response in software dev. Your best tool is good data, if you have more data, you’re able to make good judgement in short time. If AI helps you collect good data about your condition, then presenting that to a clinician is certainly better than “it hurts, idk” because there’s only so much they can do at the time
Even for the part where she discovers she was in a caloric deficit and that carbs helped, the AI didn't seem to have a role:
> Figuring out what to track is an iterative process. At first, I logged only daily calories. But when I noticed that a small amount of carbs sometimes helped me recover faster mid-episode, I tracked carbs on the same spreadsheet.
The carb realization happened before tracking carbs. And why did it take a specialist to notice she was in a 300 calorie deficit? That seems like something that any AI process would trivially notice.
I don't know. This article is puzzling. I think using AI to analyze health records is interesting, but this article didn't really have anything supporting that other than to mention that she put everything into Claude on the side.
I solved a similar issue with the help of AI. I doubt I would have ever solved it with the help of a PCP, because of really subtle nuances. It’s been almost a year and my symptoms have basically disappeared.
AI helped me figure out my symptoms were related to histamine issues. This was really hard to track down. I started to tune into chicken and eggs causing me issues, but (1) I wasn’t actually allergic (2) chicken and rice is a standard safe diet (3) chicken is low histamine itself, but can trigger histamine release . It was further complicated because a bunch of foods didn’t consistently trigger reactions.
It was only after I tracked a bunch of foods and tried a bunch of different remedies that Claude was able to track down the pattern. From there, I was able to understand what foods would trigger my issues.
Maybe I am missing it, but I don’t see where AI solved anything in this anecdote. Rather, a team of medical professionals and a patient willing to advocate for herself did.
Beside ferritin, hemoglobin shouldn't be overlooked as the determinant of iron status. Hemoglobin is a part of the CBC (Complete Blood Count) test.
Iron supplements are notorious for irritating the stomach, worsening gastritis and acid reflux. Even iron bisglycinate, while fully effective, is a significant stomach irritant. Ferric pyrophosphate might be safer, and I'm trying it now, but I have to take it twice a day because once wasn't sufficient.
With regard to addressing anemia, active B12 (hydroxo/methyl/adenosyl), methylfolate e.g. Quatrefolic, copper, and vitamin A also matter.
One can also simultaneously be low or suboptimal in many other vitamins and minerals unrelated to anemia, e.g. vitamin D. Fixing only one of them is no guarantee of a sustained solution.
With regard to blood tests, it can take many iterations of doing them to find anything useful. Doctors are ridiculous to work with in this regard, so I bypass them and get the tests myself. Doctors work for the insurance firm, not for you.
> As long as it is in the normal range for your sex and age, you're fine even if your ferritin remains at 0.
This is simply wrong when one considers that iron has many bodily functions outside of hemoglobin, such as acting as a cofactor in tyrosine hydroxylase for dopamine synthesis.
Alternate day dosing is a great tip for better stomach tolerability! If ferric pyrophasphate twice daily doesn't fix it for me, or if it becomes intolerable, then I will surely consider alternate-day iron bisglycinate. Thanks.
There is a lot of authoritative sounding text here but no real explanation of how this huge process solved the mystery fatigue. In fact she says in the middle that the specialist she saw was more valuable. Toward the end she mentions experiments that were run, but the heavy hitters appear to be from specialists adjusting the medication related to her tumor treatment and a real person reviewing her food diary and noticing she was in a caloric deficit.
For all of the writing about collecting data and using AI, she never explains what the AI did for her. Maybe it suggested trying iron supplements because her ferritin was slightly low on a blood test? I hope it explained that these blood tests can have transient highs and lows and that excess iron from supplements can have major negative health effects if not monitored.
For all of the lead up about using AI, there was never really a delivery where the AI was used to solve the mystery fatigue.
I’m glad this person is feeling better, but this post reminds me of those LinkedIn posts where someone is trying really hard to show their AI skills but used an example where the AI didn’t contribute to or accelerate the outcome.
My gut as a physician says AI will revolutionize primary care the most and the premise of AI having more specialized knowledge than a PCP holds water - I think the future of primary care is AI equipped midlevel providers - so I was very excited by the intro only to be let down. A lot of buzz about nothing sadly.
The only takeaway here is logging helps detect patterns, we already knew that.
For simple things, if an AI (combined with blood tests) can do what the doctor can do, then technically I don't need the doctor. I also don't need the friction that comes in working with a doctor, considering the AI is close to frictionless.
> ... "considering the AI is close to frictionless" ...
I feel like this bit is more of a problem than many folks realize. I know that people see "frictionless" as a desirable trait, but for many of the things folks are using AI for these days, you're gonna want "push-back" sometimes. Especially in cases of health care and computer programming, there are times when the user is straight-up wrong about what they're trying to do, say, or believe, and the last thing you wanna hear when your life or job or the lives or jobs of others are on the line and you're about to make a huge mistake is "You're absolutely right! Let me get right on doing that for you!"
You're absolutely right! What I do is have the LLM rank and score candidates for me given my situation. This helps me focus on what's more likely versus less likely. Blood tests are not free; they're in fact fairly expensive in aggregate, so this filtering has been essential.
I get it, and it draws parallels with debugging and incident response in software dev. Your best tool is good data, if you have more data, you’re able to make good judgement in short time. If AI helps you collect good data about your condition, then presenting that to a clinician is certainly better than “it hurts, idk” because there’s only so much they can do at the time
But where does that happen in the article?
Even for the part where she discovers she was in a caloric deficit and that carbs helped, the AI didn't seem to have a role:
> Figuring out what to track is an iterative process. At first, I logged only daily calories. But when I noticed that a small amount of carbs sometimes helped me recover faster mid-episode, I tracked carbs on the same spreadsheet.
The carb realization happened before tracking carbs. And why did it take a specialist to notice she was in a 300 calorie deficit? That seems like something that any AI process would trivially notice.
I don't know. This article is puzzling. I think using AI to analyze health records is interesting, but this article didn't really have anything supporting that other than to mention that she put everything into Claude on the side.
I solved a similar issue with the help of AI. I doubt I would have ever solved it with the help of a PCP, because of really subtle nuances. It’s been almost a year and my symptoms have basically disappeared.
AI helped me figure out my symptoms were related to histamine issues. This was really hard to track down. I started to tune into chicken and eggs causing me issues, but (1) I wasn’t actually allergic (2) chicken and rice is a standard safe diet (3) chicken is low histamine itself, but can trigger histamine release . It was further complicated because a bunch of foods didn’t consistently trigger reactions.
It was only after I tracked a bunch of foods and tried a bunch of different remedies that Claude was able to track down the pattern. From there, I was able to understand what foods would trigger my issues.
Maybe I am missing it, but I don’t see where AI solved anything in this anecdote. Rather, a team of medical professionals and a patient willing to advocate for herself did.
If AI enabled her to advocate for herself to the right professionals, and do so with data in hand, that’s huge.
AI guiding a process is valuable even if it doesn’t replace all the parts of the process.
But the article doesn't even make that claim. The doctors who adjusted her medication were already her care team for the brain tumor.
She said she started logging calories but recognized the carb connection before she logged carbs.
I kept waiting for the payoff where the AI led to something, but it never came.
Beside ferritin, hemoglobin shouldn't be overlooked as the determinant of iron status. Hemoglobin is a part of the CBC (Complete Blood Count) test.
Iron supplements are notorious for irritating the stomach, worsening gastritis and acid reflux. Even iron bisglycinate, while fully effective, is a significant stomach irritant. Ferric pyrophosphate might be safer, and I'm trying it now, but I have to take it twice a day because once wasn't sufficient.
With regard to addressing anemia, active B12 (hydroxo/methyl/adenosyl), methylfolate e.g. Quatrefolic, copper, and vitamin A also matter.
One can also simultaneously be low or suboptimal in many other vitamins and minerals unrelated to anemia, e.g. vitamin D. Fixing only one of them is no guarantee of a sustained solution.
With regard to blood tests, it can take many iterations of doing them to find anything useful. Doctors are ridiculous to work with in this regard, so I bypass them and get the tests myself. Doctors work for the insurance firm, not for you.
> As long as it is in the normal range for your sex and age, you're fine even if your ferritin remains at 0.
This is simply wrong when one considers that iron has many bodily functions outside of hemoglobin, such as acting as a cofactor in tyrosine hydroxylase for dopamine synthesis.
Apologies. I will atone for it.
Have you tried supplementing iron every other day instead?
https://pmc.ncbi.nlm.nih.gov/articles/PMC12372323/
Alternate day dosing is a great tip for better stomach tolerability! If ferric pyrophasphate twice daily doesn't fix it for me, or if it becomes intolerable, then I will surely consider alternate-day iron bisglycinate. Thanks.