Shocking Truth About AI, Chronic Disease, Toxins, Diet & Lifestyle For Longevity | Eric Topol

It's going to be the biggest shake up.

I think it's going to redefine medicine. I've been kind of dreaming of this moment in healthcare where we actually could make sense of human biology with the help of artificial intelligence because it's so complex. The human body has got 37 trillion cells. There's 37 billion trillion chemical reactions that happen in the body every second. And, you know, when you go to the doctor, you basically get your panel of 20 or 30 or 50 analytes. He does, he pokes around, listens to your heart, lungs, and maybe does a little x-ray here. there, sticks a scope in. But it's really the dark ages compared to where we are headed. And you're at the forefront of understanding the intersection of the rapid changes in our understanding of human biology and the omics revolution and systems medicine with technology and with artificial intelligence. So, I think most people don't understand that doctors are working in an analog way, even though we have electronic medical records, the way we practice medicine really hasn't changed much in 150 years. And we don't use computers to help us sort through data, make decisions, and figure out what to do for our patients. And there are now examples of this. So maybe you can start by talking about, you know, how did you first get interested in artificial intelligence in medicine and how do you see it having made change already? with certain things like imaging or retinal scans or interpretations of EKGs that are helping us understand and diagnose diseases better and earlier. And then we'll talk about a little bit later about where do we see all this going? Because right now we're doing this sort of medicine better, but is it the right, only the right thing to be doing?

Sure. Well, I think the interesting thing about the AI scene is it really didn't get real until, let's say, seven, eight years ago. And it really, for our space of medicine, it was confined to medical images, scans. And that was the deep learning phase of AI. And it really has been formidable. That is just about every type of scan you can imagine, but past slides, electrocardiograms, the retinas you mentioned, skin lesions, they could be interpreted as well or better by machines that were trained with so-called supervised learning, meaning that of course you had to have thousands, tens of thousands, hundreds of thousands images that were annotated by expert physicians, and then you could train a model to do better than humans. So that was really great and back in 2019 when I wrote Deep Medicine, it was about that phase of deep learning.

But that's like ancient history now, right? 2019.

Yeah, I know. It's amazing how quickly that has gone. Yeah, really, Mark. But what's interesting is, you know, I wrote in the book that what we need is a new model because we didn't have one that could take all the layers of what makes us unique. You know, you've alluded to that, not just electronic health record. but our genome, our gut microbiome, our sensors, our environment, our immune, the works, right? And the fact that that data changes over time and the fact that we could get the corpus of medical knowledge into that as well. So that's where we are now with this transformer model, also known as large language model phase, which is of course got major jump in a year ago with ChatGPT and now of course the GPT-4, Gemini and future models, GPT-5 sometime next year in fact. And that of course is getting us to that state where we could take all that data for a given patient or individual and be able to not only define what is so critical about predicting a condition, better treatment, better prevention. So we're on the cusp, but we haven't done it yet, to be honest. So, no one has actually done multiple layers. They've done electronic health records and a genome, electronic health records and a scan. But to take multiple layers, including sensors, That's an analytical AI challenge that has yet to be solved. It will be imminently, and that's exciting.

Yeah, I mean, you know, when you're talking about it, you wrote an article that I thought was just so prescient and it was such a good description in a short amount of time, and I encourage people to read it called As Artificial Intelligence Goes Multimodal, Medical Applications Multiply. And you talked about how we're going to be getting high-dimensional data that underlie the uniqueness of all of us and how it can be captured from all these different sources that you mentioned, including all the biomarkers we have through biosensors, wearables, implantables, our genome, our microbiome, our metabolome, our immunome, the transcriptome, proteome, epigenome, it goes on and on. And then, our electronic health records, our lab tests, our family histories, unstructured text from our medical records, and also things that are air pollution sensors we could be wearing. I just got one of those that someone sent me to try to wear to my air pollution, environmental stressors, All these things are going to be then informed by the whole National Library of Medicine database of peer reviewed data. And it's going to create so much information. And it seems to me there's so some intersection of a number of trends right now which are going to transform medicine in a way that we can barely imagine. And it's going to happen very soon, which is the omics revolution, the systems biology and medicine revolution. the biosensors and wearable revolution, and then the AI, machine learning, and big data analytic capacity that we have. And so those five basic trends are all converging in a way that I think is within even four or five years, we're going to see medicine be profoundly different because the acceleration of this is happening so fast. And I'm excited about it because I feel like I've been trying to, with my little brain, put my head around all these immense complexity of human biology, which we've managed to navigate through this reductionist model of medicine and science into siloed specialties where you're super sub-sub-sub-sub specialist on X, Y, or Z topic, but you don't understand how it all connects and interacts. And so the first time with AI, it seems like we're not able to do that. So how do you see this unfolding and how is this kind of happening and where are we going? Because I feel like I'm sitting on the edge of my seat and right now, I feel like we're about to kind of get out of our little dark ages and enter into an era where we're gonna be able to make a real transformation in people's health.

Well, I think you're right. It's extraordinary, this convergence that you're getting at. And it's gonna happen in phases. So the first one is more of the practical, which is, I've been calling keyboard liberation.

Thank God, I heard that you say that. I'm like, hallelujah. 'Cause every doctor is stuck on their keyboard looking at the computer instead of looking at the patient. That is so huge.

It's hated mutually by doctors and nurses and patients. I mean, everything that people love to hate because it's destroyed that bond, that human bond. And that's going to be basically history of data clerk function because we already see now in many health systems around the country that you can do all of this. through the conversation. The only adjustment you have to make, Mark, is to articulate the physical exam findings with the patient. But other than that, the notes are far superior than the ones that are pecked along. And what's great is once you have that note digitized and it's got all the juice in it, two big things happen. One is that of course you could put in any format conducive for the patient, you know, in terms of educational level or language or, you know, whatever cultural bent. You could also, that patient has the audio file, so if they don't understand something in that note, they can link it right to the audio file, listen to it again. And you know how many patients that you see where they're confused or they don't remember things. But the other big thing is on the clinician side, instead of having to peck through all this stuff, the orders for new tests and labs and return appointments, prescriptions, billing, pre-authorization, it's all done, it's all done. And the nudges to the patient subsequent about the things that were discussed, like blood pressure, did you check what were the results? You know, the AI picks that up, gets it back to the physician, you know, all these things are now automated. So that will in itself be well you know, instead of the things that all clinicians want to hate, this is, I think, something that will be widely embraced. And there's no, you know, as you know very well, Mark, there's a lot of concerns about confabulation, hallucination, but that doesn't apply here. I mean, this is not that the AI is not going to be making things up about this kind of thing.

Do you have that in your office yet? Do you have that in your office?

I've used it. At Scripps Health, where I have cardiology practice, they haven't used what I consider the best of these, but they have done a pilot. The largest one is a Microsoft Nuance, but the company that I've advised is a Bridge Health, which should just derive from University of Pittsburgh and Carnegie Mellon, but there's been several. I mean, there's about 20 of these out there, various testing.

Sam, I want to get one right away for my practice.

I think this is inevitability because this is finally the payback for all these bad years of having to become data clerks. But it's just the beginning. It's just one thing that's going to be remarkably different.

That helps us to care better, but it doesn't change what we're doing. In other words, we're going to be able to read X-rays better and MRI imaging better and pathology reports better and EKGs better. retinal imaging that tells us so much about a patient's health. And these are incredible advances that are going to create much more refinement and understanding of how to be precise in our diagnosis of patients. And that's going to up-level medicine for sure.

But let me, if can I just-.

Yeah, go, Jon, take one thing. Yeah.

Because the retinal image is something that is extraordinary. So before we just pass over that.

Yeah, yeah, no, it's not.

You know, I just want to point out that, you know, the original task was to see if the AI could interpret the image as well as a clinician. But what wasn't envisioned is that the AI could see things that humans will never see. So with the retina, as you touched on, the ability to predict Alzheimer's disease, Parkinson's disease five to seven years before there's any symptoms. the issue of, of course, the hepatobiliary tract, kidney disease, cardiac risk, risk of, you know, across all systems, diabetes control, blood pressure control. Someday we will be taking pictures of our own retina and get as a checkup with an AI. So it's pretty amazing. And that course that extends to cardiograms and chest X-ray, each of them, there's all this stuff that the AI I can see, if you will, that we will humans will never see it.

So it's even better than humans, right?

Yeah, yeah. I mean, this is why, you know, when I interviewed Jeff Hinton recently for the podcast I do Ground Truths, he said, you know, he's worried about AI because it's getting advanced so quickly, but not for medicine. He thinks this is the sweet spot. This is really where the good is extraordinary. I agree.

I mean, you know, I remember in medical school you had the ophthalmoscope and you had to look in someone's eye and you, okay, you learn about AV nicking and high blood pressure and diabetic retinopathy and macular degeneration. You could see all that stuff, but there wasn't a whole lot else you could kind of figure out, you know? And if you were an ophthalmologist, you might have a few more refinements in your ability to see things, but what you're saying is you can see things like Alzheimer's.

So how, how does it pick that up? What is it actually seeing and looking at, for example, for Alzheimer's? Well, you know, this goes back to when the realization was made and that was when you showed the retina picture to ophthalmologists and you say, is this retina from a man or a woman? They got it right 50% of the time and the AI got it right 97% of the time. And the answer is we don't really know, okay? That is, there's explainability work to define these so-called saliency maps to try to de-convolute the model. But as far as what is it picking up to see the risk of Alzheimer's or Parkinson's or a pattern biliary disease, it isn't clear. I mean, there's some aspects that have been determined, but basically because these models are so extraordinary in terms of what they've learned. And this is all from deep learning. This isn't even from, you know, this Transformer model era.

So can you just stop there for a second? You're talking about deep learning Transformer model. Can you just explain the sort of shift in what you're thinking? And because I don't think most people understand what that is.

Right. So what was the phase of AI that lit up the world that Jeff Hinton and his colleagues like Jan Lacun and many others, they basically found that there was this ability to input data that was supervised. That is, that for our purposes, it was labeled by experts, so-called ground truths. And so they put it, what they knew was the actual image interpretation and trained with tens of hundreds of thousands of these images so that the machine could see stuff.

So this is a knowledge base or expert informed AI, right?

Yeah, yeah. So that really was, you know, deep neural networks. That was the story. It required a single task, unimodal, and then what happened, a Google team in 2017 discovered what they called transformer models, the title of the preprint, Attention is All youl Need. And basically it changed the attention from a single bit of information like a word in a sentence to basically the context of the entire sentence or of course much broader than that. What turned out to be unsupervised putting in the entire internet, Wikipedia, 100,000 books, 200,000 books. That's what the transformer model, large language model, generative AI era that we're in now. It didn't start when ChatGPT was released. last year, but it actually was in incubation. It was being pursued about six years now, but it's now blossomed. We basically have two big types of AI now, the old, if you will, the old and the new.

Yeah, I mean, it just seems it's going to accelerate the pace of medical discovery because if a simple retinal scan can pick up things that we didn't even know we were missing, you know, we didn't even know we didn't know. They were unknown unknowns, as Donald Rumsfeld said. And, and that's just the back of the eye. Imagine when we put in all these things that we just mentioned, the, the whole omics field, the biosensors, the, you know, your, your pictures of what you're eating, your movement pattern. I mean, it's, it's just an enormous amount of data that's going to pick up patterns in that data that we've never seen before and that are going to inform what's happening on a biological level that I think is going to redefine medicine just as we sort of redefine physics from a Newtonian or world flat view to, you know, quantum view to even, you know, beyond that. It's like we're kind of in that era of biology where we're basically have a profound revolution that's going to upend medicine. And, I'd love to hear your perspective on as we sort of enter that era and we start learning these things and understand the body is a network, understand the body is a system, instead of these siloed specialties. How do you see that shifting medicine, medical education, medical practice, reimbursement. I mean, these are massive shift.

Well, it is seismic. It's going to be a challenge. Medicine, as you know, doesn't change easily. And then you throw in all these other practical matters like reimbursement and education, regulatory, trust, implementation. I mean, there's a long list here. of challenges. So, you know, this isn't going to be easy, but it's going to be, you know, the biggest shake up in the history of medicine. The question is how we adapt, how we, you know, our problem at the moment outside of a practical thing like we discussed with the keyboard thing is to get things implemented, we've got to have compelling evidence. And there's a dearth of that because just like you can't get thousands of doctors to annotate images, and that's why this new form, the transformer model, doesn't require supervised learning, it's self-supervised. So it basically is the bypass to what was holding back medicine. But just like that problem, we have the problem of lack of dedication to do prospective trials, whether they're randomized or not. But getting the compelling evidence, which basically says to everyone in the medical community, this is it. This is going to lead to better patient outcomes, better everything. And there's always going to be some risk, of course, when there's never going to be total positive side of the story. But we, except for the gastroenterologists who have done 33 randomized trials colonoscopy with machine vision and a few other randomized trials and radiology that have been quite impressive, particularly mammography. There hasn't been much compelling evidence so far.

Yeah, it's true. It's true. But on the other hand, you look at the amount of deaths are caused by medical practice. Probably a third or fourth leading cause of death are complications or reactions to drugs or medical errors. It's huge. And I was listening to Elon Musk talk about cars and AI and self-driven cars. And he says, about 40,000 people in America die from car accidents every year. What if that was reduced to 10,000? But that's a dramatic drop. But still, you're going to have some people dying from self-driving cars and are we willing to accept that? So I think that's really a point where we have to kind of understand the value proposition and understand that there is some risk, but the upside in terms of reducing our healthcare costs, the burden on our healthcare system is going to be profound.

Well, just to amplify what you just said, we as physicians don't tend to want to acknowledge the problem of medical errors. So I'm glad you brought it up. Yeah. Because the Johns Hopkins study that was published earlier this year in the British Medical Journal, about 800,000 Americans are either dying or seriously disabled each year from diagnostic medical errors. So you would think we'd want to invest in ways to bring that number down to the lowest possible, but that hasn't been the way medicine has worked. The medical error of diagnosis is something that's got to be confronted.

Yeah, I think you're right. I think, you know, when I think about what actually medicine is today, you know, you go see the doctor, they ask you a bunch of questions, they run a bunch of tests, it filters through their one little brain. They went to medical school. I went to medical school in 1983. I don't know what year I went to medical school. I've tried to keep up and, you know, I've had my own narrow experience. and that patient is relying on me to create all the associations, pattern recognition, sort of being able to sift through all the medical literature, whatever the 9 million published articles on PubMed, make sense of what's going on with them. It's kind of embarrassing. I feel like I'm basically groping around in the dark and I still obviously help people and obviously you do and doctors do a good job, but when you think about what's possible in terms of decision support, for practitioners. It's not like the AI is going to be treating the patient, but it's sort of an AI assisted doctor where the doctor then will have to get all that data, then filter through their understanding of medical knowledge, and that decision support from the AI will help them create better outcomes, more personalized treatments, specific care, and really overall improve healthcare. So, how do you see the AI decision support process happening unfolding? Because I think this is where I'm most excited about.

Well, this is happening really quickly, the proof of how the large language model exceeds the doctor's ability to make complex diagnoses. And what is an extraordinary recent preprint, just a number of days ago, was from the Google research team where they took over 300 of the New England Journal clinical pathologic conferences, which are the master clinician trying to get the diagnosis of the really complex, challenging case. And what was amazing is that the large language model got the differential diagnosis randomized against 20 internists with nine years of experience. So these are not just rookies, these are experienced internists. the AI got twice the accuracy, twice than the Internus. And then when you gave the Internus a Google search, it didn't make that much difference. And even when you gave the Internus then being able to go to GPT, well, it was a MedPalm 2 fine tune. We don't know exactly what it is, but it's some large language model. It still didn't get to the large language model alone level. So what we're seeing and that experience is getting replicated that large language models do very well. Now, you and I, in seeing a patient, we wouldn't only rely on the output of the large language model, but what it does, especially in these challenging, complex, it could be rare conditions, just very difficult cases, what it does is it may bring to mind some conditions that didn't pop into our head. And so, you know, it basically this is an exciting thing to see. And as you know, there's been amazing anecdotes about how patients have put their symptoms.

Yeah.

For a mother to her son, for her son, you know, put symptoms of after seeing 17 doctors for three years for a boy who is just, you know, debilitated and growth arrest, and then gets the diagnosis of a cult spina bifida. herself through ChatGPT. I mean, this is happening now. You can't hold back patients from using. Yeah, it's like a Google search, but why do a Google search when you're going to get so much enriched information, which you have to verify, of course.

Yeah, I mean, part of the challenge with this is also that it's going to pick up things that we didn't know and challenge the current model. As we now know, diseases aren't homogenous, and our current diagnostic model is based on symptoms and pathology and geography. You know, what's the symptom, where is it in your body, and what does it look like under a microscope? Not causes and mechanisms and not the infinite complexity of our whole phenome that we just described. And as we start to sort of ingest the phenome into this AI monster, whatever you call it, or AI genius, it's going to come up with a lot of different associations and patterns that reflect a way to personalize medicine to create precision or as you say, individualized medicine. That's what excites me because, you know, Alzheimer's isn't uniform. Diabetes isn't uniform. If you have breast cancer, we know this. Years ago, I read a paper in I think a journal and it was like, you know, the staging of cancer, breast cancer is far less predictive of the outcomes than if you look at the genetics of the cancer and that there is no such thing as breast cancer. There are breast cancers and they're all incredibly heterogeneous. And so, this is not just true for cancer, it's true for everything. and whether it's rheumatoid arthritis or whether it's irritable bowel syndrome or whether it's migraine headaches or whatever the condition is. And we in medicine just treat them all the same. You've got a migraine, that's the diagnosis. And I think we're really good at this diagnostic differential diagnosis process. That's where we get trained in medical school. But it doesn't really take us down the next layer, which is looking at mechanisms and causes and personalization. So how do you see where we're headed in medicine sort breaking that old paradigm and helping us create a more individualized sort of approach to different diseases.

You know, I'm really glad you brought this up because that's kind of the reductionist, you know, just because we just can't deal with this difficulty of heterogeneity of people, of diagnoses, of response to treatments and on and on. We basically dumb it down. and we just say, well, you got this condition or that condition. When, as you said, things like type 2 diabetes, just type 2, there's about 20 subtypes of type 2, and on and on, too. So we will have this kind of pinpoint precision and accuracy, which will help promote very specific individualized treatment. and that will, that can also extend to, you know, what would be the best optimal diet for any given person because, you know, we, we respond to food remarkably differently. Everything we do, we are unique. Even two identical twins are unique. They have very different epigenomes and, and, you know, so we have to appreciate once we get to the ability to deal with the information, which we couldn't until now, and we're just starting to in many respects, then we can start to get to once you recognize that each human is unique, then you can hopefully get to be much better at preventing, diagnosing, treating, getting, you know, better outcomes and promoting health span.

Yeah, I think it's so true. We, we, we in medicine, you know, wait until there's something to look at or do. And, you know, when your blood sugar goes up or your blood pressure goes up or, you know, you get a chest pain or whatever, which your cholesterol goes up. But the truth is that disease is a continuum from optimal wellness and perfect health to this transition slowly over often decades to symptoms and then diseases. And in medicine, we just, we're jumping in the game so late. And I think what AI and machine learning is going to help us do as we start to ingest the phenome is to start to look at these patterns and these transitions from wellness to disease much earlier and be able to intervene earlier before there's ever a symptom. Like you mentioned, you can look at someone's eye and see whether they're going to have Alzheimer's or look at various, even things like voice patterns or, you know, typing on a computer or all these things are actually signals that can be interpreted that we have no clue as doctors what to do with or how to make sense of. to proper application of AI and medicine, we're going to be able to start to sort and sift through this, right?

That's exactly right. The fact that you could predict a condition before any symptoms ever manifest, and we have ways to prevent that condition, to apply for that person. That's what makes this era so extraordinary, right? That we couldn't really have this personalized preventive medicine. And, you know, frankly, Mark, preventive medicine has been a fantasy throughout our careers. We want to prevent primary prevention of illness is like almost unheard of, right? But, you know, secondary prevention, once you've had a heart attack, you know, prevent the second heart attack. Okay, sure. But primary prevention is going to be actualized in the era of AI. for the conditions in which there's a way to truly prevent. So right now, yeah, I wouldn't say our Alzheimer's drugs that have recently been approved are the end all to do that, but we're chipping away at that. Yeah, yeah, really. But we are going to have drugs and interventions that will prevent conditions like neurodegenerative conditions. We are going to have ways to prevent cancer, like the earliest sign to rev up the immune system to squash it, or to go into high mode surveillance to get frequent cell-free tumor DNA blood samples and all sorts of ways to get on top of it before it ever would show up on a scan, before it ever could go to the chance of spread. So many diseases that we have to confront, that are common, will have an altogether different approach once you know that the person's at risk. If you know a person's going to have asthma in their life, prevent that they ever have a wheeze. I mean, you know, heart failure, no, they're never going to have heart failure. So you just have to know what the person is at risk for and then have something on the menu that works.

Yeah, I think it's true and I think One of the things from my perspective is this whole idea of I wrote a book called Ultra Prevention 20 years ago. Essentially, it was about this. It's like, well, we don't really have true prevention because we're just looking for things after they've occurred through a pap test or a mammogram or a colonoscopy or whatever we're doing. And we're entering an era where we're going to be able to see these subtle changes and then learn how to modify them and tweak them and through diet and lifestyle and some of the preventive things that we know work. in a much more personalized way. And I think one of the things about your work that I love is that you do talk about nutrition. Very few doctors talk about it. You're very focused on what that looks like. You recently published in November an article on your substack called Towards an Optimal Diet, looking at the role of diet nutrition and chronic disease. And for me, it's this best of times, worst of times moment. We have the best of times with the advances in technology, all the exciting things we've been talking about that are going to revolutionize medicine. But at the same time, we're also seeing this explosion of chronic disease and diet-related illness is the number one killer in the world now. And, you know, we're seeing increasing mental health disorders and all sorts of burden of chronic disease that are exploding at the same time we're seeing all these advances in medicine. So how do we reconcile these two? It's like how do we kind of use one to solve the other? Because I feel for me, this is the biggest problem that no one's really addressing is we have better healthcare than ever before, and we're getting sicker and dying younger than ever before. And our life expectancy is going down. So we got to do something about it.

Well, yeah, I mean, it isn't one simple thing, of course. There's lots of moving parts here. On the one hand, you know, as you well know, we have a monstrous problem with obesity and diabetes and diobesity, the twin pandemics. And even though we're starting to chip away a little bit of that with these GLP-1 family drugs. Yeah, we're talking about, yeah, you know, we, we Mandurro and then the triple receptors that will follow. But, you know, we're only at the, at the beginning of that. And who knows how many people ultimately will have their weight issues addressed. And of course, the toll of obesity on every system in the body is, is profound. Now, at the same time, we ignore our environment. Yesterday, my friend Sid Mukherjee had a great article on the carcinogens and air pollution. And I actually think that ultra-processed foods and a lot of other things that we take in our food stuff and beverages is potentially a liability for bad health, cancer, cardiovascular risk. We don't pay attention to it.

I'd remove the word potentially. I think the data is so clear on this ultra-processed food link.

Actually, I shouldn't have used that word. I'm glad we're in accord. Yeah. No, not potentially. The data is overwhelming, but nobody does anything about it. What are we doing to get rid of the carcinogens and the things that are accelerating major chronic diseases like you know, like heart disease and cancer and neurodegenerative diseases. So we've identified them, but we do nothing about it. And you remember, Mark, how long did it take to get trans fats to be, you know, outlawed?

But they're not even, I mean, even the government has ruled them as non-grass generally, they've not recognized it's safe to eat, but they're still in the grocery stores even after almost eight years later having that ruling, which is the food industry's lobbying efforts.

Yeah, so this is amazing, right? So we know some of the culprits, but we don't do, we have no teeth. We do nothing about it. Europe, they actually have done better on this than the US. So this is really disheartening.

Yeah, I think it's true. I mean, I think we're going to link in the show notes to this article in your newsletter called Ground Truths. I love this. Ground Truths is like, what are the fundamental laws of nature is sort of like how I think about it. What is sort of the indisputable truth? and, and I think we're getting closer to understanding that link between all the diseases that we're seeing exploding and the things that we have control over our diet, our lifestyle, our sleep, stress, environmental toxins, things that we actually can modify that doctors actually pay very little attention to.

Right. Yeah. Well, you know, just, just look at, you know, you're well aware of this. How come people in their 20s now are getting colon cancer? how come people in their 30s who never smoked are developing lung cancer, breast cancer? You know, it isn't their genes, you know, what is what's going on here? And so you have to really start looking at things like our environment, which includes, you know, what we're eating, what we're breathing, and, you know, again, there's like a blind eye to this stuff.

It's true, it's shocking to me that Fatima Stanford from Harvard, who's a professor is now on the US Dietary Guidelines Committee, said on 60 Minutes that obesity is genetic. And I think there are genes involved for sure that put you at risk or predispose you. But where were all the obese people when you and I were born? There was 5, 10% obesity, now it's 42%, 100 years ago it was nothing. And it's a It's kind of frightening to see that level of scientists who's now on a federal guidelines committee coming up with this concept and not talking about the real issue. And she's got funding from Mozambique and it's just like, oh my God, it's so corrupt. It reminds me and kind of where we are now in a way, I think it's a little pejorative for doctors, but I think it kind of reminds me of Voltaire's saying where he said, Doctors are men who prescribe medicines of which they know little to cure diseases of which they know less, and human beings of whom they know nothing. I think, you know, we're going to look back maybe in five or 10 years and go, what the heck were we doing? What were we thinking? Because, you know, I feel like it's such an, like, I'm just sort of sitting on the edge of my seat. It's such an exciting time. And I think if we can, if we can cross this threshold and we can get the paradigm to shift and we can use technology to do it, I think the world's going to look very different. I think our ability to treat disease is going to be very different. And what it was, so tell me in the, in that context, what are you most excited about in this? Is it just making healthcare improvements around how the process goes, around decision support, around healthcare costs, or is it around really a scientific paradigm shift?

Well, I mean, actually for me, we've touched on things that I think are really the big frontiers, not only including a virtual coach for people who want to use it to keep them healthy, and hospitals at home where you, unless you need to be in the ICU, you don't ever go near the hospital. because you can be monitored with multimodal AI. But to me, the far-reaching objective is to restore the patient-doctor relationship. Because I've watched this go down the tubes. I mean, back when I finished med school and started training, post-graduate training in the early 80s, the relationship with the doctor and patient was a precious thing. I mean, you knew the patient knew that you were there for them, that when you met with them, you know, you were there.

You were present.

Yeah. Yeah. And, you know, you had their back and they, it was an intimate, compassionate, empathetic trust relationship. What is it now? There's very little of that. I mean, you know, there are some exceptions. but there's not the time, you know, and that's why I, I really think this gift of time that AI can bring us whereby, you know, we give more autonomy to patients. So that decompresses the time of clinicians. I think this would be our, the most exciting thing to derive from the AI era. And we have ways to go because right now things are definitely going in the wrong direction and we have shortage clinicians, but ultimately, if we do this right, if we stand up to administrative overlords that tend to rule the roost, a lot of health systems and health practices around the country, then maybe we can get this on track.

Yeah, I think it's so important and it also speaks to the need to help those who are not having access to healthcare and medicine, who don't have experts in their area. I remember when I was a family doctor in Idaho in the in the middle of nowhere in a logging town. And, you know, we were it. There was, like, five family doctors. And so when you're on call and something came in and it was weird or strange, you had to deal with it. And so I remember, like, I would call up the doctors in Spokane, Washington, which was the closest sort of major Hospital. And I said, okay, what do I do? I got this baby. I said, premature birth. I got to put in this, you know, umbilical catheter, which how it was. And I would be coached through it. But, you know, I was a rural doctor back then, and there are now people. You know, globally, we're facing a doctor shortage, a nursing shortage, problems with accessing healthcare. And I think these technologies will actually help people in those areas and some of these health disparities and scale up medicine in a way that I think we haven't seen before.

I really agree with that. I mean, we've got to do something to address, you know, the inequities. And they include not just, you know, socioeconomic, but, you know, rural versus urban, like you're getting at. this is essentially software and it doesn't have to be a problem to accentuate or exacerbate the gap, but it hopefully could actually reduce the problem which is very, it's so severe that we have today. But, you know, I think this is something that we have an opportunity. We can see where this is headed if we really, you know, go after this. You know, right now there's this big controversy about accelerating, decelerating this, it seems to me, you know, we need to be accelerating for the most part because there's too much good here that can.

Be actualized, particularly in medicine, you know, I think, yeah, I think that's the key. And I think, you know, the doctor and the healthcare system is a bottleneck for people who are suffering and not just taking time to get an appointment or lack of access, but but most of the health issues we have, I believe, can be solved by self-care if we had the right information. Like you're mentioning, what are the foods that affect your blood sugar different than this person? And I think having a kind of AI-powered personal health co-pilot is going to be possible by ingesting this large phenomic data, all the stuff that I said at the beginning, all your lab tests, your wearables, your symptoms. and it goes into a system that'll create some intelligent guidance for how you can handle this on your own, what to eat, how to exercise better, simple practices you can do. And I found that as a functional medicine doctor, most of my patients can get better from things without me. If they just had the roadmap and could follow it and do the self-care practices, because most conditions that are chronic illnesses often can be solved through, you know, really simple basic lifestyle or other, other changes. And, and I think that's going to, in some ways, relieve pressure on the health care system. You can go to your doctor when you need to, but right now, it's almost like a guild. You have to go through the health care system to get your data. You don't own your data. You don't own, you can now, I wear an aura ring or a watch. You can get your biosensor data. But what about what's under the hood? You know, under, you can't. And, and so now I think we're going I co-founded this company called Function Health, which is designed to be an AI-powered personal health co-pilot where all this data is going to be ingested and you'll be guided on what to do for self-care and then when to go get medical care and what kind of medical care you might need. And then it'll help the doctors have a decision support tool. But it's happening outside the healthcare system because it's so hard to change the current paradigm. The Structure of Scientific Revolutions by Thomas Kuhn was the first discussion of this idea of paradigm shift. And he said, how hard it is to change normal science, essentially what people are doing now. But I think through sort of coming from the outside, from the consumer-driven aspect of healthcare, I think we're going to see a greater acceleration of this. And I'd love to hear your thoughts on that movement and that effort.

Yeah, no, I think this is really interesting what you're bringing up. So firstly, there are now 20 or 30 companies, mostly startups, that are offering coaching chatbot with humans in the loop as a backup for most chronic conditions like hypertension, depression, diabetes, you know. So they are, some of them have done randomized trials to show they're really helping people. So, you know, we were talking about unimodal, so it's narrow, they're helping one condition. The next phase is what you are getting at, which is a virtual coach across all conditions to prevent them. And I don't see theoretically why that can't be done. It has to be proven that it's working. Now, one of the things that helped me in recent days to see, well, this is inevitable. There was a paper by the Stanford Group in Nature about organ specific clocks. it was to cover last week. And, you- like aging clocks, you mean?

Or aging clock.

Aging clock. So you know how you can get so-called Horvath or epigenetic clock to say, what is your biological age? And hopefully it's similar or better than your chronological age.

Well, mine was 43 and I'm 64, so I'm going with that.

Oh, okay. There you go.

Good for you.

Wow. Wow. Anyway, that isn't that helpful compared, I think, with organ-specific age. So what they discovered, which is using AI.

Yeah, my back is about 110, but...

Right, but yeah, you didn't need the proteins to know that. But what they discovered, which was amazing, they validated through three different populations, was that there's signatures and proteins that can be assayed in each of these organs to know for you, what are you at risk for? Are you at risk for your kidney, your liver, your heart? And this is just yet another layer of data which we didn't have before to add to the mix. So, you know, I do envision that this virtual coaching with periodic, whether it's retinal assessment or organ specific clocks or you name it. And what they showed in that study in the report in Nature was the proteins correlated with what happened to people. That is, people with accelerated aging of their heart had heart attacks and heart failure. People with brain aging, of course, went on to Alzheimer's. And so this is yet another way to anticipate an individual. You won't get that from an overall body clock, but you'll get it from organ clocks. And I think that to me was really one of many important in a chain of recent discoveries.

Well, that speaks to the real issue here, right? Which is what are we looking at as doctors? We're looking at so basically things with a magnifying glass, not a microscope. And now we have an electron microscope, you know, to even look deeper. And I, I think, I think, you know, there's tens and tens of thousands of molecules and proteins and metabolites and, you know, even peptides and, and things that are floating around in your blood, many of them even from your microbiome. Stan Hazen said he thinks about a third of all the metabolites in your blood are from your microbiome, and we look at that. If you do an extensive panel of tests, you might even get 50, 100 blood tests, but that's kind of minuscule when it comes to what's really going on. And so what you're talking about really is on this precipice of being able to look at all these analytes from all the different sources and be able to track what's going on in your body and use that information to optimize your health and to do research in a different way. So, just the last few minutes we have, I'd love to sort of have you kind of pontificate on how we are going to reimagine research in this era of AI enabled medicine?

Well, I mean, I think the research has to be more to get a change in practice. So another really good example of the same, you know, major discovery, the Icelandic group publishes that when people in Iceland had their genomes assessed, they found that, you know, four percent could have had their lifespan expand extended up to seven years if they had known the information about, like, for example, BRCA2 mutation recently published in the New England Journal. So we, the, what I'm getting at, Mark, is we have abundant medical research that sits in these high tier journals goes nowhere. Never goes into patients. Never is proven in real patients in the real world that it changes their health, it promotes their health. That's what we need now to basically get this into high gear. We have too much stuff that's stuck in the I think you recognize there's paternalism in medicine. And patients don't tell them about their genomic data because they won't be able to handle the truth.

It's a little patronizing. It's a little patronizing.

It's their data. If they want to get their data and they want to know what caveats there are and so-called variants of unknown significance, that's their right, but we don't let them have their data. So we have to stop this and that's holding us back.

Yeah, it's true. And I think I just want to point out that you're the founder and director of the Scripps Research Translational Institute, which is really designed to translate from the bench to the clinic research that's going to change people's lives. And we're so slow in medicine. I mean, everybody knows the story of Semmelweis who sort of figured out that we should wash our hands before a child, you know, attending to births. And he saw that the midwives patients were not dying of childbirth fever and all the doctors patients were. And he said, you know, guys, we should probably wash our hands. They're washing their hands. It might, maybe it's gonna help. And they were like, God, how could you even suggest that doctors would be causing any harm to their patients? You're excommunicated and he died in, you know, in disgrace. And it took 50 years to figure out that we should wash our hands before surgery, you know? I know. It's kind of like that right now the advances are happening so rapidly in medical science. And my daughter's in medical school now and I'm just so embarrassed. I said, Rachel, are you learning about the microbiome? Did you learn about insulin resistance? Did you learn about nutrition? Well, we learned about amino acids and fatty acids. I'm like, what are you gonna tell your patient to eat? I'm like, what is happening out there? So I'm just so excited by this moment because I I think it's going to be a way to disrupt healthcare and that we haven't been able to do yet. And I'm so excited that you've been thinking about this, leading this work. You've done so much in your career that's advanced science and thinking so much. And it really makes me very inspired to know people like you who are at the peak of healthcare and understand these concepts and are trying to push medicine forward into the place where it can really be what it's designed to be, a profession that really is there for helping and healing and transforming lives and not just a big medical industrial complex.

Oh, thank you. Thanks, Mark. That means a lot. And, you know, healing, caring, those are the, those are the key. We got to get to that and do it better and, and, you know, leverage what we can out of AI and machines to help us. get there.

Yeah. What was that quote? The secret of caring for the patient.

The care of the patient is caring for the patient. Peabody, 1927. Yeah, my favorite quote, actually.

Yeah, that's right. That's exactly true. And it's such a beautiful sentiment. I think this revolution in medicine and AI and systems biology and all the things we're working on, I think are going to allow us to bring humanists back into medicine, which seems almost paradoxical given given that it's technology, but I think it's going to allow a much more humane approach to our patients. And, I can't wait to have keyboard liberation on it.

Yeah, it's coming. It's going to happen fast, you'll see.

All right. Any final thoughts, Eric, on where we're headed and what you're inspired about next?

Well, no, I'm just excited. I was talking to my fellow recently about how he's going to really see it because it's going to take some years. Every time I forecast something, I realize that my impatience tends to accelerate when we're going to really get there. But we will eventually. It will be a far better way that we practice medicine and patients will derive far more benefit. They'll have renewed faith in physicians and the relationship they have with them. And this is really a unique time to look forward to. I hope we'll get there. The sooner the Well, I think, you.

Know, both of us obviously take care of our health and are focused on our health. So hopefully we're going to extend our life long enough. We'll be alive when all this happens.

I hope so.

Well, look forward to having more chats with you and following this along the way. And thanks for all your work. If you love that last video, you're gonna love the next one. Check it out here. If we trace back every experience, every.

Perception, every sensation, every image, every thought.

To its source, you'll end up with the quantum field.