Agentic AI and Longevity - Key topics at NextMed Health 2025

[00:00:00] Dear listeners, welcome to Faces of Digital Health, a podcast about digital health and how healthcare systems around the world adopt technology with me, Tjasa Zajc.

[00:00:19] In today's episode, you will hear a short reflection and a recap of NextMed Health, which took place in San Diego in the United States. NextMed Health is a conference that looks at what are the good practices in healthcare technology today, what's needed and what we can expect in the future of healthcare.

[00:00:44] It's a four-day intense course in everything from neurotech, AI, longevity and more. So in today's episode, you will hear from me together with Gary Monk, who is a digital health consultant and a thought leader on digital health.

[00:01:07] And additional interviews I did with Anthony Cheng, founder of AI Med and Chief Intelligence and Innovation Officer at Children's Hospital of Orange County. Daniel Craft, founder and chair of NextMed Health. Beo Curry-Winchel, founder of Beyond Clinical Walls. She's also an urgent care medical director at St. Mary's Health Network.

[00:01:34] Steven Brown, AI developer, coder, founder, investor and filmmaker. Jennifer Garrison, professor at Buck Institute and co-founder and director of ProductiveHealth.org. Eric Topol, founder and director of Scripps Research Translational Institute. Anthony Cheng talked about the current state of AI in healthcare and the needed changes for adoption.

[00:02:02] Beo Curry-Winchel talked about bias in healthcare. I spoke with Steven Brown about his personal projects on building agentic AI. Jennifer Garrison described the current state of women's health and the knowledge that we currently have about ovarian health.

[00:02:24] And I took some of the segments of the discussion Daniel Craft had with Eric Topol, specifically the parts around longevity, the industry, and what can we take into account for our own improved well-being already today. Enjoy the show. This episode is also available in a video format. So do go to YouTube and search for Faces of Digital Health.

[00:02:52] The link is also in the show notes and see the video episode as well, if that's your preferred way of consuming content. And if you will enjoy the show, make sure to subscribe to the podcast or leave a rating or a review wherever you listen to your shows. I will really appreciate this as well as your suggestions for what else we could cover and how we can improve further. Thank you. Now let's dive in.

[00:03:21] Dear viewers, dear listeners, welcome to a very special episode of Faces of Digital Health. We're in San Diego at NextMed Health, a four-day event. And Gary Monk here, it's a great pleasure to be doing this with you here.

[00:03:50] So let's reflect on what's been going on in the last three days, starting with day one, when the focus was on AI. What were your key takeaways? Yeah, great. Great to see Anthony Chang. And I think hard to sum it up, but the key thing is how AI has been moving, is moving from more of a, I guess like an expert source of reference to, I guess they call it agentic,

[00:04:15] that it's actually going to be able to have so much power to be multiple different consultants, specialties, to support patients, to support doctors. So it's going to, we're going to see that level of control and power increase rapidly. And I think we're seeing that already. We heard a brilliant, not a brilliant, but a very mind-blowing story from... Mind-blowing sounds brilliant.

[00:04:41] Steven Brown, who was diagnosed with multiple myeloma, but very late. So basically the doctors did all the tests they could. Didn't find anything a month after he appeared in the ER because of abdominal pain. And that's when they basically discovered what his real problem was.

[00:05:07] And he then started thinking, would the AI discover this faster? And he built this whole agentic system. And I did a short interview with him. So let's hear that now. So I had been having some abdominal discomfort and I lost like 20 pounds and felt like something was going on.

[00:05:28] So I went to the doctor to get everything tested, full body scan, colonoscopy, endoscopy, and every blood test they would allow me to check the boxes for, or that they would check the boxes for, just because I wanted to find out. And the doctors thought it was a fishing expedition and couldn't find anything. So in the end, we were in...

[00:05:56] Our house was burned down in the Palisades fire in Los Angeles and we were displaced from LA and from our doctors. And I ended up in the Palm Springs area and ended up in the emergency room for what I thought was something completely unrelated. And I got some additional tests and they sort of pieced things together and got to a diagnosis of a plasma cell disorder related to multiple myeloma. What did you do then?

[00:06:26] Like, how did that make you feel? Well, that was on one hand, it's... On one hand, it's kind of a relief to actually have an answer. It wasn't really the answer I wanted to have. And then I was admitted to the hospital for... So they could do more biopsies and tests and sort of figure it out. And while I was in the hospital for, I guess, nine days, I started using AI. I've done a lot of work in AI.

[00:06:55] I've developed a lot of applications with AI and... I mean, I've developed a lot of AI applications in other fields and education and entertainment. And so I had a lot of experience in there. So I've been working with AI. So the first thing I went to is... You know, my question was, well, you just figured out this now, but I got tested for all this kind of stuff just three or four weeks ago. So when would the AI have figured this out?

[00:07:23] So I started testing on my own data and kind of working with AI, but testing on my own data and realizing that the AI would have figured it out earlier where my doctors didn't. And the AI can see patterns that people have a hard time seeing because we know what we look for, we know what we're specialized in, and we have a fairly narrow field of view.

[00:07:52] Even when we think we know a lot, the medical field is very large and complex. And it's really not possible for one human brain to capture all of that. So it's very difficult. But AI has read all of the literature, and it has a much broader perspective. Mm-hmm. And which tools did you use? So how did you go about this?

[00:08:17] You know, many patients in the room or just like all the participants started wondering, like, where can I get this tool? Because you also demonstrated your solution. Right. But did you start with CHET-GPT? Like, how did you design everything? Well, initially, the exploration was just how good is the medical knowledge in the existing models, in the latest anthropic model, in the latest OpenAI model, in the latest Gemini model.

[00:08:45] Just how deep does the knowledge go? So the first exploration with almost everything I do is like, what does it already know? But I'm building applications on top of that. Because it's one thing to just, you know, have a conversation with CHET-GPT, which is kind of designed to be a little bit different every time that you have a conversation. It's designed to be, to have some kind of unpredictability and variability so that it's interesting.

[00:09:13] But, you know, if you're developing an application, you really need to kind of control and manage the whole process. So once I saw that the knowledge was in there, and in there in a way that gave me confidence that it was solid. I mean, you can ask AI questions and it will make something up. I mean, the most human thing that AI does is make stuff up. You know, when asked a question, it will always give you an answer. That doesn't mean it's necessarily correct.

[00:09:41] And that's just inherent in the design of just how a neural network works. It's going to give you an answer. But it's, the question is, can you trust it? So the first exploration is to probe the models for the knowledge and kind of QA, you know, as much as you can. See, can I trust the knowledge that's in there? How deep does it go? You know, what does it know? What does it not know? What are the limits? What has it been trained on?

[00:10:10] So once I saw that it had been trained on the things that were, the first check was how well is it trained on medicine? And it's pretty clear that the leading models have already been trained on lots and lots of medical literature and information. So you can see that the knowledge is in there.

[00:10:33] The question then is, how can you, I'd say, reliably get the right knowledge out based on what your situation is, what the context is? So then I started kind of cross-referencing with my own data and some of the own particulars and digging deeper and deeper and deeper. And once I saw that it was, it was, it gave me confidence that the training is pretty solid and it's getting better every day. So I started building an application on top of it.

[00:11:01] And the application on top of it is partly to, to, I don't say guardrail, but, but to make sure that it stays within the task that you're seeking, that it kind of doesn't wander out of the, what the kind of the area, the zone of the model that you're, that you're trying to probe. And then also personalize it with your own questions and with your own data. So what surprised you, you know, when you were building this out, did you test different large language models?

[00:11:30] Um, how did you decide that what you built is good enough? How much refinements did you have to make, you know, even in the beginning, when you said that you cleaned the data? Uh, did you mean that you cleaned your, uh, the data, you know, from your medical records and how did you know how you, what you have to do, given that you, you know, you personally do not have a medical background?

[00:11:55] Well, it's, you can export your data from my chart and their PDFs and their PDFs with a lot of other information in there and, you know, tables and different kinds of formats. And the question is, is it understanding what's in the PDF?

[00:12:10] So when I say cleaning the data, it's really more a matter of making sure that when you convert that PDF to plain text information, that it's, that it, it's a one to 100% match, that it didn't mix something up or didn't miss something. So I just, it's more a matter of understanding the data in the chart and making sure that it understood that it understood that it accurately understood what was in there.

[00:12:36] Cause sometimes it's like, it looks some, some of the things look like a, you know, faxed in report. Well, you know, so it's, it's like, you need to optical character recognition. You need to make sure that, you know, it's what comes out is structured data that is reliable, that represents what actually happens. So, you know, in the end, it's a bunch of medical test data and it's a bunch of imaging reads of imaging. And then there's, you know, sort of the patient's symptoms and kind of chief complaint.

[00:13:07] So it's, it's all the same stuff that the, the, the, the doctor is looking at. I'm just want to make sure that it's translated into a kind of a plain text format that the AI is going to understand. Cause the AI is operating on plain text, even when you upload a document or a PDF or an image, it's converting that into a plain text interpretation. So you need to make sure that it captured the right data. So that's, that's partly, you know, make sure that your inputs are correct.

[00:13:35] And as far as which model I'm in, everything that I've been doing in education and entertainment, the model is interchangeable. I mean, I basically, it's a setting you can choose any one of a dozen different models. So you choose. And the reason I do that is because the models are always changing. It was GPT-40 and then it's GPT-4.5. It's Gemini, Gemini 1.5 and now Gemini 2.5. And these are getting better every day. And, and sometimes one is better than the other. And then somebody else comes out and it's better.

[00:14:04] And so I'm, I'm actually testing the models against each other. So I ran an analysis where I, on stage, where I had multiple different doctors, which are basically multiple different perspectives, multiple different routes of navigating through the knowledge that's in the, in the large language model. But I also can run the analysis of like, hey, same doctor, but you know, seven different models and see what, what comes out.

[00:14:32] So part of that is, it also answers your question to like, you know, when does this get released? Right now, this is, this is an R and D phase because I am testing model versus model. Navigation strategy or specialist versus specialist. And, you know, so there's a lot of, still a lot of more testing. And I'm happy to do that on my own data or with volunteers who understand where this is at.

[00:14:57] But, you know, this needs to be something that we can count on for our own kind of decision support, you know, with the right caveats and the right understanding that this isn't perfect. These are generating ideas. It's generating ideas that maybe you should talk to your doctor about. So, so there's a lot of that. It's helping educate you, helping you learn the language related to your condition, helping you know what things mean and helping you talk to your doctor. Do you use AI, Gary? Have you ever considered using it for, you know, health purposes?

[00:15:27] For health purposes? Yeah, I, I, I have used it in terms of looking up specific symptoms, trying to stay well, what sort of things are going on, what could they be caused by. I think I did a blog post about some side effects I got, which I thought from a medication without going into a lot of detail here. So muscular pain was it to do with sport medication. So I put that into AI.

[00:15:49] But I think what, what, what for me was really powerful was this opportunity to put all my health data, health records, even if now I think it was Eric Topol said it's things aren't centralized, being able to throw in whether it's PDFs from my medical records, etc. And then get more of a holistic view of my, my health and all previous, I guess, longitudinal tests and blood tests and what that could mean in the future. So I think that's, that's super fun. Rather than just taking this moment in time, what's wrong with me?

[00:16:17] What can I optimize to looking at my whole sort of history and how that might affect the future? Yeah, that's the personal side. As we mentioned earlier, Anthony Chang had a three hour workshop on AI, which focused basically on how healthcare systems can adopt technology or the challenge with the lack of knowledge among physicians and the challenge with upskilling.

[00:16:45] So I, let's, let's hear it from Anthony Chang on the key insights from the workshop and why he is more concerned about AI not being used effectively rather than the dangers of AI. Mr. Chang, you just had a really engaging session on AI in healthcare.

[00:17:09] And I think one of the most, the strongest points that you made is that it's soon going to be unethical to not use AI in healthcare. Can you maybe expand on that a little bit? Well, of course, that statement has to be interpreted correctly.

[00:17:26] I just mean in certain situations, which there are cases that have used AI for quite some time and proven to be effective, that we are reaching the point in which the standard of care mandates that you use AI in a timely fashion to make the correct diagnosis. Radiology is probably the prime example.

[00:17:48] The one that I'm seeing some activity in terms of people holding the hospital accountable would be radiology, particularly brain scans and probably cardiac scans as well. Mm-hmm. Another thing that you also mentioned is that, you know, sometimes you consult hospitals around their AI approach. And it's not that rare that if you ask, can you show me your AI strategy, that the strategy actually isn't there.

[00:18:16] So what's your advice to hospital leaders in terms of how to approach that? What if they already have a data strategy and a data government strategy? What's the key thing that's usually missing? And what's your advice on where to even begin? Yeah, my comment was that oftentimes when we get asked to talk to a hospital about their AI strategy, they don't even have a data strategy or data governance yet. So I think you start with a foundation, which is the data governance.

[00:18:44] And then you build on top of that an AI strategy or governance and also a committee. So it's something I said during the talk, which is AI agenda in healthcare has to be driven by human-to-human interactions and relationships. Mm-hmm. And also that change is not technology-driven, it's human-driven. I thought that was one of the better ones as well. I think it's one of the beautiful things about AI in healthcare is the technology is pretty robust and mature right now.

[00:19:14] But it does take humans to learn about it, to adopt it, and to use it effectively. Mm-hmm. Where do you see the biggest danger of AI at the moment? One of the participants mentioned the fear of, you know, the data sets that are not diverse enough, which doesn't seem that it's going to improve in the future. So her call to action was for people to gather as much data as possible.

[00:19:40] So how do you see that development and, yeah, basically the biggest challenges that AI is going to face before it's more broadly distributed? Well, I think AI is going to hold us accountable to make sure that the data sets are balanced and there are technological workarounds that we can do, like use of synthetic data that's affected.

[00:20:01] I'm actually much more concerned about the mature AI tools not being used efficiently and effectively to help mitigate the burden that caretakers have, as well as the disease burden that we have in general. Where do you see the reasons why the tools are not used effectively? And which reminds me that you mentioned that basically out of the 4.2 trillion U.S. dollars that's used in healthcare,

[00:20:31] basically 20, 25% could be reduced with, you know, the use of AI in administrative tasks. So why is this not happening? Well, I think one reason is that not enough people are educated in this area, which we're trying to solve here partly. And the other reason, and there are many reasons, but one of the other reasons is that the financial incentives are not aligned between the payers and the providers and the hospitals and the patients.

[00:20:59] So it's, you know, having an AI as an elegant solution can't solve the human-driven problems, which is alignment of the financial incentives. Is there anything that you think that people should currently know about agentic AI, which was discussed at the end of the workshop? And you mentioned that it's often confused with AI agents. Yeah.

[00:21:26] It's an easy thing to misunderstand or confuse, be confused about. I look at agentic AI as a very capable captain of a team on the field. There's quite a bit of autonomy with the coach's direction, by the way, which is the general AI part.

[00:21:44] So the agentic AI would be the leader on the field with autonomy and the individual players could be, we could draw the analogy that they're the AI agents, that they do things autonomously, but with the overall autonomy of the agentic system. So we've got all this powerful data. Dr. Baio mentioned bias is such a key, key issue. And it kind of got me thinking that Biobank was mentioned as such a powerful source, maybe by Eric Topol yesterday. Am I allowed to say that?

[00:22:14] Yes, we're not there yet. It's connected. It's all connected. It's all connected. Okay. Should I start again? Yeah, let's just start again. See, I think we should mention bias because Dr. Baio talked about bias being really endemic within the health system. And we heard later in the conference about Biobank, I think Eric Topol mentioned it, that there's such a powerful resource, yet it's 95% white people.

[00:22:44] So a great resource, but I've got to say, it sounds inherently biased to me as well. It seems like it's a great resource that's actually fueling bias. Yeah.

[00:22:55] It was a heartbreaking story of Dr. Baio, who almost died in her own hospital during labor because her pain was dismissed as a black woman, despite the fact that she's a medical director in that institution. But she's a ray of sunshine, I think, so a very optimistic person.

[00:23:24] So despite that experience, here's some of her things that she believes that we can all do when we try to advocate for ourselves in the healthcare system and when we try to contribute to basically decreasing the bias. Okay. And bias is something that we all have. And in my situation, it really is reflective that what happened to me can happen to anyone.

[00:23:49] And when we look at the statistic that black women are dying at the highest rate during childbirth, I'm reflective of that statistic. And I had to advocate for myself. And what happened was there were multiple areas of dismissal that happened, not only based on I'm a woman, I'm a woman of color. And the fact that although I said something is wrong, I still wasn't heard.

[00:24:18] So it's my race. It's my gender. It's just all of the things that encompass who I am that unfortunately almost cost me my life, even when I was advocating for myself. And I share my full story on my TEDx title, Do No Harm. So I encourage everyone to take a look at that talk because it shares my story. It shares what happened.

[00:24:44] And it also helps you know more about what you can do if you face a similar situation. And that was my goal is if I can save one life through sharing what happened to me. That's what it's about, because in the end, yes, I'm a physician. But at that moment, I was a patient and I faced something that many women face giving birth in this country, especially black women.

[00:25:13] If we expand on that a little bit, you know, when you're a patient, you are in a bit of a subordinate position. It's a vulnerable position in the power dynamics. You're the vulnerable party and it might be difficult to advocate for oneself. So any tips, any advice on how people can do that? Yes. So the three tips that I love is the three C's. Be clear, concise, and confident.

[00:25:41] Be clear about your situation and what's happening. Be concise, meaning when you see the doctor and you feel dismissed or you're trying to share your symptoms, instead of saying, well, I've been feeling like this for a couple of weeks. Say, I've been feeling like this for one week and it is impacting my quality of life. If specifically, I'm not able to get out of bed. I'm not able to spend time with my kids and then be confident in the fact that you know your body best.

[00:26:11] So those three C's can really help you every time you go to the doctor and just being clear, concise, and confident. And those are things that you can carry with you. Everything we've mentioned so far, Gary, was just day one. Really intense program. Day two was, to me, very focused on lifespan, healthspan, women's health. What did you get out of that?

[00:26:41] Yeah, it was, I felt the longevity came out. And I think it was Maggie, Maddie, I can't pronounce her surname, Dietwald, talked a lot about longevity from a female perspective. But I think there were lessons that we could all learn in terms of healthy lifestyle, healthy aging, extending our lifespan. Key takeout for me was the fact that often it feels like we're getting older on this conveyor belt towards sickness,

[00:27:08] things we can't avoid, whereas actually we have a lot of power. There's a lot of opportunities to stay well, to stay healthy, to avoid some of the pitfalls of aging and have that longer, healthier lifespan. So I felt that was very, very empowering and positive. A lot of focus was also on women's health, which is a field that's very well under-researched. So I spoke with Jennifer Garrison, who also presented on the ovarian health.

[00:27:38] And let's hear why that is important and why we should not consider ovaries just as reproductive organs anymore. My talk was about reframing women's health through the lens of ovarian function. And the reason for that is that in addition to being absolutely essential for reproduction, ovaries are sitting at the center of a complex signaling network that's kind of like Wi-Fi.

[00:28:08] And they're talking to a lot of different tissues in the female body. And what they're doing is promoting health. We don't understand a lot about those axes of communication, but it's, you know, you can think of it like a symphony. You're probably at least mildly familiar with the idea that your menstrual cycle is set up by a conversation that happens between your brain, your pituitary, and your ovaries, right?

[00:28:32] So imagine that there's a conversation like that that happens between your ovaries and your skin, or your ovaries and your bone, or your ovaries in pick an organ. And while we don't understand very much about the details of those chemical conversations, they really are important for overall health, no matter what your age is.

[00:28:53] So beyond fertility, beyond menopause, which are obviously two very important things that happen that are related to ovarian function or dysfunction. No matter what your age is, your ovaries and how they're working is really important for your overall health. Mm-hmm. How did these findings come about?

[00:29:15] And what that makes me wonder is which other organs or, you know, body parts are also as mysterious and we don't potentially know much about when it comes to women's health? Well, I guess we really don't know much about female physiology. Full stop. Like, forget about ovaries. We just don't understand how female bodies work.

[00:29:39] Um, and what we can infer about ovarian function is really what happens when either they're missing, right, when they're either removed through surgery for whatever reason, or when they're not working properly. And, um, that basically gives us a sense for what they're doing normally. So we don't have a lot of studies. This is a place where we have a huge data gap. And that's part of why I was talking about the need for more research on women's health. Mm-hmm.

[00:30:09] So how does research look like at the moment? Um, funding, as you mentioned, is a crucial problem. So what can you then do? Uh, I'm almost afraid to ask, you know, what do you anticipate in terms of what's going to be possible to research, especially in the U.S. with the current, um, you know, changes that are happening in terms of the support to science? Yeah, I mean, we're, uh, we're at a crisis moment right now for all of biomedical research.

[00:30:38] Um, I think, you know, women's health, the word woman, the word female, those are on the list of words that have been used to, uh, revoke funding, pull grants, you know, this, um, censorship that's happening, which is absolutely ridiculous. Um, I think at the end of the day, science shouldn't be political, right? It shouldn't be political.

[00:31:02] You know, I'm, as a woman, as a scientist, um, I believe that knowledge saves lives. And biomedical research, when we politicize, um, science, when we, um, weaponize research funding for political purposes rather than for the public good, that's bad for everybody.

[00:31:25] And so, you know, I think we're just at a crucial moment where we have to decide, are we going to choose knowledge or ignorance? And, you know, that's, uh, right now, I think, um, the state of science funding is really under attack in the U.S. So we're either going to lose a lot of really great scientists to other countries, um, or hopefully something will turn around soon.

[00:31:52] Based on the last eight years that you have been researching, uh, you know, women's health, um, what are you observing? Uh, what kind of progress do you see? Um, maybe even what surprised you during this time? Yeah, uh, it's been great. I mean, we started, um, we started the global consortium five years ago. We started the center at the Buck Institute six years ago. And, um, the progress has been really amazing.

[00:32:21] Like, even better than I could have hoped in terms of just the number of researchers and clinicians who've come into the field. And the progress that's been made, uh, so many studies have been published now around ovarian aging. It's, you know, it went from like an, a niche area where there were maybe a dozen people in the world who were really focused on ovarian function outside of reproduction and menopause. Um, to now there are hundreds of people, right?

[00:32:49] We have the fourth annual reproductive aging conference coming up, which is the only international meeting, um, in the space. And, you know, we anticipate there will be a hundred people or more there. And so there's a field now where there wasn't one before. So there's a critical mass of people who are devoting their attention to this problem.

[00:33:10] And, um, the fact that we're working on developing this X prize around, uh, women's health and ovarian function, that's going to bring a whole different group of thinkers and innovators to focus on this problem. And so hopefully, you know, we can, we can continue to accelerate progress. Yeah. Good. Um, Gary, and we covered a lot already. Yes, a lot.

[00:33:40] We haven't yet shared with the audience what Eric Topol was talking about on stage. Uh, the things that kind of stood out most to me was, were the things around longevity and what actually works today and what doesn't work. Um, so let's hear it from him directly from the stage.

[00:34:04] As everyone knows, there's been intense interest in anti-aging intervention, you know, from cellular, uh, uh, reprogramming, partial epigenetic programming to, you know, simulitis, genealogies, long list. Uh, that's interesting, but the real, uh, science of aging advances have not been from those interventions, which all have some serious potential side effects and don't have any human data yet that show that they really were.

[00:34:34] But the real is the ability to, uh, determine our, the clocks and metrics of aging. And, uh, like I mentioned, the organ clocks, which is amazing work. It's come out of Stanford. Tony, it was Cori and now replicated through the UK Biobank in several groups that each of us at any point could have a tube of blood with up to 11,000 proteins in the plasma and say, this is your organ that's out of kilter, aging faster than everything else that's good.

[00:35:04] You have that with, you know, genomics and then you add on other layers of data like the epigenetic clock of Horvilla. Then you start to say, hmm, we have so much data on any given person. We can partition risk. And the three major age-related diseases are Alzheimer's neurodegenerative and cancer. Like we've never done before. It used to be, you'd say, in fact, uh, 23andMe really pushed for the polygenetic risk clock.

[00:35:34] Uh, and the problem with the polygenetic risk score is, okay, you have a risk for, let's say Alzheimer's. But when? Is it age 98 or is it age 16? It really didn't help that much. Now you can pinpoint percent of a person's risk and it's how high and also say when. And all three of the diseases I've mentioned are diseases that take 20 years to incubate and finally clinically manifest.

[00:36:04] So if we can't work with a 20-year run run with the kind of layers of data that we have now that are going to become widely available and inexpensive, with the factors that we know make it difference, then, you know, we are missing perhaps the greatest opportunity. Now, the key point is that you can't use all this data if you didn't have AI, the models of today, including those with reasoning.

[00:36:30] Because that's what really cracks the case is when you go for all a person's data, and that includes things like set points in labs, whereby the labs are normal, but the AI picks up the trend along with all the other things in that person are going and are pointing to something. So there's, the data is so rich that we have or will have, and if we can't use that to prevent the big major age-related diseases, then there's something wrong with us.

[00:37:00] So that's why I'm so excited. The key is that it's not just that you are at risk for one of these three major diseases, but it's pinboarding, you know, when and which and, you know, the details. And, you know, there's been breakthroughs in each of those three diseases. The P-tau-219 and other plasma markers that are as good as a PET scan for hemorrhoid buildup and tau-14 aggregates. That's a huge difference.

[00:37:27] We never had anything like that for picking up risks. And when you have two of those, you can define the art of potential neurodegenerative disease. And, of course, what's great is it responds to exercise and lifestyle factors. So, you know, you get it. You know you have a potential burden integrated with all your other data, and you can take action. The same thing for cancer. We didn't have wanting cancer early detection tests. That were worth much. Now we do.

[00:37:57] But they're being used in the wild people. They're being used in anybody over age 50, which is completely stupid. But what they're used for people who are at high risk for cancer as a screening, you know, on a periodic basis, that's when we nip this at the microscopic stage, and we shouldn't be doing total body MRIs. So another reason to do the book was, you know, Peter Atiyah's album, Liz's book, was very popular. How many of you have read that book yet? Yeah. Yeah.

[00:38:25] One of the most popular books in this space for many years. And many things in there were just absolutely wrong. It's a good part. Really good on metabolic syndrome and, you know, that. But he advocated taking reprimicin. He advocated everyone should have total body MRI and many other things. You see, it's a company biograph. And by the way, any person or company that's selling supplements is off the list for me. Okay?

[00:38:53] Because none of these supplements have shown benefit from Tulsa. Thank you. Thank you. That aligns with a sort of dark belly of longevity, right? All this sort of snake oil, in a sense. Go to sort of countries and get our stem cell out there. Yeah. I'm a stem cell biologist and bone-like transporter. We've done stem cell transplants for a long time. And there may be some applications. How can it help, I mean, in your book, to address some of this? Sift through that noise and find ones that are real. Yeah.

[00:39:22] Well, I mean, it's funny you mentioned snake oil. The most extreme form of this is this Brian Johnson guy. And he even has all these problems. Longevity mix. He is a prodigal snake oil. He's incredible. I can't make this stuff up. I mean, yeah. So how do you... I recently did a subset on 12 companies that are basically selling longevity at HealthSpring, right?

[00:39:50] One of them, Function Health, Mark Hyden. I mean, he's a very bright fellow. And the Function Health is valued at $2.6 billion. For what? They basically are testing several hundred lab tests. And giving you, you know, some counsel on that. The point about these companies, they have the right idea. We do need more data in individuals. But they're not even getting the right data.

[00:40:18] So I do think that's where we're headed is deep data and using AI. It's funny. If you look at some of these companies, they say they're using AI. But they're not. I mean, not the AI that we know of. So we're moving in the right direction. Drafts jumping the gun in some of these companies. I think they have good intentions, generally. But they're not up to speed. Slowly, slowly. I'm going to have to go to the airport. So we're going to have to wrap up.

[00:40:44] To conclude our whole, you know, reflection on NextNet, is there anything else that you would like to add? You just heard from Ray Cartel on day four. What do you remember from his thought? A couple of things really stood out. One was this idea, I think you said, by 2035, we'll be able to do all drug research computationally. No need for, I guess, animal testing, even human testing,

[00:41:13] all be able to be done in silico. And the fact that we're getting close to artificial general intelligence and the implication that's going to have for medicine. So, yeah, a very, very profound, profound talk. I'm definitely going to summarize that, but yeah. Yes. So thank you for being on this journey with us. And for more insights, we'll do Darius Group. Yeah, I'll write more of this stuff up on LinkedIn.

[00:41:43] So yeah, make sure to follow me. I'll do some more. I'll write about Ray, some of the other takeouts and trends that have come out of NextMed. And just before we conclude, I think it's fair to say congrats to Daniel, Shauna and Tim for again creating a magnificent event that is memorable, full of new knowledge and is definitely going to keep us thinking. Yeah, for sure. Without them wouldn't be possible. And so much work's gone into it.

[00:42:13] So it's been an amazing conference. You've been listening to Faces of Digital Health, a proud member of the Health Podcast Network. If you enjoyed the show, do leave a rating or a review wherever you get your podcast, subscribe to the show or follow us on LinkedIn. Additionally, check out our newsletter. You can find it at fodh.substack.com. That's fodh.substack.com.

[00:42:42] Stay tuned.