AI: How it Works + AI in Medicine with Harper Carroll
Today I'm speaking with Harper Carroll, an AI and ML software engineer, on the Longevity Optimization Podcast. In this episode, we explore the fascinating intersection of artificial intelligence and health optimization. Harper shares insights into the concept of AI as the ultimate polymath, detailing her journey into the field and the transformative potential of AI in revolutionizing health diagnostics and treatments, particularly in cancer care. The conversation delves into the implications of AI in personalized medicine and how technology can shape the future of health optimization.
Harper Carroll is a forward-thinking engineer with a deep passion for leveraging artificial intelligence to enhance health outcomes. His expertise lies in developing innovative solutions that address the complexities of personalized care, particularly in women's health. Throughout the discussion, he emphasizes the limitations of traditional medical practices and highlights the transformative potential of AI in optimizing health and creativity. Harper also discusses the role of AI in enhancing empathy through virtual reality experiences, the implications of AI hallucinations, and the environmental impact of AI technologies. The conversation concludes with a forward-looking perspective on energy solutions, particularly in relation to nuclear and solar power, underscoring the need for sustainable advancements in technology.
Follow her on Instagram : https://www.instagram.com/harpercarrollai/
Timestamps
00:00 Introduction to AI and Longevity
05:08 The Polymath Concept in AI
09:58 Harper's Journey into AI
14:47 Understanding AI: Definitions and Applications
19:56 AI in Health: Opportunities and Concerns
25:00 AI's Role in Cancer Treatment and Diagnostics
30:10 Future of AI in Personalized Medicine
32:23 Innovative Approaches to Women's Health
35:11 The Role of AI in Health Optimization
36:51 Empathy and Virtual Reality Experiences
41:05 AI's Impact on Creativity and Accessibility
43:59 AI as a Personal Assistant in Daily Life
48:54 Understanding AI Hallucinations and Their Implications
53:28 The Environmental Impact of AI and Future Solutions
Transcript
[00:00:00.320] - KAYLA BARNES-LENTZ
Welcome to the Longevity Optimization podcast, where we discuss longevity, optimal health, nutrition, peak performance, cognitive excellence, and so much more. Today, we have Harper Carroll on the podcast. Harper Carroll is an AI, ML software engineer and educator who specializes in making artificial intelligence and machine learning concepts accessible to a wide audience. She holds a master's and bachelor's degree in computer science, specializing in AI from Stanford University. Her educational background includes a strong background in mathematics. She has industry experience with senior teams, having worked at Metta on projects related to augmented reality and civic integrity, specifically addressing misinformation and community protection. She has developed AI model coding tutorials and educational content, which have gained extraordinary popularity among online learners. During her time at Stanford, Carol served as teaching assistant for advanced computer science and AI courses, including the PhD level, decision making under uncertainty. Her online educational presence spans platforms such as YouTube and Instagram, where she has accumulated over 300,000 subscribers. Harper, welcome to the podcast.
[00:01:13.570] - HARPER CAROLL
Thank you. Happy to be here, Kayla.
[00:01:15.730] - KAYLA BARNES-LENTZ
It's so exciting to have you. You are like my new friend. I know. It's been so fun. We're great. We tried to film a podcast last time you came over. It didn't work. It was like four hours.
[00:01:26.300] - HARPER CAROLL
We talked for four hours, and then we were like, Wait, we forgot about the podcast, but now we're tired. We talked about too much.
[00:01:33.680] - KAYLA BARNES-LENTZ
Let's just do another day. I know. We had to do a reset. Yeah, we did PMF and talked for four hours. I loved it. How did you feel, by the way? Did you feel good? Oh, yeah.
[00:01:42.380] - HARPER CAROLL
I love to put a video on my story. I was feeling so great. I was like, I think I've been sitting here for over an hour. I'm chilling. I have my amazing water from Kayla.
[00:01:51.880] - KAYLA BARNES-LENTZ
Yeah.
[00:01:52.330] - HARPER CAROLL
Was that like a nanomolecular?
[00:01:54.620] - KAYLA BARNES-LENTZ
So yeah, like nano-restructured, hyper-oxygenated, medical-grade water.
[00:02:00.240] - HARPER CAROLL
Yes. Amazing. I felt great.
[00:02:02.380] - KAYLA BARNES-LENTZ
We tried three different waters while you were here.
[00:02:05.820] - HARPER CAROLL
Oh, yeah. And then you got your shipment in. And you're like, I'm so excited. It tasted great. And this is that, I think.
[00:02:10.350] - KAYLA BARNES-LENTZ
We had that one. This is the Alive water. I know. So basically, we have so many common interests, health. I'm interested in AI. You're an expert in AI. You taught me about this Polymath thing. Oh, yeah. And I was so excited. Maybe we'll just jump in. Okay. Real quick, before we really get into the details, you were It's explaining to me that AI is going to be the ultimate polymath. I had never heard this term, and I'm so excited to learn from you, too. What does that mean?
[00:02:38.690] - HARPER CAROLL
I love this example. A polymath is someone who is an expert in multiple fields. There have only been a few or a few hundred polymaths ever that we know of. A couple of them were Thomas Edison and Nikola Tesla, and they're experts in multiple domains. What that allows is is drawing correlations between these two fields and finding insights that you can't find when you're just looking at one or the other by itself. It leads to extraordinary breakthroughs. But we haven't had that many people who can do that because it's difficult to be an expert in one, let alone two, and then to find that intersection. What AI can do is it can be the ultimate polymath. I mean, it can be an expert in thousands of fields and then find that correlation between all these fields. We'll be able to have advancements that we can't even fathom at this point because of this knowledge, this mighty brain that has knowledge of thousands of different fields and can synthesize across them. Another example, so I read that example, this book, I forget what it's called. I'll let you know. Yeah, it's a great book, and they gave that example.
[00:04:01.700] - HARPER CAROLL
But my example that I like to give before, before I heard the polymath example, was that if you imagine you have a room and you have five people in one, or you have two rooms, five people in one room, experts in different subjects, like what we're talking about. Rather than a polymath, it's five different single subject experts. Then you have an AI that is the polymath in the other room that is one single AI that has exactly the same amount of knowledge as the five experts, but in its just one brain. The AI is going to come up with so many insights and breakthroughs, and we can't even imagine what it's going to come up with, whereas the five people in the other room, even though it has the same amount of information, because it's across different brains, they could spend their whole lives talking and never find the same correlations that the AI can find in seconds or minutes or whatever, what have you. It's pretty amazing I mean, we are just going to see advancements that we can't even fathom. It's really exciting. We'll have breakthroughs in health and longevity and cancer treatment and drug discovery.
[00:05:17.560] - HARPER CAROLL
I'm so excited about the health benefits that we'll see in AI and also just optimization across like grid systems and energy management and food. We'll just be able to optimize. It We could have a utopia from the extraordinary scientific breakthroughs that will come from this master brain.
[00:05:40.200] - KAYLA BARNES-LENTZ
It's exciting. I'm so excited about it, too. Let's go back to the beginning. You started studying AI way before this was cool, I feel like. Yeah. Why got you interested in this career path?
[00:05:55.850] - HARPER CAROLL
Great question. When I was in high school, I've always been a math person ever since I was a little kid, loved math. I also simultaneously was acting. I've loved to act ever since I was a little kid. I would watch the TV and repeat back the lines in a way I felt was more realistic on Disney channel and all that. Anyway, so my senior year, I was interning at Columbia Business School with a professor, Professor Kosti Glaris. He was rad. He helped me learn about social and economic networks. What I was doing was I was modeling how Amazon reviews over time converge into single reviews. People start with very differing reviews, and then over time, they tend to converge, and looking at why that is, and you can model it. Anyway, so I was doing that, and then simultaneously, I was acting on the side, and I got really far in this role. That is the one role I wanted the most my whole life. I was mainly auditioning for... I had just gotten an agent. I was mainly auditioning for film and TV, and I This was one that it was just a play, actually.
[00:07:19.760] - HARPER CAROLL
But it was the best script I'd ever read. It was a five-person play. Granted, it was with Sarah Jessica Parker and Wleith Danner. It was going to be a legit play. And I was so excited about it. So excited. It was just awesome. Me and the director got along so well. And she was like, I can't believe you've never been in anything. You're perfect for this role. Anyway, it was between me and this one other woman, girl, at the time. We were 17. And she got it. And I remember thinking, I was devastated. Normally, I'm like, whatever. But I realized that Sarah Jessica Parker and Blyte Stanner are both 5'3, 5'4. I When I was 5'8 at the time. Now I'm 5'9, I grew an inch. When I was 5'8 at the time, and I was supposed to play their niece on stage. There was just no way they could ever cast me, honestly. It was just not going to happen. It was wishful thinking for me and the director because we were so excited, but it just wasn't going to work. I remember thinking, I wish I could train an AI model. Or at the time, it wasn't AI.
[00:08:24.620] - HARPER CAROLL
While I was working with Columbia Business School professor, I realized, Oh, my God, you can model social situations with numbers. I was like, This is crazy. At the same time, I was taking calculus, and I don't think it was multivariable at the time, but I was taking calculus, which is all about optimizing, finding the minimum and optimizing different functions, finding the minimum on a graph. I thought that was the most fun thing ever was optimization. I thought it was the coolest math ever. I just loved doing differential equations and all these things. At the same time, learning about how you can model social situations with numbers. Then I had this social situation that I wanted to model, which was, what if I took stats from IMDb and was able to actually train a model to tell me, you actually have a shot at this audition. Based on the people who are already cast in history, whatever, this is something that you should audition for and rule out a lot of auditions because when you're an actress, you're auditioning all the time. Anyway, so that was my thinking. But I didn't still have a word for what was.
[00:09:30.500] - HARPER CAROLL
I just thought it was math. Then I went to Stanford, and I had never done computer science, although I did use HTML as a kid. I had a website briefly, and we can talk about that at some point, but I got in trouble for that website.
[00:09:44.670] - KAYLA BARNES-LENTZ
I'm What was the... Okay, we have to struggle back to that. I would have to say.
[00:09:50.460] - HARPER CAROLL
I was basically new to computer science, and people in my hall were like, Take it, try it. Because at Stanford, that's the thing. Everyone does computer science, or at least they try it. They see us 101. 106a is the actual course. Anyway, so I take it and then it's super hard. I had a really hard time with 106A because it's basically like learning a language. Then 106a is algorithm. So 106A, you're learning Java, you're making games and graphics and basic graphics and things like that. And then 106B is algorithms, which is basically processes how to get from A to B and do it in a really elegant amazing way. And that was the coolest thing ever. Again, super challenging, but it trained my brain to think in a new way. And I thought that was super cool. It was my favorite class at the time. And then over time, as you had to decide, what major do you want to do? Sorry, this is a super long story, but- No, I love that. Okay, good. Over time, at some point, you have to choose a concentration. There are different ones, human-computer interaction, which is basically people who make apps and want to make them really beautiful.
[00:10:58.450] - HARPER CAROLL
They'll go that route. One of them was systems, like how computers work at their core. Then there's AI. I'm looking at AI and I realize it's literally everything I loved in high school. It's optimization, it's calculus, in this case, multivariable, and linear algebra is added, which was super cool and challenging and also awesome. It's modeling social situations with numbers, and then you can create these predictive machines. It was just perfect for me. It's like algorithms as well to tie it back to that. It was just everything that I had found interesting from senior year until that point. I was drawn in and it just felt like the perfect emergence. Now when I look back at how I was in high school with that problem in the social and economic networks and working with that professor, and I'm like, Oh, my God, it was AI. All that time it was AI. I didn't know. I didn't have a word for it.
[00:12:00.020] - KAYLA BARNES-LENTZ
Oh, my gosh. I love that story. Was it a really popular course at that time?
[00:12:07.540] - HARPER CAROLL
Which one? 106a or just the AI curriculum? Oh, yes. Okay, I forgot the original question. It was not super popular. No. It was... Not a lot of people were doing it.
[00:12:26.090] - KAYLA BARNES-LENTZ
I love that. Well, I mean, now, fast forward, it's like everybody's talking about.
[00:12:33.280] - HARPER CAROLL
Totally. I mean, people did it, to be clear. People did it, but it was not the most popular.
[00:12:37.080] - KAYLA BARNES-LENTZ
Yeah, that makes sense. Okay, well, what was the website and how did you get in trouble about it?
[00:12:42.340] - HARPER CAROLL
Okay, so I made the website in first grade, I think.
[00:12:46.720] - KAYLA BARNES-LENTZ
First grade? Oh, my gosh. I love a little Harvard.
[00:12:49.710] - HARPER CAROLL
I loved to create. I was always just making something. But in first grade, my dad bought harpercarol. Com. You didn't I was going to do AI back then. I still have the website today, and I wanted to be on the internet. The internet was super cool. I learned basic HTML to make this site that was like, Hi, I'm Harper. Here are my Tom I got you. I got you tips and tricks. I had a whole section. Apparently, you could do different... There's like ABC buttons, like key strokes and get different sheet codes and things. I had a whole section for that. I had a joke section. But it was I guess relatively early in the days of the internet. Maybe, I don't know how old I was, seven or something. It was relatively early in the days I was born in '95. My principal called me in and was like, I heard you have a website. You need to take it down. I was like, Why? She was like, It's not safe. You shouldn't be on the Internet. You're a child. You don't know who's on the internet. You don't know who's on there. He's going to find you.
[00:13:58.690] - HARPER CAROLL
I I had to take it down. Honestly, maybe I would have become a prolific coder because I was having a great time working on my website, and I had to stop. I just stopped coding. But it came back to me later in life, you know how the universe does that. I got in trouble. There was nothing bad on the site. Unfortunately, it's not that juicy of a story, but they were just afraid.
[00:14:23.000] - KAYLA BARNES-LENTZ
Well, I love that, and that's so cute. You were just like little coding in first grade. It's so cute. For anyone that doesn't really understand what… On this show, I typically talk a lot about longevity and health. If you don't really know exactly what AI means, what's the most basic definition of it?
[00:14:44.130] - HARPER CAROLL
Ai is simulated intelligence to buy a machine by the most basic definition. That could come from hard coding a machine. Say I could code the machine where if it receives the input high, it should output high. If you only say high, it seems like it's intelligent because it says high back. But that's not really true intelligence, right? It's just hard coded in that sense. I could say, if it says high, say high. If it says what's up, say what's up. Then there's machine learning, which is when the machine is actually learning from examples. Say I passed it a ton of transcripts and it We saw that when it receives hi, it might say, what's up? It might say, hello, nice to meet you. It could say all these different things. Then when inference time, so when it's out in the world, inference means when a model is deployed and it receives an input and it gives its output. At inference time, it has this array of options that it could do, and it probabilistically chooses based on what's most likely. It's determined in its training phase what is most likely. But yeah, so AI, at its most basic definition, is simulated intelligence, like by a machine.
[00:16:08.330] - HARPER CAROLL
Then now, today, you typically think of AI as... Most AI is machine learning. So chat bots and things that you use or image generation models, they all use machine learning. They all learn from data. But yeah, there is a little distinction there. But I think over time, maybe that distinction will go away. Like I think all AI will also be machine learning, but I think the purists, there's a difference, technically.
[00:16:39.880] - KAYLA BARNES-LENTZ
Before we jump into health, so I watched the movie that you told me about the other day.
[00:16:44.680] - HARPER CAROLL
Oh, my Which is a Megan Fox one? Yeah. What's that called? What's that called? I don't remember.
[00:16:48.190] - KAYLA BARNES-LENTZ
Something like an S. Yeah.
[00:16:52.010] - HARPER CAROLL
I know. I forget.
[00:16:54.440] - KAYLA BARNES-LENTZ
But yeah, I can't remember either. But it started with an S. It's a new movie. Essentially, the premise is it's an AI robot. Yes. The robot, things are going well in the beginning, and it's just showing up to serve and make the owner happy. Then a reset occurs, and then it goes totally rogue.
[00:17:16.300] - HARPER CAROLL
No sense, by the way. But okay. The fact that she goes rogue after a button reset is just crazy. But okay, yeah.
[00:17:22.850] - KAYLA BARNES-LENTZ
Was it because he told her that she could erase code? Is that where it went wrong?
[00:17:28.280] - HARPER CAROLL
Maybe. I mean, he told her to erase data, not code.
[00:17:35.610] - KAYLA BARNES-LENTZ
Okay, yeah. The data from that movie.
[00:17:37.890] - HARPER CAROLL
But then she would have to retrain. I don't know. There are some stretches here, but like, continue.
[00:17:43.630] - KAYLA BARNES-LENTZ
So okay, that happened And then this AI robot goes just totally rogue and starts doing a lot of bad things, right? Like seducing the husband, committing murder, a variety of different things. I didn't even think about being scared about this, honestly, until I watched this. I know. Sorry, but no, that's okay. You obviously hear way more about AI, right? And so some people have these concerns, is it going to get so smart and then just try to attack us. What What are your thoughts on that?
[00:18:16.720] - HARPER CAROLL
I am not really into the humanoid space. That's a whole thing that I've never really been interested in or like, robotics never really drawn me in. But yeah, I think I've always said this is that that thing spooks me, but maybe it's because I don't know anything about it. Or I don't know nearly as much as the experts do on it or even someone who's a little bit versed in it. But it does scare me. The humanoid stuff is weird to me. When I think of how AI is going to help the world, I think of just optimizing systems. I think of the pure math and taking in data and looking at how to help things get better. I'm not really thinking about making new humans or new humanoids. I'm less excited about that. I watched that movie and I was like, I need to really think about this and pray on this. What is my response to this? Because people are going to ask. People are afraid of AI. And watching that, I'm like, Oh, okay. I see why. But I feel like something like that is so unlikely to happen. Well, first off, I think the robot would have to learn how to do all the things that it did.
[00:19:51.390] - HARPER CAROLL
It would learn how to shoot a gun. But I guess you could say once the models get smart enough, if they even watch a movie on how to do it or a video on YouTube or something, they could figure it out. So I don't want to say that there's no risk there because I don't know about the humanoid bot world, and it does spook me. But the AI that I'm really focusing on is not in that area. But I am curious. I would like to dig into that more because I want an answer, and I know that there are experts in it. Like Elon, he's coming out his humanoid bots, right? I wonder what he would say about it. I know also, though, Elon is concerned about extinctionism. Elon has come out and said, We need to slow down AI, but then he still speeds up AI. So I don't know. But yeah, I wish I had a better answer for you on that. Maybe in a couple of months, I will.
[00:20:52.760] - KAYLA BARNES-LENTZ
Yeah, we'll keep this up, dude. I will. But yeah, truthfully, I have only thought of AI, not to the level that you have, but in the way that I've been very beneficial, helping to solve a lot of the... Some of the biggest problems. But then you watch the movie and I'm like, Okay, well, this could be something. Totally.
[00:21:12.530] - HARPER CAROLL
It's not hard to be spooked by That's what I thought.
[00:21:15.660] - KAYLA BARNES-LENTZ
Yeah. What about AI and health? Obviously, that is a big interest to me. We've already been chatting about some of the things that are going around in my mind in terms of how can we best use it. I just, truthfully, I don't know a lot about it. My extent, I told you, I've been trying to train a little AI bot to take my data and then give me some information. But where do you see AI really being able to make a big impact on longevity You mentioned potentially cancer treatments.
[00:21:48.470] - HARPER CAROLL
Okay, so AI is able to identify special cells and protein structures and go over massive amounts of data and find special things that we wouldn't be able to find by the fact that it's able to analyze just enormous amounts of data. That doesn't really make sense for our human minds, but let me just give an example. Glioblastoma is a type of cancer that is the most deadly cancer. I think, I forget the percentage, but was it like only 10% of people survive within five years of diagnosis or something like that? It's a brain cancer. You get tumors in your brain. Immunotherapy, which is an extraordinary amazing breakthrough in cancer treatment recently, hasn't been able to help glioblastoma because the immune cells that are being transported in the blood aren't able to reach the tumor because there's a blood-brain barrier. So the brain cells are outside the blood, so the blood doesn't reach it. What AI was able to do was identify cells in the brain tumor that could cancer cells, that could transform from cancer cells into cancer-fighting cells.
[00:23:15.390] - KAYLA BARNES-LENTZ
Mom.
[00:23:16.810] - HARPER CAROLL
What? They were able to find there are certain types of cells that can transform from cancer cells, not only into healthy cells, but into cancer-fighting cells. Then they were able to do that.
[00:23:31.910] - KAYLA BARNES-LENTZ
Like, activate them. Yeah. Like, natural killer cells, essentially.
[00:23:35.610] - HARPER CAROLL
T cells, I think is what they're called. Yeah. And so what? And so AI is able to analyze enormous amounts of data and find these little optimizations and these miraculous cures. I know that it'll be huge for drug discovery, and I bet we could Maybe there's some cell that is capable of aging backwards or pausing or distributing healthy mitochondria. I don't know. This is your thing. To preserve the health of the other cells, I don't know. But AI will help us find those things because it's able to analyze enormous amounts of data.
[00:24:22.770] - KAYLA BARNES-LENTZ
I feel like early diagnostics, too, is an area in which- Yes. A thousand %.
[00:24:28.070] - HARPER CAROLL
Have you seen the images of breast cancer I don't know if I've seen those ones, but I know, for example, I already get this thing called a clearly soft plaque analysis.
[00:24:37.980] - KAYLA BARNES-LENTZ
The standard of care for cardiovascular health is getting a calcium score. That will tell you the amount of calcified plaque in your arteries, but they already have something better, which is this soft plaque analysis. It wouldn't be easily detectable to just a standard radiologist, for example, is what I'm understanding. But the AI can go in and see if there's a soft plaques, which that's like precalcified plaques. Then you can do more about it. If you go in and instead of waiting to get calcified plaques, you now know way earlier with AI to be able to identify these soft plaques.
[00:25:13.800] - HARPER CAROLL
Oh, my God, I love that. There's another one. I forget the type of imaging that they do, but to detect tumors and things like that, breast cancer. Like an MRI? Maybe it's an MRI.
[00:25:27.820] - KAYLA BARNES-LENTZ
But anyway, the imaging- Breast cancer would usually be a mammogram program?
[00:25:30.630] - HARPER CAROLL
Yeah, or it's like a black and white image. But basically, there are any imaging, any type of medical imaging. Ai is able to detect smaller things that the human eye wouldn't necessarily be able to see, and it's able to detect breast cancer, I think five years early, which is an enormous time saved, right? Yeah. Your chances of survival go up so much. There's actually, I was reading somewhere recently, in the near future, it will likely be illegal to not diagnose someone using AI, to at least run an AI to double-check your diagnosis or to assist your diagnosis.
[00:26:10.820] - KAYLA BARNES-LENTZ
Wow. That's exciting because I've seen some stats. I don't know if I haven't confirmed them 100%, but that AI is making better diagnosis than doctors. Of course.
[00:26:21.670] - HARPER CAROLL
I mean, of course. It's looking at... If you think of it, say you are like, think about the best doctor you know. Likely, they have 50 years of experience or something crazy, right? It's experience, not just doctors, for any field, like therapists or whatever. They're like, I've done this for 30 years. I've seen so many samples. I see this. I notice patterns. I can tell you that in this situation, this is going to happen, or this symptoms, this is the diagnosis, or this is the treatment. Ai is that It's all of those people who have been learning for 50 years, but times 100,000 or more. It has so much experience. It has the experience of eventually, my hope is that we'll have a database that's anonymized of people's health. We were talking about this, their health stats, and then following them over time, like what illnesses did they get? What treatments did they use? Did the treatments work? And ultimately, we'll be able to diagnose people accurately, Every time, every time. If we had the entire world or even just the United States health information and really tracked it and put it in a database and trained a model over time and anonymized it, we could have unbelievable health in the United States because we would also know what treatments work.
[00:27:50.470] - HARPER CAROLL
How do we help people, what works, what doesn't. It's just amazing. Imagine it's the expert, but it's thousands of experts over thousands of years all in a brain.
[00:28:03.930] - KAYLA BARNES-LENTZ
It's so exciting. I mean, yeah, you brought up drug development. I think it's actually a very low percentage, I don't know the exact number, but of drugs actually working for people. It's a relatively low percentage. It could be for a variety of different reasons, one of them, genetics. If you have your genetic screening and then you compare it with the activities in the drug, essentially, you can figure out if that's going to work well for you if there's another drug. But now at this stage, it's just a try and see type model. Some people are prescribed three, four, five different types of medication all in the same family prior to actually finding one that they find works for them. That seems like a really broken model.
[00:28:49.340] - HARPER CAROLL
Totally. If you had the data on what works, what does this person look like? What's their genetic makeup, what's their health history, whatever, which drugs work for them, which don't, this is so valuable. Then you could help people, ultimately, not have to try all the different ones. You could give them one that you think that you have a very high confidence will work for them. If it doesn't, then that's just more data for eventually the model to get better.
[00:29:13.920] - KAYLA BARNES-LENTZ
Yeah.
[00:29:14.650] - HARPER CAROLL
Yeah, it's just so exciting.
[00:29:17.480] - KAYLA BARNES-LENTZ
It's crazy how exciting it is. I know. I'm thinking about all the things that it could identify for longevity. Yes, definitely. Maybe it would start in an organ level and I think about how to reverse each organ system or the whole body systemically, one of the two. I think definitely it's going to be so fun. I take so many supplements. I take like 40 supplements a day. It's wild. If I could put all of the supplements into one AI, which is what I'm going to hopefully be able to do soon. And also see, are there any contraindications with any of those ? Are they counteracting each other in any way? Then understanding I think what is going to be so interesting. At the Buck Institute, they're working on putting cells into this aging device. It will age the cells and tell you essentially what disease or your pathway of decreased health. What's really going to take you out and how? Wow. Okay. Super interesting. If you have that data- So it take people's individual cells. Yeah.
[00:30:25.050] - HARPER CAROLL
They're putting- Because it's different per person, what's going to take them out. That's also AI to age AI. Anyway. Yeah.
[00:30:31.310] - KAYLA BARNES-LENTZ
Well, yeah, it must be AI because they're basically putting it into this compartment where it's aging it.
[00:30:37.740] - HARPER CAROLL
Oh, maybe it's not AI. I was thinking it was a simulator. But anyway.
[00:30:40.040] - KAYLA BARNES-LENTZ
I'm sure that they're using AI in there. But it's like, okay, maybe you might have cardiovascular disease, based on your current health and your cellular makeup. Cool. Then if you could overlay with AI, knowing that that is the potential pathway that you're going down, and then overlay which molecules, exogenous molecules, could help to counteract that and prevent it. That's so exciting. That's brilliant.
[00:31:06.020] - HARPER CAROLL
Brilliant. Then over time, as you're giving more data to the model, it just gets better and better.
[00:31:10.580] - KAYLA BARNES-LENTZ
Because I think in the future, AI will be developing those protocols on a daily basis, and then we just have a 3D printer of all of the molecules, hundreds, and then it creates it into an elixir that you drink every night. Oh my gosh.
[00:31:24.290] - HARPER CAROLL
I hear a 3D printer and I'm thinking of that horrible killer of the United... You know he printed a gun?
[00:31:33.080] - KAYLA BARNES-LENTZ
Who? Which one?
[00:31:34.360] - HARPER CAROLL
The United Health care seat. Really? The killer printed his gun and a silencer.
[00:31:39.750] - KAYLA BARNES-LENTZ
Oh my gosh. What I have earned.
[00:31:41.140] - HARPER CAROLL
What? Okay. I don't know. That's just crazy. Also horrible. I think it's horrible.
[00:31:46.680] - KAYLA BARNES-LENTZ
Yeah, that's horrible.
[00:31:48.270] - HARPER CAROLL
Anyway, so 3D printer. I know. 3d printers are just crazy. I would like to use them more.
[00:31:54.820] - KAYLA BARNES-LENTZ
I know. Me too. But how cool would it be and fun would it be instead of just like, I amend supplements every quarter. But no, this would be on a daily basis. We have a little... I just saw Aura Ring and Dexcom have now come together. So CGM with Aura Ring, and who knows if it'll be Aura or if it'll be someone else. But to have ongoing data, blood data, hormone data for women.
[00:32:19.520] - HARPER CAROLL
I'm so excited for the hormone test. We women, we need better hormone testing.
[00:32:23.520] - KAYLA BARNES-LENTZ
I know. Then it just pings you when it's the best time to eat this specific thing in your cycle, and it It gives you each day of your cycle a different unique supplement. Just an elixir, so it's super bioavailable, and you don't have to take the pills with the fillers, the cellulose. Just a little beverage.
[00:32:41.930] - HARPER CAROLL
Oh, that's why. Because you've been saying elixir a bunch, and that's why. It's because you don't have to take the pill.
[00:32:46.710] - KAYLA BARNES-LENTZ
Yeah, exactly. The filler stuff. Yeah, because the fillers and then even the capsule make it easy on your liver and kidneys as possible, really.
[00:32:55.600] - HARPER CAROLL
Yeah. I'm really excited for women's health. I recently, or not that recently, but there was one time that I was getting super hormonal before my cycle. I went and I was told, see a psychiatrist or something. I was like, okay, I guess I'll go. Because I thought, I want to get my hormones checked and see maybe I have too much estrogen or something. I saw her and I told her what was going on. She was like, okay, we have two options. We can put you on birth control or we can put you on an antidepressive.
[00:33:32.460] - KAYLA BARNES-LENTZ
No, you told me this.
[00:33:34.600] - HARPER CAROLL
I was like, Are you kidding? I literally was like, That's it. You're charging my insurance like $1,000 for this session. You're offering those two options. You're not even going to run a test on my hormones. You're not even going to think about it. And she was like, No, we don't do that. Sorry.
[00:33:52.320] - KAYLA BARNES-LENTZ
I was like, Screw this. It's so crazy. I know that this happens, but it always It hurts me more when I actually hear that it happened to someone that I know. It's wild.
[00:34:05.010] - HARPER CAROLL
It's crazy. That was my two options. Antidepressants or birth control. Oh, you're feeling hormonal? Let's put you on birth control.
[00:34:13.510] - KAYLA BARNES-LENTZ
What? Oh my gosh. For a million reasons. I'm so glad you could do that. Oh, yeah. I'm so glad.
[00:34:18.430] - HARPER CAROLL
Of course not. I went and saw a naturopath, and then she tested my hormones. She's a doctor and naturopath because she has an MD and also does alternative medicine. I know, right? Checking your hormones.
[00:34:31.460] - KAYLA BARNES-LENTZ
It's so alternative. Totally. Or maybe just how medicine probably should be practiced. Let's see what's going on and then think about the best. And let's see why. Right. We don't ask why. We don't know. It's just like, here's the diagnosis, here's the prescription. That's best suited.
[00:34:50.450] - HARPER CAROLL
Right.
[00:34:51.270] - KAYLA BARNES-LENTZ
Hi, guys. I'm going to interrupt this episode for a brief announcement. As you may or may not know, I started a community for females by females, and it's a a female longevity optimization community. This is a place that you can connect with like-minded women. We are all here to support each other, and there's a variety of different benefits to being a member. You get a monthly ask me anything, so submit your questions, and I'll answer them directly. We also have an entire library of courses on all of the important components of longevity, such as labs, nutrition, exercise, sleep optimization, longevity optimization protocols that I'm doing, along with real-time updates to my personal protocols. There are so many benefits of being a member of the community. We'll also be doing in-person live events here in California and virtual events for anyone that can't attend. But if you're interested in joining the community, I would absolutely love to see you there, and I will include a link in the show notes. What are your favorite aspects of AI? What makes you most passionate about it?
[00:35:53.690] - HARPER CAROLL
I think it's just what we've talked about, the fact that you can optimize any system and that you can have these insights across domains. It's like a super brain. It's a super brain. I think that we can really help people. I feel like my purpose with all of this, with the education, is because I really believe that AI can help bring a utopia. I really do. The impact that we can have on humanity is absolutely enormous. I feel like my purpose in all of this is to help people understand that. In the meantime, how they can use it to their benefit because it can help streamline a lot of processes and take away the tedious tasks that you don't like doing. For example, I have a friend who's a photographer, and she told me, she's like, Oh, my God, I'm so excited about this new AI update. I can now just click an area that I don't want to be there. For example, if I take a photo and there's this poll in the that I don't like. I just click it and it's gone. I used to take four hours doing that. Now it frees me up to take more photos, edit quicker, edit more.
[00:37:11.890] - HARPER CAROLL
She's freed up to do more. And so I think what I really want people to know, everyone is like, AI is going to take our jobs. It's going to ruin everything. No, no. It will enable you to do more of what you love to do in your job. It will free you up to do different kinds of work. It's democratizing also creativity. So for example, say I wanted to make a movie. I'm so excited. I'm like, I have this really strong vision of what I want in my mind. Okay? And I'm just have this. I'm so excited. And I have this movie I want to make. But I don't have a ton of money. I can't hire a filmographer. I can't hire actors. I can't hire this or that. But instead, I can make this movie start to finish completely on my own with no cost. And so then, I hear people in my head saying, then what's going to happen to the actors? What's going to happen to the filmographers? They're still going to exist. They're still going to exist. People are still going to want to make movies with real people. I'm talking about just enabling creativity in a way that it's never been enabled before, democratizing it so that people can create without needing a lot of money or expertise or a beautiful location to film at.
[00:38:37.950] - HARPER CAROLL
I'm just excited about it. Also, another thing, this is with virtual reality and augmented Reality. But I really like that virtual reality. I'm just thinking about this right now. You can suddenly be in a beautiful place. You just put on the headset and you're like, You can design a place that is perfectly for you, and you're just there, and you can rest and whatever and experience whatever it is that you want to experience. Your brain really doesn't know the difference, honestly. I used to work in the VR lab at Stanford, the Virtual Human Interaction Lab, where they studied VR's impact on empathy. For example, they had this one experience where you experience being a homeless person So you're in your house and then you get an eviction notice and then you have to go to a homeless shelter and you're turned away and all these things. There's a survey before and after about your empathy towards people who are homeless. They just found that across all different domains, VR greatly increases people's ability to empathize because your brain really doesn't know the difference. There's this super famous VR experiment where our experience Where you're walking across a plank with a headset on.
[00:40:04.350] - HARPER CAROLL
People can't do it. You're in a plank elevated in the sky. You know, your brain knows. Like you're in a room or you know, consciously, but your your whole system cannot... It's too scary. You can't do it. Anyway, back to what I was saying, it democratizes access to beautiful places and beautiful experiences and being able to travel the world. Yes, people are going to be like, Well, just make things less expensive. Everyone has whatever. But I think it's great. I do. Yes, things can be better as well. People can experience the real world as well. But this is a great option for people to have. Also, just in a short time, say you don't have time to take a vacation and you just get home and you're tired after work and you just want to be on a tropical island for a few hours. Suddenly you're there. Anyway, there's just so much that we can do with AI, and that's virtual reality. But there are AI components. Ultimately, you can probably point and say, I want to put a desk there. That will be AI. Once it becomes really interactive, that We need AI for that.
[00:41:32.110] - KAYLA BARNES-LENTZ
The other thing that will be so cool is if you're wearing contacts or glasses, like some of the glasses, to actually track what you're doing on a day-to-day basis. Yeah. Then it can I'm seeing so many people that are maybe not doing the healthiest habits because it's an interesting phenomenon. I don't know how much science is actually around this. I just know it exists because we hear it all the time in my clinic. People tend to think they're doing really good in health perspective. Or we've all experienced this. People go to the gym and they stay on their phone and they're literally doing social media at the gym. But because they say that they went to the gym, they then maybe eat more or are lax about other things because they technically went to the gym. Well, if you were wearing glasses that were actually tracking every movement that you make, every supplement that you take, every piece of food that you put into your mouth, it could give you such a better picture as to how you are where you're at. Yes. Then layer on the AI, too, so we can say, Okay, this is what you're doing.
[00:42:36.980] - KAYLA BARNES-LENTZ
But if you make these small tweaks, then you can do better and feel better and have better health. I love that.
[00:42:43.270] - HARPER CAROLL
I hadn't even thought of that example. Absolutely. That actually makes me think about just on a tiny scale, but something that exists already today, there are these glasses that came out for people who are deaf, where they wear them, and on the inside, as people are talking, it puts it into text, transcribes. Oh, wow.
[00:43:04.040] - KAYLA BARNES-LENTZ
That's amazing.
[00:43:04.570] - HARPER CAROLL
Right. They're able to... They don't need sign language anymore. I watched this video of this woman opening them, and she's deaf, and she puts them on, and she's crying. She's so happy, and it's just so emotional. The people who will be helped with AI and the tech that will make their lives so much better, people who with locked in syndrome who are able to communicate or or have limbs that work really well, controlled with their brain. There's just so much that we're going to do for people. It's just really exciting.
[00:43:41.920] - KAYLA BARNES-LENTZ
I know. I can't wait. I need to get the glasses I just start tracking all my movements and hauling all my labs together.
[00:43:50.840] - HARPER CAROLL
I'm excited. Yeah, real-time suggestions.
[00:43:53.220] - KAYLA BARNES-LENTZ
Yeah.
[00:43:53.900] - HARPER CAROLL
Someone I know that makes me think, too, is building something called, I think, friend. It's a pendant that you wear and it listens to you. I don't know the details of it. I think I talked to him once, or someone who's working with him, and he said they don't store the text, the words specifically or the audio specifically, but they make graph representations of information. It's privacy-safe. I don't know. But basically, it's this friend who will listen to your conversations and you can say, Oh, how was what I said, Is that okay? Or what did you think about what Kayla said about this? Did you think she liked it? How did my presentation go? Based on what I did today, what's a movie I should watch tonight that will help me learn about this and that. It's this life assistant. I don't know about that, but I do know in terms of, I don't know whether I would use that or not, but I could definitely see it being useful for some people. But I have a friend, and I love this example for Claude. I think AI will be enormously helpful in terms of therapy, and a friend in some cases.
[00:45:12.660] - HARPER CAROLL
I have a friend who is in the dating pool, and she really likes this guy, and he is very avoidant, like textbook avoidant. She would rotate through her friends. She I was super nervous about texting him and rotate through her friends like, Was this okay? What do you think? Why didn't he respond? Whatever. And get her friend's disparate responses to whatever is happening at that moment with him. I mean, her friend is like, blind leading the blind. No one, you know what I mean? They're like, Yeah, no, that's bad. He should do this or that. She called me and I said, Hey, I want you to try Claude. She's like, What? I'm like, It's like ChatGPT, but it's more personable It has just a nice personality. Once you start using Claude, you can't really go back to Chapter 2PT. Claude is just really sweet. I was like, You can start a thread with it and you can upload screenshots and you can talk to it about, Hey, I went on a date with this guy this many times. He's avoidant, whatever. And talk about your relationship like you would to a friend or to a therapist.
[00:46:18.330] - HARPER CAROLL
But then, Claude, it will build this massive history over time. Eventually, you reach a limit, but that's after a while. And he'll be able to talk to you, but also with expert guidance. So it's like, whereas her friends would be like, That's not cool that he's disappeared for a week or whatever after being very close. Claude would be like, Well, that's actually textbook avoidance and if you give him some space to process. Claude is an expert. Whereas before your friends are like, They don't really know, and they're just trying to help you, and they love you. But Claude is this friend that's there 24/7. She doesn't have to call them. She just write at the moment and get insight. She's like, Harper, Claude has changed my life. Thank you. I'm so much more chill. I don't have to call all my friends about this all the time. I feel reassured instantly. I just think everyone should be using Claude for things like that.
[00:47:16.640] - KAYLA BARNES-LENTZ
I should move over to Claude. Why is Claude better than ChatGPT? The personal.
[00:47:21.900] - HARPER CAROLL
It just has a good vibe to it. Anthropic is a company that makes Claude. So OpenAI makes ChatGPT, Anthropicic makes Claude. Anthropic's whole thing is about AI alignment. So making AI that is really helpful and supportive for humans. They spend a lot of money. Honestly, when you think about research or salaries, it's ultimately billions of dollars on Probably, on research. Researching how to make safe AI. Ai that isn't going to go rogue in this movie that we watched. They put a lot of time into making this AI that is really personable and sweet and kind and very human-like. It just has a good vibe. I don't know how to explain it. It just has a good vibe. I think that reflects the vibe of the company, too. Like the creator of the company. I just like Anthropic a lot. So yeah, Claude is great. Chatgpt, it has more features, like voice mode. But I think Claude might have just integrated I did that or added that. Chatgpt can search the web, which Claude cannot do. So, oh, this is an important thing to cover. Hallucinations. I really like more people need to know about this.
[00:48:41.680] - HARPER CAROLL
So can I just get into it? Yeah. Okay. So Models are trained, like a base model that isn't augmented with the ability to search the web is trained, just large language models, are trained on massive amounts of data, and they are trained to find patterns text and then output new text according to the patterns that it's learned before. It can make things up. It doesn't really have a concept of this is fact or this is not. If you have a model that cannot search the web and validate what it's outputting, it can make things up. If you are looking for research or you need a real fact, you need to be positive that the fact is true, you probably don't want to use a base model like Claude, which doesn't currently search the web. Perplexity searches the web. I think ChatGPT, the pro-version, searches the web. Perplexity Pro, you have to pay for it, also searches the web. But this is very important. Someone at Stanford, a Stanford professor who's actually an expert on misinformation, and tech and misinformation, is now facing perjury because there was this law that was considered considering, I think it was ban of political deep fakes or something like that.
[00:50:05.570] - HARPER CAROLL
In some state, they were considering banning deep fakes that are about politicians or something. I don't really think that's correct. Anyway, this professor at Stanford, who's his expert on AI and tech and misinformation, he wrote this 13-page or 12-page paper in support of the ban. It had, I think it was seven sources, and two of them weren't real. He signed something that said, Under penalty of perjury, or... I don't know. Yeah. I declare that this is all true to the best of my knowledge. He used something like ChatGPT. I think he actually recently admitted to using ChatGPT to find sources for his paper. Find me research that shows like this. There are AI services that do this. If you want research that supports something that you think might be true, you can ask an AI and it will find you real research papers that support that. It's amazing. It's amazing. Or you can just ask it to gather info on a certain topic or whatever. Anyway, he asked it for resources. Two of them weren't real, but he I didn't know that. But the reason that happened was because ChatGPT can hallucinate these models. If they aren't trained to search the web, can hallucinate.
[00:51:37.760] - HARPER CAROLL
They just output text according to prior patterns. For example, if you have a bunch of examples of case numbers, and they're 10-digit case numbers, and you ask it to output, Find me a case number that proves that... I don't know. X is true. It might output a 10-digit case number and then a title that seems to prove what you're saying, but it's just making that number up because it's not referencing real data. So if you want something, if you have to know that something is true, make sure it searches the web and also be aware of training limits on models, like when they stopped getting training data. I was just I was talking to someone who wanted to have a summary of a book, and it was a recent book. I said, you definitely want to use a model that searches the web because the book didn't come out when it was being trained. Anyway.
[00:52:49.830] - KAYLA BARNES-LENTZ
How quickly are the models being trained?
[00:52:53.160] - HARPER CAROLL
I don't know. I think I was seeing something like coming out with a new model once a year. Oh, wow. Maybe once every six months, something like that.
[00:53:02.750] - KAYLA BARNES-LENTZ
Things change so quickly now.
[00:53:04.270] - HARPER CAROLL
Exactly. And so ultimately, all these models will have searching the web capabilities built in.
[00:53:09.710] - KAYLA BARNES-LENTZ
Is it really hard to do that? How come cloud just doesn't do that?
[00:53:14.150] - HARPER CAROLL
Yeah. I mean, it's an added feature where you have to have it detect, should I search the web for this? Then it needs to... I mean, it takes a lot of power. That's why a lot of them are paid models. Cloud is free to a certain limit I've never reached, I think, the limit. But anyway, it's free. But if you're going to search the web, it has to look at a bunch of sources, synthesizes this relevant, scan the whole page, put it together, attach only the important parts to the model. It's doing a lot of computer work, which is expensive. Compute is expensive. It takes more coding to get it to do something like that.
[00:53:59.190] - KAYLA BARNES-LENTZ
So What are people concerned about the environment with AI? Is it it's using too much water or something?
[00:54:04.870] - HARPER CAROLL
Yeah, it's using a lot of energy to power the machine. Ai models use computers, basically, GPUs that are chips that are optimized for AI processes. These chips and these computers It takes a lot of energy. It takes a lot of energy to run inference. We talked about that at the beginning, which is when an AI model is deployed or even training, that takes a lot of energy, too. But there's one stat that says, if Google search, augmented with AI takes 10 times the energy of a normal Google search. It's like in Google where you can search something and it gives you a little AI synopsis, that's 10 times the energy. It takes so much energy to power the computers that run the AI. We basically need to find energy sources. I recently did a 10-day series with Isabel Bumekhi, who is a nuclear electricity influencer, a leading nuclear electricity influencer. We did a series on why nuclear can help and why they're really good for each other. That's something that a lot of companies are investing in right now is nuclear. Isabelle talks about why... She's coming out with a book. She talks about why it got a bad rap that wasn't necessarily justified.
[00:55:44.220] - HARPER CAROLL
We talk about that. I've also heard, I think Elon was talking about this recently, solar. I mean, honestly, if we could just get all our energy needs from the sun, that'd be so cool. I mean, the sun has unlimited energy. If the sun disappears, so do we. It's a very reliable source. I mean, AI will be able to enable these technological breakthroughs. Right now, solar panels or the systems we have now to get the energy from the sun aren't reliable enough or can't store enough. For... It's not working. But ultimately, I think with AI helping us make these technological breakthroughs, we'll be able to create a technology that harnesses enough energy from the sun in a small space or however you need. I'm excited about that as well. I think AI, although it uses a lot of energy and people are worried about it, it will also help us make these energetic technological breakthroughs as well. We'll probably cure any carbon emission issues or any issue that we have with the insights that AI can provide.
[00:56:52.090] - KAYLA BARNES-LENTZ
So exciting. Yeah. Yeah. So exciting. Well, this has been so fun. I know. Do you have any last notes on AI that you want to share?
[00:57:00.820] - HARPER CAROLL
We talked about hallucinations and the importance of using a model that searches the web because that's something that I really want to get out there. I I think it's open source versus closed source. I don't think we need to talk about that. Just scanning over some of my recent videos. I wonder if we should talk about how AI models work, like input and output, but maybe we don't need to.
[00:57:36.200] - KAYLA BARNES-LENTZ
Why don't you give us the short version of how they work for anyone that is interested.
[00:57:42.350] - HARPER CAROLL
Okay. When you're using a chatbot model that is using a neural network. Neural networks are a type of AI model. What they basically do is they take input and output, so examples of They have input and output, and then they find the equation that transforms input into output. One second, sorry. Oh, yeah, no, go for it.
[00:58:08.760] - KAYLA BARNES-LENTZ
Sometimes when you swole, does this come the wrong way, though?
[00:58:11.070] - HARPER CAROLL
Oh, 100 %. A hundred %. Yeah. Or like, I've heard...
[00:58:19.150] - KAYLA BARNES-LENTZ
Or like it's too much or something. Yeah, I know. Totally. Too much.
[00:58:22.330] - HARPER CAROLL
Yeah. Okay. Actually, yeah, this is worth talking about because it ties to the brain. Okay, good. Okay. Neural networks are the models that are used to power systems like ChatGPT and large language models and image generation models. What they do is they are trained by taking input in the expected output and finding the equation that maps input into output. In the case of a large language model, the input would be the in words, technically tokens, but words before the next word, and the output would be the next word.
[00:59:07.200] - KAYLA BARNES-LENTZ
Okay.
[00:59:09.260] - HARPER CAROLL
Again, it finds this equation itself, and it does that by learning features in the text and the data that then indicate what the output should be. But it learns all those features on its own. In this sense, you could basically model everything with a mathematical equation. Any input can be mapped to an output with some sufficiently complex equation. That's why these models have billions of parameters in them, which is like many of those, that number when you hear it has 70 billion or 407 billion, that's how many, mostly how many variables there are. It allows the model to capture different nuances in the input data to then properly transform into the output data. These are just super complex equations that are capturing... These parameters are also similar to neurons. You can think of it like a math equation, or you can think of it like neurons. If you have someone with a million neurons, I'm sure they're way more than that, those neurons have to... You can capture less information than if you had like 100 million neurons or a hundred billion neurons, because those neurons can be specialized for different things. That's what these models do, is they learn to map input into output, and they learn how to do that.
[01:00:47.910] - HARPER CAROLL
They train these neurons that pick up on different features of the data that can then properly imply what the output should be. What's interesting is that in neuroscience, historically, they have modeled the brain by looking at neurons and interactions between neurons. But new neuroscience is now learning from a neural network with AI and modeling the brain mathematically. It's really interesting how they're playing into each other. But our brains also make our own mathematical equations. When we learn, we're also running gradient 100%, which is how AI models learn and tweak their equations to make them better at mapping input into output properly. It's just really interesting. We're similar to models in that way. Ultimately, they say math is the language of the universe, and that seems to be proving true.
[01:01:52.300] - KAYLA BARNES-LENTZ
Well, I love this, and I've loved our chat about AI today. Thanks to having me. Thank you. Thanks to having me. Of course, you got to come back. I'm excited once If you make more advancements in longevity, you'll have to keep us updated on how... I want people... Because I think it has the ability to really democratize functional medicine and access to high-quality information about health. Because right now, someone needs to go essentially to a clinic like mine, where we're practicing curative medicine because you're unfortunately just not going to... Your example was a great one, right? You go and they're not looking at what's actually going on. All doctors are incredibly intelligent and so on and so forth, but they're not taught to even read a lot of the labs that we offer, for example.
[01:02:37.120] - HARPER CAROLL
You're doing such great work, and the data that you are able to retrieve from your clients is going to help your clients in the long term and help people in the world. I love what you're doing. However, we can continue to collaborate, I'm happy to.
[01:02:56.570] - KAYLA BARNES-LENTZ
I love it. I love it. Well, thank you for coming on. Thanks, Kayla. Of Of course. This podcast is for informational purposes only, and views expressed on this podcast are not medical advice. This podcast, including Kayla Barnes, does not accept responsibility for any possible adverse effects from the use of the information contained herein. Opinions of their guests are their own, and this podcast does not endorse or accept responsibility for statements made by guests. This podcast does not make any representations or warranties about guest qualifications or credibility. Individuals on this podcast may have or indirect financial interest in products or services referred to herein. If you think you have a medical issue, consult a licensed physician.