Transcript 0:00 Technologists, you know, how they can use technology, how-- what are the new areas that they can start investing in-- their effort and money into, in the healthcare- Publishing groundbreaking information of how AI is, uh, impacting healthcare across three, I would say, broad areas. 0:20 [ on-hold music] Hello, everyone. Welcome to one more episode of Unlocking the AI Advantage podcast. 0:31 As you know, we have been focusing on AI in health tech as well, in addition to AI broadly, how it's impacting. So today's, uh, podcast episode is going to be on health tech. 0:45 And, uh, by the way, if you have not subscribed to our newsletter, AI Health Tech Insider newsletter, there will be a link during this podcast that I will actually mention for audio listeners, but in the YouTube, uh, you will have it in the description. 0:59 Okay? So, uh, we've been, uh, publishing groundbreaking information of h-how AI is, uh, impacting healthcare across three, I would say, broad areas. So number one is drug discovery, right? 1:15 And disease dis-de-detection, right? H- AI, as you know, it's, it's one of those areas where new drugs are being discovered faster because of AI and machine learning are helping. And secondly, 1:28 it is also helping, uh, to detect the diseases much earlier than they would, uh, otherwise find the diseases out. So that's one area, so that I'll be covering. 1:41 The second area is the personalized healthcare platform, right? So AI-- how AI is helping individuals like you and me, uh, to help manage our own, uh, you know, healthcare without having to go to hospitals or clinics. 1:54 And there are platforms out there that are coming up. Um, so it's, it's lots of advances have taken place. And we are focusing on those kinds of areas in our newsletter. 2:04 And then I'll be go-going over some of those in the podcast as well. Okay? 2:09 The third area is robotics and surgery, where, uh, AI and robots, um, are actually, uh, you know, uh, coming up with a precise application, precise, uh, you know, surgery to treat some of these, uh, you know, uh, diseases. 2:26 So I'll be going over some of the advances. And, uh, now we will be navigating between the actual practical applications where it's already happening, where it's, uh, the-- in the field, these things are happening. 2:40 And then secondly, also where the research is happening, where advanced research has shown a lot of progress in those areas. 2:46 But again, um, as with any technology, as with AI, and especially with the healthcare, uh, so caveat emptor, buyers beware. So at the end of the day, you have to make your own decisions what's right for you. 3:00 But, uh, this particular episode, uh, will be for technologists, you know, how they can use technology, how-- what are the new areas that they can start investing i-their effort and money into, in the healthcare, which areas are showing progress and... 3:14 But also to, uh, individuals like you and me, how we can use, uh, this, some of the health tech for our own benefit, right? So it's, uh, it's-- it, uh, transcends, uh, all those areas. 3:26 So without, uh, much further ado, so let's go into the newsletter and some of the areas that we've been covering more recently. Okay? 3:33 So for people who are watching on YouTube, definitely you can, uh, get the screenshots. And for people who are, um, only listening on the podcast, I will be going over this information as well. 3:44 So let me start, uh, sharing some of the, you know, uh, information that we have here. Okay. So now, uh, so let me just go into two, uh, recent episodes that we have published. So this particular one 4:00 is, uh, um, it's, uh, issue number ten, uh, AI Health Tech Insider. And, uh, so you can go into the aientrepreneurs.com and subscribe to newsletters. 4:12 You will get, uh, both, uh, AI Health Tech as well as, uh, um, you know, regular standard AI newsletter, uh, the Entrepreneurs newsletter as well. 4:21 So in this particular episode, the very first thing we covered is a new hope for prostate cancer patients, right? So prostate cancer remains a significant health challenge worldwide. 4:33 And one of the areas, uh, it's androgen deprivation therapy, ADT, has shown promise. But, uh, you know, it's a long-term, it can lead to, uh, uh, what's called a CRPC and NEPC. 4:47 Uh, so some of those ar-me-technical terms, but let's not worry about them. 4:51 But, uh, how this particular P three hundred, you know, plays a cru-a crucial role in promoting on-oncogenic signaling pathways in various cancers. 5:01 And this is where AI has shown, um, AI-assisted design has shown a lot of progress here. Okay? Effective degradation of P three hundred. So, uh, this particular, uh, drug showed cancer cell killing capabilities. Okay? 5:16 This is-- this follows into the, uh, how AI is helping with, uh, you know, um, drug discovery. 5:24 In this particular case, a particular method of treating cancer that it is showing a lot of progress, and then AI has helped, uh, uncover, or, you know, uh, th-this particular, uh, method. Okay? 5:36 Tumor suppression in, on mouse models. So what are the implications of this particular one? So this development of, of this AI-driven drug represents a significant advancement in prostate cancer therapy. Okay? 5:48 So that is, uh, one area.And, uh, the one that-- another, uh, issue that we're, um, publishing number eleven, so that actually coming out, uh, as we speak. 5:59 So-- And this particular one is a second disease detection that we are covering, which is a AI tongue disease detection, and it, uh, uh, with this particular model is showing a ninety-six point six percent accuracy in real-time diagnostics. 6:14 Okay? So this is a Middle Technical University in, in, uh, in Iraq and University of South Australasia. 6:20 So they, uh, they use some of the machine learning, and supposedly this particular method of using tongue color, uh, is from the traditional Chinese medicine, right? 6:30 And an innovative imaging system trained five thousand two hundred and sixty images and tested with sixty real-time patient images. Okay? So let's find out how, you know, what happens here. 6:42 The study utilized machine learning algorithms, including XGBoost, SVM, and KNN that we know, right? People in the AI space know these algorithms, to classify tongue colors under different lighting conditions. Okay? 6:55 And XGBoost emerged as the top performer because many times you're trying to, uh, you, uh, figure out which algorithm is best, and in this case, XGBoost emerged as the top performer with an accuracy of ninety-eight point seven one percent, right? 7:10 The system was trained in five thousand two hundred and sixty, and people who in the AI space know that the data is, you know, very crucial in terms of what kind of data was used, uh, to train the models and, and then if it's, uh, only comes from certain segments of the population or it's a certain, uh, um, you know, age or sex-related, related, uh, patients, then, you know, it can be skewed, it can be biased. 7:34 So the data has to, uh, pretty large-- It has to be pretty large, and it has to, uh, b-be, uh, representative of vast, uh, you know, areas of population. Otherwise, you know, you would have skewed or biased results. 7:48 So in this case, five thousand two hundred and sixty images were there, and it's a real-time diagnostics. So it's, um, uh, it's used the MATLAB-based G, uh, you know, GUI. 7:58 And so it can, it can be used if it's extended diseases like diabetes and asthma as well. But this particular one, um, you know, use the, uh, tongue, the colors to figure out. Okay? 8:11 So, um, so what happened here, let's say in the picture, uh, let me just, uh, verbalize it. So there is a image processing that it's going on, and then there's a trained model, right? 8:21 It's using five thousand two sixty images that we talked about. And, uh, um, and the disease are predicted, uh, with certain accuracy, right? At least in the training space, and then this can be app, uh, applied. 8:34 It, uh, has a potential to revolutionize non-invasive medical diagnostics, right? 8:39 Again, remember our caveat emptor, which is this particular method does, you know, it doesn't mean that it is applicable to-- for everybody. It, it comes from a tr-traditional Chinese, uh, method. 8:52 So people who kind of practice or believe in that one, you know, can use this one. Okay? So now those are two examples of disease detection, drug discovery area, how AI is helping. 9:05 One was the prostate cancer, one was using the tongue colors, uh, you know, how to, uh, uh, diagnose diseases with certain, uh, accuracy. Okay? 9:16 So now let's talk about, uh, uh, what are the platforms that are out there, uh, that can help, uh, you know, manage our own, uh, issues or healthcare kind of thing, right? 9:27 So, uh, there is a AI hybrid survival assessment. This is for heart failure and renal dysfunctional patients. Um, in this one, um, this particular model, they used seventeen or so. 9:40 So they used twelve AI algorithms, right? And for accurate survival, uh, predictions, this one uses combination of many algorithms. Okay? So twelve AI algorithms, a hybrid AI model. 9:51 Uh, again, it utilizes data from over one thousand seven hundred patients across multiple cohorts. 9:57 Whether it's one thousand seven hundred is the representative sample or not, to be seen, uh, but as we keep, uh, refining this particular model, you will have more patients. 10:07 But, uh, the more the merrier in this case to train the model, which of course comes with costs as well. Okay? So that's, uh, that's one, uh, platform that we came across that we put it in the newsletter. 10:20 So-- And then, uh, let's see. 10:21 In issue number, uh, eleven, um, so we covered in the tools and technologies in the same platform area, medical imaging and research with the world's most powerful all-in-one Aleph Alpha IVAS, right? 10:34 So this is a, a high-performance personal supercomputer aimed at revolution AI-driven healthcare, right? 10:40 So it, uh, uh, you know, it addresses complex medical applications such as digital pathology, genetics research, and cardiac imaging, and it accelerates digital pathology, right? So it's 10:53 ability to figure out, uh, healthcare issues, uh, much sooner than, um, you know, it was taken before, right? So processes whole slide scans up to hundred times faster, right? 11:04 Enabling rapid pixel classification and cell segmentation. Right? So that's all. Um, it's about making it faster, um, with certain, uh, accuracy. Okay? So those are the areas and the, and the tools and technologies. 11:19 And the last section that I will cover is, you know, the broadly categorizing as, uh, so robots or surgeries or, you know, where devices are, uh, you know, playing, uh, very well i-in healthcare. Okay? 11:36 So this particular one is a Neuralink. We all know about Neuralink. This is Elon Musk's company. And, uh, some time ago, um, they have done the first implant, and then this recent one is a second brain implant, right? 11:51 What is the progress and what's potential? 11:53 So let me cover this.It's a second brain-computer interface, BCI implant, marking a significant milestone in the company's journey to create a symbiosis between humans and artificial intelligence, right? 12:06 This development has sparked both excitement and questions, of course. Uh, how effective these are, uh, the ethical implications, right? So those things will be questioned, asked. 12:16 So what are the key points of this Neuralink's second BCI brain, uh, computer, uh, interface? Okay. 12:24 Um, so successfully implanted in a second individual, um, where the first one there were some setbacks, but they've learned from the setbacks, the second one. 12:33 And, uh, it refined its recording algorithm and surgical technique to avoid issues like air pockets during the first implant that they had. So they fixed those issues. 12:43 And the, the BCI telepathy features a coin-sized hub in the skull with threads in the motor con-cortex to record neural activity, transmitting signals to an external device via Bluetooth, right? 12:58 So this is where there's a device inside, and then it's actually transmitting, uh, the real-time information via Bluetooth. 13:05 Okay, so this is-- You can call it telepathy, but, uh, this is a, a brain-computer interface telepathy. Okay? 13:12 So those kinds of things and, uh, s-and-a-a, as the second version, uh, implant, uh, definitely made, uh, enhancements over the first one and fixed some of the issues that they found out. 13:25 But also they're doing, uh, you know, telepathy, where the-- they can record the data, uh, that's being transmitted from the brain to an external recording system. Okay? So that's about, uh, Neuralink. 13:37 And, um, uh, let me see one other point I want to say. Uh, bio, uh, implantable BCI. It's, uh, fully implantable and cosmetically invisible, so it's not there. Uh, it's-- It cannot be seen. Uh, biocompatible enclosure. 13:51 Uh, the N1 implant is hermetically sealed in a biocompatible encla-enclosure, and the implant is powered by a small battery charged wirelessly from the outside, making it convenient for everyday use. 14:03 Advanced chips and threads. Custom low-power chips process neural signals. And surgical robot. Threads are inserted by a surgical robot with precision, ensuring minimal damage during implantation, right? 14:17 So it's combining not just the device, but the ability to do implant the device also, uh, using a surgical robot. So that's, uh, uh, new, uh, Neuralink. 14:27 And in this category, let me also talk about the AI and research personalized brain health insights with the BrainAge UI platform. Right? So, uh, Diagnomed Holdings has introduced a BrainAge AI platform. 14:43 Uh, it's a groundbreaking tool designed to assess and monitor brain health, particularly in relation to the impact of weight loss drugs like, you know, semaglutide. 14:53 So this particular one is checking how the brain is functioning and the brain health, monitoring and assessing that in, in this particular, uh, platform. Okay? And again, it's a personalized treatment. 15:06 It's, it's meant for each individual. An AI-powered assessment, as we talked about. And, uh, there's also clinical validation that's happening. It's proving effective in large scale screening. 15:16 Um, so we'll see how, uh, it progresses further. The last one I'll be covering today is the OpenAI and Thrive Global's personal AI health coach. 15:28 The collaboration between OpenAI and Thrive Global, um, is set to transform personal health management through the development of an AI health coach, right? 15:37 So in this area also, we are seeing lots of, uh, progress where personalized health coaches. Okay? 15:46 And then so OpenAI has lots of information, lots of data that they can leverage, and just like a GPTs are there, but now applied to he-uh, healthcare. So now I have my own personalized health management, right? 16:00 The AI health coach offers tailored advice in five key areas of, you know, sleep, nutrition, fitness, stress management, and social connection, right? Enabling users to make informed health-conscious decisions. 16:15 And again, so there's a lot of data out, out there, so data-driven insights. And, uh, Thrive AI Health, right? 16:21 So led by DeCarlos Love, it collaborations-- uh, collaborates with institutions like Stanford Medicine and Rockefeller Neuroscience Institute and of course, OpenAI in this case, to put this plant, uh, uh, this kind of a platform out there. 16:36 Okay? So that's that. And the way, uh, in this category, the l-the last one we'll talk about is the Med-Gemini. So we all know about Gemini. Okay? Google researchers and DeepMinds have the Gemini platform. 16:50 Now they have unveiled a healthcare-focused one called Med-Gemini, a powerful suite of AI models trained for medical applications. Okay? 17:02 These advanced models excel in processing complex medical data, delivering state-of-the-art results in radiology, pathology, genomics, and more. 17:12 So if you start combining these things, like we saw the personal health ca-- health coach that OpenAI is doing, uh, working with Thrive Global, now you're seeing Google is putting out a Med-Gemini, right? 17:26 So it's a, it's a suite of AI models trained for medical applications. 17:31 Some other startups, some other companies can use this AI, this Med-Gemini platform to, uh, focus on solutions in, uh, uh, these areas: radiology, pathology, genomics, and more. Okay? 17:46 So those are the kinds of things that we are coming up with. Um, and so subscribe to the newsletter. 17:53 It's a health, AI Health Tech News-Insider newsletter, and of course, AI, uh, Entrepre-AI Entrepreneurs N-Newsletter, both of them. 18:03 So one is a very broad applications of AI, the AI Entrepreneurs Newsletter, and the AI Health Tech Insider is a very healthcare focused on where AI is making impact, right? So that's what that is. 18:16 So I hope, uh, so this helped you and, uh, so we will be switching back and forth between, uh, healthcare and the broad applications of AI in, uh, many areas. So I definitely would like to, you know, hear back from you. 18:31 So if you have any questions, you know, let me know. And this is Ramesh Danta signing off, uh, for today. Take care. [upbeat music]