Full Transcript 0:00 We're going to have some fun and I only have 20 minutes and just know I usually have about an hour. 0:04 So let's jump right in. 0:06 Let's set the tone. 0:10 Come on. 0:11 If we can up the volume. 0:13 That's one of my mission parameters. 0:17 I'll drive. 0:19 Trust me. 0:20 All right. 0:23 And let us just jump into the next one. 0:24 In the Next one, let make sure we get the audio up a little more. 0:31 As a patient advocate, I'm excited about the potential of AI in health care. 0:35 I enjoy seeing following Dr. 0:36 Castro on LinkedIn. 0:37 Accurate early disease detection, groundbreaking. 0:40 This technology promises a future where care is customized to each individual's unique health profile. 0:45 Keep up the good work, Dr Castro. 0:47 Yes, predictive diagnostics within reach. 0:50 Thanks to innovation in this field, personalized health is becoming accessible to all. 0:55 I'm happy to see that Dr. 0:56 Castro wrote the first book on chat, GBT and healthcare back in 2022. No more injuries undetected. 1:01 Hi, as an AI researcher, I collaborate with medical experts like Dr Castro to develop technologies that predict health risks, 1:09 personalized therapies and support surgical procedures. 1:13 Our aim is to leverage AI to not only improve health outcomes, but also to save lives. 1:18 VideoGen AI medical imaging transformed. 1:21 The new iPhone 18 will leverage small language models. 1:24 I am excited to see what applications Dr. 1:27 Castro will create. 1:28 He has several Apple Vision Pro applications out. 1:31 Yes. 1:32 Predictive diagnostics within reach guided by Dr Castro's expertise in AI and health care. 1:37 Now let's make sure all of us at this conference follow Dr Harvey Castro lead in revolutionizing health I am happy to see Dr. 1:44 Castro as the speaker today at this conference. 1:47 His insights are invaluable. 1:50 All right, well, we have to start with some fun, right? 1:53 All Right, so, and this is an AI conference, I can get away with this kind of stuff. 1:56 So every slide that you see has used AI. 1:59 This video that your about to see was just text to video, even down to some of the wording. 2:04 this image right here just shows you how many startups as of March of this year. 2:08 I mean, look at that, that's just ridiculous. 2:10 And these are all AI startups. 2:11 and so obviously we're seeing them in every industry. 2:14 i'm sure if I polled everybody here, were in all the industries and you guys can have heard of some AI in your industry or something coming. 2:21 There's that video animation. 2:22 I literally took this slide and I put it into AI and said, give me a video for this line and that's the image and the video. 2:29 As we move forward, what is the current state? 2:32 As a doctor, I'm really passionate about this because number one, there's not many people that are really pushing it as far as doctors. 2:39 So I am taking medicine and AI, and putting it together and saying, no, these are the ways that we can help our patients and let's do that now. 2:46 And so when I look at how we're able to increase our lifespan, increase out healthcare professionals, there's so many people that are wanting to retire. 2:55 Tons of nurses are like Harvey, I'm out, and I've done. 2:57 And I wanna give you some quick solutions of how can we make sure that our healthcare professional stay here for a longer time. 3:03 So this is a story about Alex. 3:05 Just curious, has anyone heard about this case last year, you guys? 3:08 Okay, good, not many, awesome. 3:10 Basically this poor child, imagine going to the doctor, 17 different doctors, to be exact, and none of the doctors gave you an answer to what was wrong 3:19 with you. 3:20 And imagine you had a little child named Alex, your little kid was like, Mom, I can't play, can go to the playground, my leg hurts, 3:28 so much pain. 3:29 I mean, how horrible that would be as a parent. 3:33 Then for three years, three year not knowing what's wrong your child, that really devastates me when I think about it. 3:42 Obviously, I'm about to say, ChatGPT gave her the answer. 3:47 She took all the medical data, all of the reports, the history and physical for 17 doctors, uploaded it to Chat GPT. 3:55 Now, granted, you're not supposed to do that, HIPAA and all that stuff, but play with me for a sec. 4:00 But then Chat GPTs said, You know what? 4:02 This could be tethered cord syndrome. 4:04 And then she went back to the neurosurgeon and said hey, chat GPTs told me this. 4:07 I almost died there by the second. 4:10 And then the bottom line, ChatGPT told her. 4:13 And I thought, man, what a story. 4:14 This is the one slide that I'm like, this is why I do this. 4:18 I think it's so amazing, and this why we need to do something. 4:21 As we move forward, another thing, I wish I had an English accent, like a really cool English. 4:28 So I cloned my face, made this video, said, you know what, here's my clone, read my slide. 4:33 What are the AI components in healthcare? 4:36 Definition, artificial intelligence, AI, simulation of human intelligence by machines. 4:43 Key components, machine learning, ML, algorithms that enable computers to learn from data and make decisions. 4:51 Natural language processing, NLP, allows machines to understand and respond to human language. 4:58 Computer vision, enables machines, to interpret and, make decision based on visual data. 5:04 Robotics, application of AI to perform tasks that typically require human intervention. 5:11 All right, well, I wish I had that guy's accent, but that's for another story. 5:14 In the interest of time, it's the same slide. 5:16 I'm just going to flip to the next one. 5:17 So everyone, not everyone. 5:19 Everyone in medicine has heard of Watson AI, and this was supposed to help us in oncology. 5:24 We're supposed come up with these algorithms. 5:26 And if a patient came to me, i was able to look at all of the data that is out there. 5:30 They would say, hey, doc, here's a treatment plan. 5:32 Unfortunately, they didn't come to anything. 5:34 If they ended up selling, then go any far. 5:36 But these other two companies, took that concept and I feel like they're doing a better job. 5:41 And I'm starting to see now signatures of now they are doing those things that Watson was hoping to do. 5:46 What's the difference? 5:48 Our GPUs, our power, I mean there's so many factors now that we can get things done. 5:52 So I am really excited to stuff like that. 5:55 Cleveland Clinic now claims that they can do predictive analytics, and they reduce readmissions by 12 percent. 6:02 Different hospital systems are doing this, New Yorktron, GatorTron. 6:06 What it is, is they just created their own large language model, let's call it ChachiB2 equivalent. 6:10 They took all the data of all of the patients that come to the hospital, then they said, okay, you know what? 6:16 You are exactly like this patient, but if I send you home, there's a 12-percent chance that you're going to come back, so let me keep you another night. 6:23 And I thought, man, what a way, because I have 20 years of clinical experience as an ER doctor. 6:29 But if you add AI, plus my clinical gestalt, and that's my big take is AI plus human is always going to beat the best human or the Best AI. 6:37 And so as we move forward, I see all these bots, all this fun stuff. 6:41 And here's Sarah. 6:42 I'm going to let her talk for a little bit. 6:44 Oh, darn it. 6:44 Maybe it's not in there. 6:46 But basically, she's a bot. 6:48 The World Health Organization has obviously prevented measures, and they want to make sure that you're not smoking, you have better mental health, 6:56 that your losing weight. 6:57 They keep updating it, I want say it in eight languages. 7:03 And this bot is 24-7 and you can ask these questions and go back and forth. 7:08 And so if the video had worked, it would just be her talking and explaining. 7:12 I thought, man, what a cool way to deliver medicine because now there might be some person that doesn't get that information, 7:18 but they may feel more comfortable having a conversation. 7:20 It's interactive. 7:21 You can interrupt her and talk and back-and-forth. 7:23 That's pretty cool. 7:24 And so Hippocratic AI is another one. 7:27 I know a lot of nurses were upset. 7:29 The news had the nurses out in California complaining about AI and what they did is they have all these different algorithms all the different nurses as 7:38 you can see nutrition pharma. 7:41 And then what they did is just help the doctor take a better intake of the nurse. 7:46 And so, it literally would call you and then you would have a conversation and go through your meds and it would be able to go back and forth and take 7:52 that in. 7:53 They're charging about $8 an hour and they're saying that this would help decrease our healthcare costs and would it help. 7:58 So, for better or worse, I think this isn't a bad idea. 8:00 I this is good because we need it. 8:03 The alternative is we're going to have some burnout health care professionals. 8:07 They may be short with you, they may not give you the time you need. 8:10 And so I'd rather have this for now to help us bridge to make sure that you're getting the best care. 8:15 And so as we start looking forward, I'm really big on communication. 8:20 I think this is the easiest low hanging fruit that we can do today. 8:23 ChatGPT, the ability to communicate is huge.I used to work at an airport and right next to an Airport and every disease in the world, 8:32 anyone that got sick in on the airplane would literally come to my hospital. 8:37 And I would have to take care of them. 8:38 And sometimes I didn't understand what they said, I don't know their culture, and I know many things. 8:43 I'd be so nice to have a bot that would help me translate. 8:46 Now I'll show you some quick demos of how I could translate with that person. 8:49 But take it to the next level. 8:51 What if that I'm in Texas and start talking about football, but what if they're European and say football? 8:56 They're like, oh, soccer. 8:57 No, no, we're talking two different languages. 9:00 If I can use AI to really help communicate and so for my patients I made this med helper because how many times have you gone to the doctor and you have 9:09 a horrible experience and then you finish you're like man I should have asked XYZ I said this oh my gosh I forgot so I make this bot. 9:19 that basically helps you figure out what questions to ask. 9:22 So when you're at the doctor's office, you already have a list of questions, and you know what to as, because you don't know, 9:26 what you do know. 9:28 But if you have AI to help you combine what your thinking with your disease process, now when we meet with the Doctor, your going to have better experience. 9:36 The average visit is only about 12 minutes with you Doctor unfortunately. 9:40 And so I thought, man, how do I use AI? 9:42 How do use it for my pediatric cases? 9:45 So here's another clone voice. 9:46 Correct inhaler technique for children. 9:49 Managing asthma triggered by sports. 9:52 Communicating asthma symptoms to adults. 9:55 And so what I did is I cloned my voice. 9:57 I took my discharge instructions for children. 10:00 And then I was able to break it down and then use the core concepts. 10:06 Then I pushed it into an AI to make a coloring book. 10:09 Now when I discharge a child, I can say here's your coloring and we can go through the principles. 10:13 So now I know that they're really understanding the disease process. 10:17 And so as we look forward, there's all this fun stuff. 10:20 I do it in a comic strip, I'd do in playbook. 10:23 GPT podcast is kind of fun, and I literally just clone my voice, with my own voice I can take any blog and have a fake conversation and within five minutes 10:32 I could push it out with me as if I did a podcast. 10:35 And it's me leveraging, then I said, what if did this for my patients? 10:38 What if just took my discharge instruction and pushed it with out voice? 10:42 And I'm going to show you an image of that in minute. 10:44 And so, this is a quick little plushie. 10:46 Tell me about this device. 10:48 And I wanted to check it out. 10:50 Hey, Grak, how are you? 10:55 Hello, Dr. 10:56 Catbro. 10:57 I'm feeling charged up and ready to explore with you. 11:01 Nice to meet you, what all can you do? 11:07 Nice meet to you too, Doctor CatBro. 11:14 All right, let's give me a joke real quick. 11:20 Why did the scarecrow win an award? 11:22 Because he was outstanding in his field. 11:25 All right. 11:26 So the skinnies, my business partner and I now are creating our own little plushie, and we're going to put it in clinical trials for kids that are having 11:33 trouble and help them walk them through their issue. 11:35 And they can have a conversation with the AI, go back and forth, better understand. 11:39 That lag time was bad, so we are improving the lagtime. 11:43 We're also helping with just the whole process where it's more fun voice. 11:48 It's a kid's voice, not so robotic. 11:51 And so as we move forward, this is interesting. 11:54 Microsoft released this a couple of months ago, and it blew my mind. 11:57 All it is is that image of the lady that you see on the bottom left plus an audio. 12:02 So it's just audio, so imagine if I dictated my discharge instruction and I just took an image and then watch. 12:15 Have you ever had, maybe you're in that place right now where you want to turn your life around and you know somewhere deep in your soul. 12:23 there could be some decisions that you have to make. 12:26 Like you know, like it's like things, something was decided for you. 12:30 And instead of trying to makes some. 12:33 It's all positive, but can you imagine? 12:35 That was just with an image and audio. 12:37 and it did the facial and the muscles and sounds real. 12:40 So Microsoft will not release this right now because of the elections. 12:43 But imagine if you put it in the hands of a doctor that could give you your discharge instructions that way. 12:49 Because how many times have you gone to the doctor and they said something and it's like Charlie Brown because after that you're like, 12:53 oh my God, they set X and I can't deal with it. 12:56 And then after you don't hear it, but what if I recorded it and use the eye and now I gave it to you? 13:02 And so like at two in the morning, be like oh yeah, now like I could go through this. 13:05 Now you saw the transcript. 13:07 That's such a better way to deliver medicine. 13:10 There's this cool thing called the pendant. 13:13 It's coming out in the fall, the multiple projects like this. 13:17 Just imagine a little tiny pendant that goes on you and then when you see the doctors transcribing. 13:22 So I won't bore you with my English accent, but that was... 13:24 Limitless pendant preserving conversations, lunch meetings, conferences, client calls or doctor visits, personal reflection, 13:32 verbal processing, thinking out loud, capturing a thought while your hands are busy, organizing the unexpected, So I'm going to keep going in the interest 13:40 of time. 13:41 So obviously, we've got to talk about ethics real quick, and I don't have that much time, but at the end of the day, I am really worried about my patients. 13:48 I worry about, especially patients that are outside of United States, that can just go to a pharmacy and say, hey, CHPT told me I have this disease and 13:55 it told I could take this drug and here I, give me the drug. 13:58 And I was like, no, were not doing that. 14:00 and so I made this algorithm for my patient, so give it to the hospital systems and just make sure that the human is in a loop and its not just any human, 14:06 its the right human. 14:07 In this case, it would be a doctor. 14:09 As we move forward, I'm really excited about AI's vision and we're going to start looking at things. 14:15 This is a company I was associated with and I am just going play the video. 14:18 There is no audio at all, but what it does is pretty amazing. 14:23 I have patients that are at risk for falls. 14:26 If they fall, especially the elderly, if they hit their head, they can break their hip, their mortality goes up and they unfortunately die or something bad. 14:34 And I thought if we could put a camera in the patient's room And then you're like, oh my god, camera, big brother. 14:38 No, it doesn't recognize their face. 14:40 It's just like that stick figure. 14:41 That's what it's seen. 14:42 And it does recognize that particular person. 14:44 But then it could send a text to the nurse or the doctor saying, hey, your patient is about to get out of bed. 14:50 Now, they haven't gone out a bed, there's still in bed but now it gives time for the AI to call the nerves, call a doctor, 14:56 and then the I can also talk to a patient and say, Ms. 14:58 Smith. 15:00 What's wrong? 15:00 Don't get out of bed. 15:01 And now we can use that. 15:03 So I just think this is just an amazing way and we're going to start seeing more and more of this technology everywhere. 15:20 So I truly think we're going to start seeing robots. 15:23 And I really believe that the vision of the robot is going hatch out to be five or six. 15:28 What it will have is it'll analyze my voice. 15:31 With my voices, it would say, that person's sugar is this. 15:34 That exists already. 15:36 It would hear a conversation between you and I, and the AI would, say he or she is depressed. 15:41 I can hear it. 15:42 She's at risk for Alzheimer. 15:44 He's a risk of Parkinson's. 15:45 I hear it in her voice. 15:47 And I said, you know what? 15:49 I hurt my back. 15:50 Can you help me with the patient? 15:51 And the robot's helping me lift the patients. 15:54 That sounds like a scary thought, but that is coming. 15:57 So I foresee that we'll be here by the end of next year. 16:00 And so as we move forward, we start looking at all these different technologies, and I'm really excited about the vision. 16:11 The meta-ray bands are these glasses. 16:13 And basically, I could take a picture of everybody like that. 16:16 I can hold this longer and take video while I am up here. 16:18 But what's really cool is I have a conversation. 16:23 If you ask me a hard question that I know, it will tell me in my ear and then I respond. 16:28 The other thing is if in medicine I put some funny rashes or weird things and say, hey, what am I looking at? 16:35 And I thought, man, What a cool tool if I start using this in health care. 16:39 The Rabbit R1 has really bad reviews, but at the end of the day, I see it more as a proof of concept. 16:45 You know, this thing is paired to mid-journey. 16:46 If I need some slides, literally say, hey, need a slide on X, Y, and Z, it'll do it, then I can transfer to these slides. 16:52 This is pair to Uber, so if I get a lot of bad looks, say hey get me an Uber and it will find me and get an uber out there, 16:59 that'll be right there. 17:00 And so this is pretty cool stuff. 17:02 In healthcare, as an ER doctor, take a quick EKG, analyze it and send it to all the people that need to know. 17:09 I'm wearing the humane AI pin right now. 17:11 I won't do the demo, but the skinny is I can hold it and translate, and in real time it would translate. 17:16 So imagine if I am at a pinch and I don't know XYZ language and can just go back and forth. 17:22 These are the things that are here. 17:24 the small language model, it'll be on your phone. 17:27 Imagine if you're in the Grand Canyon, someone kills over and you like, oh my God, I don't even know CPR. 17:32 But now your iPhone has all that information in, the phone and can look at you and say you are going too fast, you going to slow. 17:39 And then you point to the satellite and then your able to get a helicopter to save that life. 17:43 This will be here. 17:45 So as we move forward, in residency the last thing was the Apple Vision Pro. 17:51 When I was in residency, it was see one, do one and teach one. 17:53 They would be like, Harvey, intubate. 17:56 And I'm like oh my god, I just saw my first intibation. 17:58 Like no, now it's your turn. 18:01 Oh my gosh, i read about it. 18:02 I saw this guy do it, but what if I had an Apple Vision Pro where I could intobate a hundred times before I touch the patient? 18:09 How much better would I be? 18:11 And so these are the kind of fun stuff that's coming that I feel that it's the future. 18:15 Now, one of my friends sent me this slide. 18:16 The skinny is this. 18:17 Make sure you follow people on the right, the left, bottom right. 18:21 Because there's only there is no one person that knows it all. 18:23 So I really encourage you. 18:24 the more you learn, come to these conferences, The more, you can communicate with each other. 18:29 Obviously reach out to me and feel free to hit me up. 18:31 This is where I can tell me about A.I. 18:33 and health care. 18:37 Oh, this is a fascinating area. 18:40 AI is making waves in health care. 18:42 And I won't play it all because I know I got to run. 18:44 But the skinny is this that I'll give you the link. 18:48 It's basically it's a podcast. 18:49 In the next drive you're doing, you can ask the podcast, hey, what's going on? 18:53 Say his news. 18:54 Hey, What's Going On specifically to my industry in this? 18:56 And it literally will go to Google, extract the question, and it'll go back and forth. 19:00 You can interrupt that. 19:02 So it's pretty cool. 19:03 I made these quick GPTs. 19:05 Feel free to hit me up. 19:05 And on the last slide, I have a QR code, and you can get them for free. 19:08 They're just how to use GPT's in health care. 19:11 Then I cloned myself. 19:12 All my books, all my interviews, on my TV, everything I've done, made my own digital twin. 19:18 That little wand that is at the bottom right, right there, that's my digital twins. 19:23 So you could literally say, what does Harvey think about AI in healthcare? 19:25 What does he think of ethics? 19:27 What does he think about welding? 19:28 And it'll say he's never talked about Welding. 19:30 So I don't know. 19:31 Just so you know, try it out. 19:33 And then there's my QR code. 19:35 Feel free to scan it. 19:37 If you want a copy of these slides, it will take you to a form and you can download it and then this is... 19:42 Hasta la vista, baby. 19:46 I'll be back.
Comments