AI in Healthcare Series: Empowering Patients with Kimberly Powell, NVIDIA

0:15 Welcome back to the Stanford Healthcare AI podcast. We're so excited to be joined by Kimberly Powell, who's been leading healthcare efforts at NVIDIA for almost the last two decades. Uh we'll get into a lot more of what that means in a second. Uh but unless you're living under a rock, Nvidia is the largest company ever today, over 4 trillion in market cap. And we're just so thrilled to have Kimberly joining us today. >> Thanks so much for having me, Justin Matt. I'm excited to be here.

0:45 >> Well, I'll I'll signpost a few of the topics. We'll we'll go through uh we'll talk through some of just the recent AI developments. We'll talk through a few of the different things we're seeing around the world on chat with EHR and then excited to kind of get into some of the robotics angle which we have yet to talk about whatsoever. Um, but I wanted to start with this familiar topic around search and how people are starting to use these tools. Matt, can you tell us what what we're looking at here?

1:14 >> Well, yeah, I mean it's interesting. I think um you know a lot of my clinician colleagues are also seeing kind of similar trends where you know again there was that and we've talked about this on the show before but this is one of those threads that I think will continue to come up which is you know are how are our patients uh using the models and um and it's almost like you're starting to see a lot more sophistication on on the part of the patient in terms of having had a conversation with one of the frontier models and their healthcare data or their diagnosis or their treatment or whatever. ever and and it's actually been in some cases really refreshing uh feeling like I can have a a conversation with my my patient and that they have a lot more uh background and and frankly have been spending a lot of time working with the models explaining some of the concepts but at the same time it it does I I think this is a good way to look at maybe that trend from a different perspective because I think historically it was always you know you go to a search engine you start Q&A that then you go through a bunch of links you go down some rabbit hole, right? And and then you I I feel like a lot of times, at least historically in the pregbpt days, you're spending a lot of time in the clinic visit sort of explaining away some of those problematic rabbit holes, right? Because, you know, the model there's no model behind it and and maybe they didn't have the context or maybe even the the literacy, frankly, to to sort of navigate some of these really complicated uh sites orformational resources, etc. And so I guess this trend I listen does is this, you know, is this definitely because of the models, you know, hard to say it's kind of an interesting coincidence if it isn't. Um I think more and more folks I don't know if if Kim you you've I mean I ask this all the time to my friends too, like are you are you putting your data in? Are you putting your loved one? I mean because I literally do that with like with my own healthcare data and it's it's fascinating like you know um and I learned some things too.

3:07 >> Yeah. I mean, listen, it's it's totally it's has some similarities. I I uh watched the show where you guys talked a lot about like when when you had the power of the internet both as patient and and clinician, how it had to change the way that you work, right? Talk the patients out of all the things that they read. What is so different now though with these large language models um is the introduction of reasoning, right?

3:30 It's not just the memorization of anything that might have happened with some adverse, you know, effect of a drug. It's it's listening to you. It's it's it's remembering things about you. Um in fact, just this weekend, um you know, in our own family situation, it uh asks you questions and and says like, "Oh, well, you were you were up in Maine. Is it possible that you were, you know, uh bit by a a tick and that maybe you should screen for Lyme disease?"

3:58 Right? just so the fact that it can reason, the fact that it has memory and context about you as a patient, I think that's what's going to be kind of gamechanging about this. So hopefully the models are going to help um not have to do so much explaining away and actually empower the patient a lot to just all that context that we take for granted as patients but is such powerful information for doctors but you can't always remember it when you're in that 30 minute visit and you're you're feeling nervous and anxious and all these other things you know chat can help maintain all of that memory for you. I think it's going to be really stunning um a stunning experience.

4:39 >> I think what's also totally agree and I think what's also fascinating is when you compare this you know and obviously there's many many things going on here and uh you know it's hard to isolate one thing but when you compare the health queries to everything else the health queries are the most down. >> Yeah. >> Right. And and you know that's fascinating to me and like what you know we were actually laughing before we came on you know I still used a Google search like oh my gosh what what what a lite you know I'm not using you know technology for everything but I do use search for something still I kind of know I shouldn't but I'm used to it but for health queries I'll never touch Google first right it will definitely go to a model to go give more context ask more questions go down a rabbit hole and so it does feel different at least for for me personally for from a healthcare standpoint and then um you know compared to other things that I haven't fully changed changed my workflow yet on for health questions it's by far uh the first place for for me to go.

5:40 >> Absolutely. >> Yeah. Yeah. And I think part of it too is like what I would like to see hopefully in the future is a correlation also to another chart that shows like uh interest in preventative health uh you know routine screenings are up. Like I would love to see that this lead to a carryover effect to potentially, you know, I guess health literacy at large, you know, in this democratization of I guess the the care management journey like cuz ultimately we can't really experience any of the true outcomes until this causes behavior change. And as we know, that's that's hard to do in a 15-minute clinic visit or whatever.

6:15 and and you often wait for these episodic crises of care as opposed to, hey, I'd love to just talk about like let's plan out the next, you know, 6 months, 2 years, whatever. Um I I I'm optimistic though that like just that literacy and the education will lead to some of those other behaviors. And um I guess um we'll have to stay tuned to find out. So I mean you talked actually Kim you brought up context and you know asking for more information and you know were you you know traveling in Maine and I was you know kind of laughing thinking through you know this is like asking you know our infectious disease consult where these are the clinicians that spend more time with the patient go far deeper you know pull all these things together. Uh, one of the interesting things we're also starting to see now in a few different places is people trying to bring more context to models working in systems. And so the examples I'm starting to touch on are, you know, these different, you know, with EHR pieces where, hey, rather than having the patient or the clinician have to feed everything to that subsequent model, can we bring some of this data almost directly to the models to kind of query in different ways? And so there's a few studies that have come out here o over over the past and actually Matt Matt yeah what what did you want to highlight here? Well, I I think this is that chat ehr that I think you know kind of made quite a few headlines uh from from the work out of Stanford. But the but the basic gist of it is you know systems are setting up well so first of all systems are seeing the trend where clinicians and other healthcare you know folks in their healthcare workforce are realizing that there's some real advantages to using the models for the things that they do. Uh that's a obviously not necessarily a compliant there that's a compliancy risk. There's all the things we've talked about in prior episodes around that problem of folks just using the the open API or web app on their phone for example in the healthcare setting. So then health systems are then setting up alternative ways to access frontier models that are more of a you know safe compliance secure. Okay, that's but that still requires like cutting and pasting and to your point Justin like the bringing the context over from the records or whatever or having to put more information. Now it's like why don't we just connect it and in a lot of cases it's a smart on fire just connect it directly into the HR pull in the data that's relevant and I can ask summarization questions or uh maybe you know best tests or potentially reference guidelines or whatever those things are.

8:43 I think that um it does raise an interesting point because like there's a couple articles well the one on the right is a is from a group out of Europe where there's additional regulatory considerations that require in this case required them to leverage local models and I think we are seeing when we talk about the prior frontier and the capabilities of the models that uh models are getting smaller models are getting faster models are getting cheaper and models are increasingly run in local environments and so Kim I mean clearly Nvidia is leading the world in so many areas but I think one of the big ones is uh efficient running of frontier models distillation of the models on edge deployments and all the advantages there too that I think uh particularly in healthcare where there's so much you know concern around privacy and and obviously restrictions around use like this this could be a really powerful use case I don't know if you're seeing a similar thing where like folks are like hey I'd love to deploy it locally in this area uh but I need the horsepower to run it or I need to connect it a certain way. I I don't know if you have a sense of that trend, but but it's definitely something we're seeing.

9:51 >> Yeah. I mean, we're we're interested to enable these models to run everywhere. You could imagine, right? I mean, you want it to run inside your car. You you would want it to run inside your operating room, right, Matt? like you you don't want to have to ask a nurse or yourself disengage a surgery ever to maybe pull up um you know a prior um piece of information whether that be um you know a an image that you took of the patient prior to the surgery or um any you know potential coorbidities and things that might you know get tricky when your your surgery's getting more lengthy. You just want to be able to talk, you know, to an something that's listening ambiently and provide you that very relevant, very uh important real-time information. So, having models be able to just live on edge is it's a necessary condition. Um, but it's just like, you know, models run in the car, but you're going to phone home to more capable models potentially that um need to do other type of tasks that that run in the cloud. And so we absolutely see a a complete hybrid and you know being an accelerated computing company we're just always been passionate focused and um taking the most important computing technology and fitting it into the tiniest compute footprint. one so it's accessible two so it can as you say run on the edge some of the most important and and I would say you know uh risky applications are these edge applications because there's usually life at risk and if it's life at risk you need to have the models to be so close to the humans that it's interacting with um so I think there's there's reason that we have to be really focused on um taking this amazing capability and putting it in tiny packages and and running it everywhere Yeah, I I totally agree. And it's interesting, you know, thinking about autonomous cars, other places where really life's at risk. Things really have moved to the edge.

11:49 If I look back though, where we're at in healthcare today, I would say, you know, we're we're early. We're very I one, we don't have autonomous surgeries going on or anything like that yet. Most of our problems we have, you know, uh stopped. So difference just for people if you're driving a car you can't just slam on the brake if you lose internet connection you know that's real bad for a chat with EHR function yes it's not good if you don't have that data but you're not in the middle you know hopefully of an operation or something where you're depending on this data I guess Tim where where do you see this starting because if I grade us right now you know we're still very very early so where do you see you know because you're going to see you know the literal frontier of this where do you starting of things getting pulled to the edge and kind of moving that way in healthcare.

12:38 >> Yeah. And if you know let's let's think about what uh what kind of agents there are or or you know use of utility of agents in healthcare. Um we think of digital agents they could be vision agents. They could be you know ambient listening agents. They could be information retrieval agents. And so you know and and as you say it we're connected with uh it's approaching 5,000 AI startup companies uh in healthcare alone here at NVIDIA. So we do see quite a bit but vision you know as Fay at Stanford um pioneered for for us or helped pioneer for for the whole world vision is a huge aspect and I think we're we're not yet taking full advantage of it in the healthcare space but it's going to be an absolute necessary condition. Um, and then the ability now with being able to do speech uh recognition uh we're doing crazy amount of awesome research um with uh fantastic companies like a bridge where um you know you need to know exactly who's speaking um who they are the context of that person because you're digitizing that language and you're going to turn it into some kind of action right these these tokens of speech turn into some kind of task that we have to end up going to do. So you need to know if the patient said that, if the nurse said that, if the doctor said that, if the, you know, um and so all of that has to be, you know, really really looked after, but I think that um you know, vision systems are are coming into play now because there's a lot of fantastic efficiency you can have just by being able to track things in the hospital or um timing of surgeries or um you know, being able to uh see you know, the flow of of patients uh within hospitals and um and then connecting um a lot of that vision. Now you know that the trend of vision language models that have reasoning in fact um this is where you know the vision systems of the past where they might have um been a little tricky for um for nursing stations because they would have overalerted perhaps right and caused actually potentially more work um on the staff than than you would want. Um, I believe that these reasoning uh systems and and chain of thought where you can really attach the agents to how that hospital runs in in uh very particular ways. We're going to be able to really reduce some of that uh overhead that came with the first generation of of vision systems for example. Um and then you know digital agents that uh then combine into um as I said just we were talking very early on how you know why are why are the health searches going down because you can just have a conversation with your phone now as a patient and you're getting prompted and things like that. And so the fact that um we can now do that ambient listening in a healthcare setting. Oh my gosh.

15:27 Right. I can't I get so excited about the possibilities because it's not just the thing that was written down. It's all that other context that's being digitized both by the doctors and the nurses exchange and the patient exchange. Um all of that can facilitate just a massive amount of of opportunity. So vision systems um ambient listening systems and then the ability to attach them uh to what is otherwise really messy healthcare IT systems. Um we're in a whole new category of possibilities now. uh world is the oyster and that's why we see just the startup community and the AI applications that are really domain specific really vertical for healthcare um being wildly successful you know 100 to 200 million annual reoccurring revenue startup companies in the matter of 12 to 24 months is absolutely unheard of in this industry it's just been you know they used to be fraught with barriers to entry um whether that was computational footprint whether that was safety considerations, whether that was the ability for clinicians or nurses to adopt it. You know, here the adoption is human language, speaking to a phone, they're all familiar with it, right? There's no learning a new UI. The UI is speech or the UI is a camera just, you know, watching over me or, you know, so there the the barriers are really coming down uh super fast. That's super exciting.

17:00 >> Yeah. I mean I I as you know me I this is so true in terms of the multimodal aspect because I we always say like there's so much that you can do. We've we we've shown this in prior you know episodes of different data around the performance of these models and like text traditional textbased tasks and there's a lot still to do but when you get the multimodal to your point that unlocks an entirely new surface area.

17:21 there's so much more information that we process just as human clinicians with our eyes and our senses and and and we know that there's information being transmitted that may not be captured in language uh and and are do we have the infrastructure of the low latency systems I I think we're getting there right and and in terms of the computational capabilities but then also you know ju to use the voice example just as a different modality just on the voice biomarkers right we're seeing startups that can take just the audio data and have a very high uh prediction rate for you know neurodeenerative disease or depression or other things and it just adds these layers of more opportunity for us and I think some of it you might ask like an old-timer clinician who could probably walk in a room and tell it like within two minutes what's going on maybe before anyone speaks a word and because there is there's something to your point about the vision aspect to the to the sound to the anyway I I'm I'm just as optimistic as you are that we're scratching the surface and then I think on your point about the alerts, I think to me like this is where that abstraction layer with an agent that understands context, you can interact with it almost like quote unquote like a human, but can can feed the or use those tools as uh independent systems in the context of the situation and help filter out some of the potentially false positives or alert fatigue that ends up going on when you have, you know, what, like you said, like dozens of these running and they're just kind of like being triggered by a simple event as opposed to a contextual awareness. I think there's a lot of work there that um that I'm seeing that's really exciting. Again, that t taps into the idea of agents. But you you all have really thought a lot about agents. And I think what I like about the approach Nvidia's taken is it doesn't just stop with the digital world. You know, you you really are pushing into the, you know, physical world, I think. And and I don't really know of another tech company that's thought quite as much about the robotic space. And then by the way, can we create a developer ecosystem where folks can build on that no matter what the vertical? But I obviously I'm biased. I think healthcare robotics is going to explode in the next 5 years.

19:30 And you know, I I don't know what you're saying, but I just feel like there's so much opportunity to marry the context, the multimodal with now interactions in the physical world. And that could be as simple as you know supply chain and you know back you know back office type things but it can also be literally in the operating room like I you know I I don't know where this is going to go if you if you have a sense but it seems really exciting.

19:54 >> Yeah we're we're more than excited about it. I mean a lot of the stuff that we were just talking about you could just go so far as to say every hospital is going to be a robot. You know it it is it is a 3D it's a 3D space. It's a physical space. You have to understand the 3D world to do all of the things that you just described to really um have the appropriate context. The distance between things um matters. The speed at which things are traveling matters. Um the the size and volume of things matters, right? And so having that that physical understanding of the world is what will enable the future of of what I'd say an embodiment of an AI hospital and you're it's going to be a necessary condition um in the future state of of surgical surgical robotics because you if you think about this as a multiscale problem you know on the one hand you want to have you know your hospital be a robot so that efficiency like you were talking about a lot of back office and a lot of efficiency where you know obviously is absolutely necessary. We're we're tens of millions of health care professionals short uh of the demand of healthare. So we have to do something even at the hospital environment level to offload any kind of nonclinical work. It it it's just a necessary condition for us to be able to serve uh the population. But then you get into okay um you know like we were talking about the in the in the doctor's office, right? There's that office is kind of a robot in itself listening, observing, capturing um and then you know moving into a more um intense environment where you have to actually receive treatment that could be the patient room where you might have lots of monitoring systems um and and then all the way into uh where it's very very very complicated into the operating room and even there right so the operating room itself is a robot the surgical device and all of the other devices surrounding that patient is going to be robotic.

22:01 And then the human we are going to want to simulate every individual patient's anatomy and understand its essential um you know physicality and and functionality and that is essentially you know living in a in a simulation environment. So, you know, at the atomic, you know, scale, um, so that we can teach these robots how to work. Um, and so we're we're super excited about the physical AI space. Um, we've created, um, what we call the three computer platform, which is, um, truly accelerating this. There's a reason why robots kind of were steady state in not being able to be super useful and that's because we didn't have the third computer which is essentially simulation and understanding um the physical world and all physical things in the future will be born in a computer first. You won't make the robot then try to teach it in the real world because as you can see it takes decades to try to generate enough training data or write enough rules in software for them to uh be robust enough to be in our physical environment. But once you can introduce these this new concept of physical AI which um there are these new class of models called world foundation models.

23:23 Again fee Lee and and the team at Stanford doing amazing work there. Nvidia doing amazing work with our Cosmos models. Um you can essentially create millions, billions, trillions if you like scenarios in which case you can start to train um these otherwise physical things. Could be a surgical robot, could be a cobot, it could be um it could be just any kind of medical device that um will be working with patients. And so um that third computer uh has been the missing link and it is here now um with world foundation models and the ability to do um very physically accurate digital twin environments. Um so synthetically generating information and then being able to physically accurately represent it in a computer such that it obeys the laws of physics. Um, so we can train robots um in in in very real time and um everything from every medical device, whether that's an ultrasound machine is going to be uh autonomous in the future, you're going to be entering autonomous X-ray rooms and and and uh diagnostic suites uh diagnostic imaging suites all the way through to um in fact the very first um excuse me FDA approved uh autonomous um robotic task uh just came through at our GTC conference with a a company called Moon Surgical. Um this is a surgical assistant robot and they pioneered um by watching by watching the tools the scope actually follows and moves. Otherwise, a surgical technician, a surgeon is is talking to a technician to do that uh motion and it's on uh you know, obviously robotic arm. So, it has like just extreme precision and that's that's the first time through AI vision systems that they've actually allowed the robot to actuate in a surgical environment. Uh so, we're we're moving now. we're we're cooking uh in this area and I think we're going to see a lot of um uh you know very very rapid advancements uh in the operating room in the patient room and I would say in the in the hospital as a whole.

25:36 >> Yeah, I I love so lot so so much there I'm excited about. I think that, you know, the idea of digital twins and and I honestly should do a tip of the cap to another space that we don't talk about as much on the show, which is manipulating the chemical environment, the molecular environment for drug discovery, which is a whole another topic alto together. But to your point, this simulation idea is fascinating to me because like again like we encounter all these scenarists from the the inexperienced proceduralists, right? It's having that muscle memory, seeing the edge cases, being aware of that, and having that assistant the whole time. I often make uh I've been pushing my team here occasionally. I haven't gotten any takers, but to literally do like, you remember Tesla had the the dojo, right, where they like would simulate, you know, millions of different outcomes. I always thought we could have a hospital dojo >> and we just like let everything play out and just see what happens in all these different and I my my running joke my dad joke version of this is that in 99 out of the hundred scenarios the the the clinicians rise up against the administrators and take over the hospital. I I say that with tongue and cheek, but I but I do feel like there's an opportunity to take real world data, possibly physical environments and like to your point simulate this with the world model um and and probably come up with the optimal care that is like the quintessential precision medicine, right?

27:00 >> Yeah, absolutely. And and this is this is well within reach, Matt. Right. We um at at our GPU technology conference this year, we announced a platform called Isaac for healthcare. And this is all of the tools necessary. It's digital assets um that are already pre-made for you, hospital beds, you know, operating tables, lamps um that you can start to build out these digital twins. um were working feverishly and we were just at the Society of uh robotic surgery conference announcing a whole bunch of new technology and workflows for teles surgery um for being able to build an autonomous ultrasound um where you can use just a you know a commercial off-the-shelf arm and attach you know a mobile transducer to it. You can speak to this robotic arm and say please go scan the liver and it automatically scans the liver doing the vision system in the background that's seeing you know liver lesions and uh and all of this are now developer tools uh available uh to to the community all open source and so um we can build that environment and then you can take um these Cosmos models and and also we have group models uh to be able to create those corner cases.

28:17 You know, even taking we're working with clinicians to say they did experience those corner cases sometimes traumatically um for themselves and their patients. So, you can build that into these foundation models um and make sure that it it captures that pre-existing, you know, experience as well as go hallucinate and dream up a a crazy thing that hasn't, thank God, happened, but could could be, you know, physically feasible. Um and and so I think it's going to be uh just really exciting to see how advanced how fast it's going to advance. Um just like we're seeing in the digital world, we've not seen technology sort of enter the healthcare space at this pace before. I think in the physical aspect of what we're talking about, we're going to we're going to feel a similar experience.

29:05 >> Wow. So we're we're moving totally agree with you Kim also on how fast software is running in we haven't seen adoption like this you know to come more more on the surgical side other places but help help ground us how we how we get there from your perspective obviously the three of us are some of the most you know optimistic people around where I where AI is going to go what's coming next you know thinking you know about what's possible you know when when ground this in let's say the perspective of a hospital CEO today as they're looking at these tools what what do they see they see you know the financial impacts coming from you know the one big beautiful bill coming down and looking at serious cuts to staff other resources ways to make ends meet I would say the broader context historically is most of those CEOs who got there today the right answer or call it the past two or three decades has been to ignore technology.

30:09 Right? Every new shiny thing that's come along more or less the right answer has been to ignore it. Just wait. Let someone else try it out first. You know, don't get distracted by, you know, blockchain first generation, you know, of AI systems that came out a decade ago. You It's been the right answer to ignore it. How do you as you have these conversations start to bridge for those people? This is where we're going. I know where you're coming from where you haven't done what why should they do something now to start interacting with these tools. What is what is the way either of you kind of hook them in to kind of engage now with these topics?

30:46 >> Yeah, I mean I think we've we've also have the we have the great benefit of um chat GPT being an interface the whole world could understand. Okay, that's a great benefit. Um, now you're you're speaking to it, right? First you were typing. That's still a pain in the butt for doctors and nurses. Now you can speak to it. I mean, just last night I had I had to move I moved houses and I had to unscrew something and I I couldn't figure out what bit to use. I turned on the camera and I'm like, which and which bit do I use? And he's like, the third row down two and to the right and it was perfect, right? And so like, you know, it it empowers me to do uh crazy things like that. So, I think we have the great benefit that this is just so accessible to every human.

31:30 And I honestly believe that um if you're a CEO of a health system, you're going to start feeling it's unethical for me to not help empowering my existing overworked, burnt out, potentially leaving the workforce um health care professionals with these tools. they they h they are they are demanding it and they're expecting it. So it's in a way it's almost unethical. And secondly, there's no way they're going to get out of the red if they don't find other ways to treat more patients. Uh they can't hire the people. They can't hire more health care professionals. So they've got to find other ways where technology is going to fill this great gap. So so there's there unfortunate things.

32:17 There's fortunate things that are at the advantage and there's unfortunate things that are also kind of uh pushing in this way. And then as as we said, you know, the software as a service technology of the last two dec decades was tremendously challenging, right? It it was $3 billion in three years of installation and then three more years of potential change management to adopt these technologies of the past. Today when you want to engage a company like a bridge who has just a stunning um conversational clinical conversational AI um domain system that is excellent um you download an app on your phone and you ask your patient do you mind if I use this? There's absolutely no learning curve. So the the excuses to ignore technology have just completely diminished. um and the the existential crisis that the health care system is is finding itself in is is unethical and and must demand it, right? And so that's that's still like very much in this in this digital world. But the the moment we start to see all this ability where you can, you know, you're not going to be able to hire the person to check your patients in. It's it's going to be a digital kiosk. And and and why not? Why shouldn't it be? Um, you want to have your conversation with your doctor recorded both for the clinical outcome and for the the patient understanding and hopefully the preventative things it's going to help and and do. Um, you're going to want to have an agent um tell you what um next um appointment you need to make, right? Um so that you can stay on your course of care and hopefully reduce um you know the the the cost to the healthare system long term.

34:08 Um, and so then we're going to be like, of of course we're going to get more comfortable with cameras living in the hospital, right? Because it has all this contextual information and it can catch and see things and do things that uh wasn't possible before. And you know, I I am a tech optimist here. I am at NVIDIA for 20 years, but I am a patient and we all are. And when I get a colonoscopy, I want an a AI looking at my colonoscopy. When I get a mammogram, I want an AI looking at my mammogram. um you know if if I'm having a surgery that Intuitives Da Vinci is expert at, I'm going to want to go to a hospital that has that, right? And so there's just so much push and pull here that um it's creating the conditions, I think, where it just can't be ignored and it shouldn't be ignored. And the software developers of this era really have everything they at their fingertips to make um the burden or the adoption of it just supernatural. Um and then wonderful companies that are incumbents in the space like Epic, right? They've opened the doors as well to allow for third-party integration so that there isn't this you know wall that you hit. I might have a great technology but I can't actually get it to um you know interface with the systems that are the systems of record that is the operating system of of the hospital, right? Um and so you know that's a really important part is we have to recognize where these operating systems exist and make sure that you know we're creating conditions where a whole ecosystem can thrive. Um because that has been also some of the challenges in the past is the interoperability to use that fancy word the healthare the healthcare industry loves um of of all of these different systems. So that's how I think about it.

35:57 >> Yeah. you I mean that that's literally like that's the trend like that we're seeing right it's the health literacy it's the ease of implementation partially because of interopa partially because these systems are so right I mean they're accelerating the developers and then again platforms a lot of open source right I think you know a lot of folks are converging there and allowing and then finally I think the last two pieces to me are to your point expectations on behalf of the patient because they're using the tools they have access like they've never had before and then the final piece I think there's a I think there's they generate you know I talked to you know health systems and uh they're stunned at how many of their workforce are you know interested or using the tools today but then also in the education side the youngest doctors the newest generation uh they're using it all the time too and it's actually making them better at you know learning learning the concepts potentially even patient interactions having difficult conversations which is something we brought up before and this is probably the last chart Char will bring up. This is from a paper that just looked at medical students using the model using an AI model that was meant to simulate a patient encounter and then how they ultimately did on their clinical test like their evaluation. Are they ready to you know have conversations with patients etc. This is like one of those rights of passage in your medical school training. and and at the bottom you can see that the the the group that used AI outperformed those that did not at all or use the traditional you know sort of curriculum.

37:30 But the other cool thing about this paper which again I think Kim is it speaks to your idea of like this is a democratizing this is you know everyone is has access to these uh the the students actually 70 plus% of them weren't sure if they were working with an AI or an instructor in some cases with these chat interaction which is like literally the touring test in medicine but I think to the point of like I there's this familiarity there's these increasing competencies but as a learning tool and I know this is one of my soap boxes. I don't think we're thinking about this as often as we probably should at the ability to upskill folks in a variety of areas leveraging these models. It's not a question answering machine but as a tutor. Um I'm guilty of this too. I've started to do better at that and I think there's a lot more to come here in terms of medical education getting transformed and then then I think we've got a complete cycle where things will continue to accelerate in the way that uh that you've outlined and I'm I'm here for it.

38:29 >> It's it's it's awesome and I think the upscaling we have again here we have tremendous opportunity. I myself as a consumer right I was talking about how you can just talk to your phone. You know what I mean? My whole way to work, I want to learn about um navigation systems for for surgery, right? And I'm thinking about the technology behind it and what and I just had a chat with Gemini all the way to work. Um and I'm getting smarter uh on my way to work while I'm driving and I don't have to put myself in danger reading anything.

38:59 And and it isn't, you know, a pre-recorded podcast that could be two years old. It's the here and now. It's reading all the papers that recently came out. It's um giving me market intelligence. It's who are the companies working in this space. And so, you know, you can upskill on complicated areas like navigation systems and surgery in in a 20-minute ride to work. And it put me on a whole other learning path as I did that. So, and it's just such a it's to me it's a joyous experience because you're you're not afraid to ask stupid questions. Um it's totally okay. And um and then you just keep getting better and better. So I think you're right. I mean this this generation is going to be gen AI native, not just AI native. It's like generative AI native um where you know they can they can watch the chain of thought. Even even watching the chain of thought is stunning, right? We have vision language models that are are doing chain of thought for a radiologist. This is how radiologists are trained to think through you know what when they're when they're doing their study. And uh it's really interesting to watch that or or you're you're like you say a veteran in the field, you can actually embody your train of thought into your own into your own reasoning model. So it can work just like that for you. Um it's it's it's going to be um it's going to be a tremendous time. And and some people ask me like, well, what does success look like for you? You know, and I'm like, for me or for Nvidia? I'm like I'm like if it's if it's for me it's it's when I as a patient feel every day that the health system has changed for me right and the fact that you know I can go see a doctor and they are using a bridge it's changing and it's changing fast or if if if I can go you know find that colonoscopy center that has Medronics GI genius platform this is good um but hopefully you're not going to have to seek it out. Like I think you guys were talking about like AGI in medicine a couple episodes ago and it's like hopefully actually as a patient you're just your level of care and your ability to be a good patient uh is just going through the roof like we've never seen before. Uh so that's what that that's what success will look like for me.

41:22 >> No, that's that that's amazing. And I think what's scary but exciting as well is while these changes are happening so fast, we do have an opportunity to use literally some of the same tools to re-educate um both our patients and clinicians for kind of keeping up with it and where it's going. Um you know, I guess like you said, you know, talking with a model instead of listening to a podcast will will make Matt and I out of a job here uh on what we're doing, but that's okay.

41:53 I love them. Um, and I, you know, but it's it's, you know, it's very, it's very interactive. As I said, it's hard to interrupt a podcast with a with a silly question, but that's what you do in a learning environment. You might raise your hand and ask the professor something that um, you needed clarity on, right? And so, I'm oftentimes pausing a podcast because I heard something I didn't understand, hitting up perplexity, and coming back in, you know? So, it's it's it's it we need we need all of it.

42:21 uh but the ability to consume it um is what I think is going to change the upskilling factor because you can consume so much more so much more complex information at such a rapid pace. Um so upskilling is is is going to take massive effect >> uh into all of this. >> Well, I mean one of the reasons we do this podcast, Kim, is uh so we could ask dumb questions of very smart people uh like you. And so we really appreciate you coming on and sharing everything that's going on with the video. We're we're just thrilled. Uh again, I think we um I think we see a similar future and and we'll be keeping very close tabs on the work that you all are doing particularly um particular as we start to move into this robotic space which I'm personally fascinated about. Again, Justin and I don't get a lot of a lot of uh opportunities to talk about it and um I've learned something just just talking to you today about that. So I'm going to go read more. Maybe I'll talk with the model about it. Um but uh but yeah, robotics is uh to me uh the next frontier and I'm here for it.

43:23 >> Yeah, Matt Matt, Justin, thank you so much. I I I love when you say ask good questions. Um you guys know this space incredibly well. I'm just a student of the space constantly hoping that we can take some of the most advanced, you know, technologies and and bring them into this domain where I believe it's going to have, you know, the most profound impact. Uh, so I'm excited about that and and I would I would love for you guys to talk about another frontier as you were saying, Matt, which is AI for science. AI for biological science, a AI for the physical sciences and actually how just AI is going to transform science and and a lot of the the deep um, you know, personalized medicine space. It's very much science, right? Oncology and science are deeply similar in the way that they they they work. Uh, so it will be a really neat um, you know, autonomous labs with super intelligent agents uh, is a is a tremendous future that I'm also super excited about.

44:24 >> Amazing. Well, we we'll have to have you have you back to to go deeper. Kim, thank you so much. >> Thank you guys so much. Appreciate the opportunity. It was a pleasure.