AI In Healthcare Series: Leveraging GPT-5, Cosmos, and Predictive Models for Better Outcomes

0:15 All right. Well, welcome back to the Stanford AI and Healthcare podcast series. Uh again, this I'm Matt Lungren, uh and I'm joined by Justin Nordon. And today's guest uh is probably familiar with most of our listeners. Uh Seth Hayne, uh who's the senior vice president of R&D at Epic. He has been at Epic a whopping 20 years, um focused on technology, AI, and applications there. Uh, and you know, fun fact about Seth, I've gotten to know him a little bit over the last couple years, uh, as a math major, he's obviously brilliant and and a brilliant technologist, but as a as a card carrying English major myself, he is a prolific reader, and I find more often than not that his book suggestions are way better than mine. So clearly, God has given with both hands in terms of talent to our to our guests today.

1:00 We're excited to have you on the show. Welcome, Seth. >> Thank you, Matt. It's good to see you both. So we have some topics uh I think to tee up as usual I think we typically try to get to just like where are we uh in the at least at the point of recording in this exponential world we're living in and I know a lot of folks in the audience are a little different than us in sense that they don't obsessively follow a lot of this. So Justin, I don't if you have some of the latest. So u GBD5 has launched. Um I think a lot of folks maybe had a you know mixed response let's say um at least in relationship to the hype leading up to five. Um there's we can get into some nuance there. I'm curious to hear your all your your take on this. But from the healthcare AI lens you know once again it's kind of saturating these typical benchmarks. So, I don't know Justin and Seth, you know, you've seen this kind of data over and over and over again where we're we're using some of these benchmarks. We can get into that, but first impressions of five. Um, thoughts around how are you using the the models today? Has anything changed for you with uh with the latest release?

2:03 >> First impressions on five. It makes me laugh because every time one of these models comes out, um, the very first thing I do is I give it to my son who is 14 years old and I ask him if it can build the magic deck that he has always dreamt of for him. Um, and unfortunately we still haven't cracked that nut. So I think uh that there's still a ways to go. I think the the USML benchmarks it's doing better on than my son's magic benchmark um which is current.

2:32 >> Those are the two that you track. Those are the two I track. That's exactly >> Has he But the big question is is had he has he made a magic card with your skills and stats like like does he have his own personalized deck with with the parents? >> He is he is he is starting high school. He's avoiding parents at this point in time. >> Smart. >> Probably smart move. Yes. >> Yeah. We we'll steer clear of that at the moment. Um I have found that I spend a lot of time you know you brought up GPT5 specifically um using the thinking in pro models right I think that there's sort of two tiers of usage of it folks are pretty excited to see what can happen with these reasoning models and I think there's some many folks are getting exposure to it for the first time um but I think that the opportunity to know consistently you're going to get the high quality output with either the thinking or the pro obviously doing more with a little more latency tends to be my my usage patterns these days. I don't know about you guys.

3:28 >> Yeah. And I think you know and we we've seen also some in the comments people ask you know Matt and AGI camp this is this is this is the next thing. Uh I I'll say I I was kind of in more of the rest of the internet of hey it is it is better. I think it is a very interesting consumer choice to do some of the routing behind the scenes to kind of expose people to different models. uh I don't know if it was the Death Star giant giant leap forward that some people were predicting. So there's a separate question on again what is the rate of change how much better are these models getting so I think from the was it the exponential you know from three to four four like no like to to be super clear I think at least not from from my own experience there. Uh but then other you know interesting there's again it's not just that that's come out since our last there's you know Seth to your point on you know magic cards and is your son gonna gonna make one for you there's you know Gemini you know nano banana and image models and other more specialized models coming out that aren't just in this hey bigger bigger models and so I think to me this is just kind of this broader theme of model selection specialized models and other things and you know I know we're going to get to this topic you know later with some of the announcements around some of the cosmos pieces. But I think that's just another thing for for I am watching as a field as I see these models come out.

4:53 >> Yeah, it's interesting. I feel like you know definitely we track maybe almost a pathologically too much the models. But I I would say that like I was one of the rare people that liked having choice a bunch of different models to choose from. And um and I maybe was also one of the rare people that liked 4.5. And for those in the audience who didn't know, that was kind of this middle step model, but really great at writing. Um, and and just was felt much more natural, at least to me. But but yeah, I I I think that it depends on your starting point.

5:22 Like if you were someone that was a casual user, a dabbler, and and you weren't even using some of the thinking models, and now all of a sudden you're asking a question and you're getting this really, you know, quote unquote intelligent response that was a result of a thinking model as opposed to the typical, you know, 40 response. I think you will be wowed and maybe it would feel like a death star whatever the hype is but but to your point like I think the incremental increase we're getting to a place where I just don't know if I have hard enough questions to really tease out the true difference and so I you know again we look at these benchmarks in healthcare obviously I have felt pretty solid about you know putting my own healthcare data we've had this discussion before and not feeling like compared to the original four and certainly 3.5 where it's wildly off. It It feels in fact I'm learning things. I feel like that I I continuously validate myself, but I'm still like, "Wow, it really is impressive." I I do feel like though on the benchmark side, we are we are kind of drifting away from being able to rely on those to to give us information that's useful in our daily lives.

6:29 I I think one of the things on the benchmark side, it's interesting that we continue to use benchmarks that are founded in kind of the education system, if you will, from a medical perspective. and we haven't quite figured out how do we move into those later stages of a clinician's journey as they become better and better in the role and move out of school. And I I think there it's a it's an important benchmark. It's something we need to keep watching, but I think there's other modalities and there's other approaches here we we need to start tapping into and thinking through.

7:13 >> Yeah. And actually one of the other recent papers and I I'll pull up the chart here was uh there are you know serious problems with multiple choice questions as as kind of the only benchmark and this was a study done from a a colleague at Stanford uh you know Nikum Shaw's lab that showed even if you just change um you know none of the other answers from from the right answer performance dropped >> right and so like there are issues and you know we were just talking about getting a consistent response also to these things. Remember, you know, these LM are non-deterministic and so there there are issues to work through really as we think about performance and h how we're evaluating these these systems. I think one of the things that is often challenging between this type of research which I think is important and often informs for example types of development paths we take here in regards to prompting and other pieces is that it doesn't represent the way that we built software consistently right we know when we're creating say a summary uh for a physician in the context of the ED that has a patient presenting in a certain circumstance that there are a series of checks and balances you need to go through in regards to creating that summarization in the context and then also using things like citations back to other parts of the medical record to help account for the quality in those contexts and it's unclear to me how one starts to replicate that in the academic literature in a consistent manner. So I think that's one of the reasons I was excited to come on the podcast because I think this type of dialogue needs to continue to happen between both industry and academia in many ways because we can keep learning more from each other.

9:03 >> Well and I think you're raising a really important point. I think you know we can park the idea of like do the models broadly and and again at a at a place like Epic which is famously engineer you know the prototype of of practically your entire workforce is an is an engineer. So there are clearly advantages and you know I'd love to hear your perspective just on using some of those tools and you know you hear some hyperbole about percentages of code written by you know AI assisted models whatever but if you just go back into the physician use case it's an incredibly good I feel like we've gotten enough signal from the models with the benchmarks and some of the anecdotal use that they have capabilities that are useful for real things in healthcare.

9:47 The question is that you're raising is like okay now that we know their signal how do we actually craft that into something that really delivers something for our end users whether it's a patient or physician and that gap to me is still a bit of a chasm and does require the hard work of either having the domain expertise the uh the tight developer user interactions a and frankly just the knowhow of how to stitch these things together and that I think there's an overhang we talk about the overhang a lot capabilities are here getting those into the real world to show that value is still going to be messy work. Um that will take time.

10:25 >> Well, and I think you know first off you got to make sure you're asking a meaningful question when you're designing a solution, right? What is and this I don't know it's it's probably cliche at this point but I mean we deeply believe that every developer needs to get on site through some we call it immersion here but you you've got to spend time at the elbow with the physician understanding what are the problems you're trying to solve and in some cases these are um kind of diagnostic type questions about labs in those cases in many of the contexts it's an administrative question and I think you you've got to start by asking the question. And then to your point, there's this question of how do you build up the right pipeline? How do you build up the right backend structures both in regards to monitoring and ongoing kind of reinforcement loops for improving the backend process for whatever you're generating be it a summary be it a a draft of a note um be it a a SQL query for putting real world evidence at the point of care and I think that is starting to mature. I think the thing that there's a lot of opportunity to spend time talking about and continuing to think through is what are the new user experience patterns, those design workflows, particularly where having the models think longer continues to improve the outcome. But now we're not in this instantaneous I have an answer in under a second question, but it might be worth waiting for. And I think that there's a a important conversation to be had on that front that's coming in the industry and we're just starting to tip into we could talk about nothing else but but that right and and I think uh I think it's what's fascinating right now is right you know we were just talking about model selection you know at open AI there was a lot of backlash on taking that away and forcing people to an instant response versus thinking um we we we haven't solved it there. We also haven't solved it in in healthcare as we think about higher risk interactions and latency for certain pieces of data being able to wait for others. Um, and then I think the points you're bringing up, Seth, on just software and building software, the way I describe it often when I'm talking to people in healthcare is if you want something that is reliable, consistent, you know, high quality, candidly, you want as little LLM as possible because that's going to introduce noise in what you're working.

13:02 And so we have this kind of dichotomy of worlds. We have a new model come out and benchmarks and oh my gosh, look how good it is. No one is building software that's just LLM output. Good luck. You know, hopefully hopefully that was the right piece. How are you bringing the right context? How are you bringing the right data? Audit controls, monitoring, governance, valuation, all of these other pieces are how you build good software. And I think as a whole world, we've gotten so excited about the LMS.

13:28 In some cases, people have forgotten about kind of good software development to kind of actually build high quality systems. Well, and I I think that importantly highlights kind of a common misconception I sometimes hear where folks confuse the model with the actual enduser application. Right? There's so many other ways to use these models besides a consumerf facing chatbot whether you have a model selection or not. I mean, we don't have people typing prompts into our software when they're getting a summary, right? You need to build in consistent guard rails, consistent checks in regards to where the data is being pulled from, how you're using these different patterns, and then ultimately designing it into workflow so that an informed decision is being made in regards to what's being presented and that user has the opportunity to make the decision. And it's very very different than the consumer experience most of us have on a day-to-day basis.

14:32 And well, I think this is this is important too because I think again the the the one the one onetoone interaction I think uh will will have its limits but I but then I you know to your point Justin too like as I think about the tools and the software and the things that we rely on that are deterministic that that are like a calculator let's just say like I I I still feel like there's a part of me and I've said this before but like I'd love to still abstract some of that away uh and and And and to your point, Seth, like I don't maybe need to use the model for the actual interaction for the answer I'm getting. I I just need to go fetch the right tool to to then retrieve the answer or the piece of data. and and that can be kind of again pushed back from my having to choose or or set that up in in advance like and I guess maybe we're not quite there but but to me there's a multi- aent story there's a multi-tool story there's a better together story where all the work that we've done for decades in traditional LLP and and all the calculation risk scoring start to accrete additional value to having an agent that can get that information for me and and save me time or mental effort whatever we use as the as the metric I I feel like anybody that is continuing to push the bounds of using AI in new ways needs to simultaneously be as ready to push for not using AI when it's not needed. And sometimes sometimes that seems contradictory, but I think it's back to your point, the right tool for the right job. And it this does provide a generalization framework that can be applied in a bunch of contexts, but that doesn't mean you should always use it.

16:11 >> Yeah. Well, this this leads to now what's because I feel like there's a there's clearly going to be a continued slope of improvement probably on a lot of tasks. Some of them may be healthcare of the models. They will be useful in applications, but I'm I'm really curious to to bring up some of the work that you that you all announced around specific models. So can we actually re re-evaluate the task and where the gaps might be in terms of does natural language really get us to where we need to be in specificity for things like prediction for longitudinal records for the language of healthcare which we can all agree is different than the natural language of the you know Reddit threads on internet that some of these models are trained on and and and is there a way that we can take some of the lessons learned from this phenomenal technology and apply them with some domain expertise and I I think that's what you've done and and I don't know if Justin you have some of the the the graphic from the paper that that um that you all put out but but to me this is like such a beautiful example of the NLP to GPT2 story in healthcare. In other words, let's reconsider the problem of the language the tokenization of the actually the language of healthcare.

17:22 Let's think about the fact that things happen on different time scales and and consider that right in the events in someone's progression through a health system. And then let's think about the prediction task that we've spent decades in medical AI, right? Trying to do prediction models and can we actually learn the language, hopefully get economies of scale, hopefully get the scaling laws to to tell us this is the right direction and eventually have a generalist model that can do speak the language of healthcare truly and then also predict events in ways that can help us truly advance the field. Anyway, I'm so excited about this work. Um, I don't know if you have other insights that you got from from from building this model, but this is such a powerful statement um that that just came out of of the UGM meeting. Really thinking critically, marrying the domain expertise with the scaling laws and the things we've learned from LLM.

18:13 >> I think there's a there's sort of a fun historic digression here as we get into this that that's maybe worth highlighting. Do either of you know where the name Epic came from? Have I shared this? >> So, It actually, and Matt, you're going to appreciate this. There's a, you know, with your English background here, it's actually the um classic Greek poems um and these long stories. And in fact, when Judy initially developed the first capabilities for Epic, which was the database that managed the data around the patient, she called it Chronicles.

18:52 We still use that name today. And so this idea of a patient story sort of been inherent in the company from day one. Um and the the approach we took here was really could we using things like medical events, interventions, observations and I think a a critical component here and you see it in gray time that intervals that have passed. use that in a chronological order to build out the story of a patient and then take that same transformer architecture that we've been seeing in natural language be so useful at predicting next word from a training perspective and apply that here. Um and as you noted it's early um the the at the same time the data set we've trained the largest model on so far is about 8 billion encounters. Um, so it's not small, but based on the scaling laws that we found in the paper, there's a lot of opportunity to continue to improve these models by growing them and then exposing them to more patient stories.

20:06 and the Cosmos community. This is an effort across health systems that use the Epic software to build out a large deidentified data set for these purposes is two to three times that size depending on how you measure it. So there's a lot of opportunity here. Um I think the sort of exciting thing kind of a couple of exciting things that I'll just highlight at least to me. Um, one, we have evidence that this model seems to be able to not just predict near-term events that might be, for example, what gets ordered or resulted during an upcoming encounter, but also seems to be equally applicable out three, five, 10 years in regards to predictions, which can obviously help in regards to managing chronic diseases, but also things like some of the capacity problems that health systems are facing right now in regards to getting in patients beds and those sorts of things.

21:08 Um happy to dive into other pieces. I think the the kind of visibility of the simulation of what gets simulated those future trajectories um will be interesting to explore from a UX perspective going forward as well. Um, but there's we're just I think scratching the surface as a community on how these new modalities can be used. I know I threw a lot out there. Sorry. I'm excited about Thank you, Matt, for the kind words, but it's been great to work with the the team at um Yale, Andrew, and Daniela who's on the Cosmos governing council, as well as some of your colleagues at Microsoft um to get this initial research out. So, It is amazing and it fits in it fits in this world of right healthcare is different. It's a different language that we're talking about and it's amazing just to see the push push push on the research side. I'm curious Seth as you brought up you know the other pieces on UX, how physicians, patients, others will start to interact with AI systems. There was another paper and I'll actually bring up one more uh chart that's actually gotten a lot of attention over the past uh few weeks which is around you know deskkilling desklling of users with w with with AI and and Matt I don't know if you have any more uh comments you want to talk here as as the practicing physician but between between the three of us um but but you know at large I think like I felt like I saw this and like yeah it makes sense you have a physician rely on AI a bunch for for imaging you take that away. By the way, maybe they weren't as good as it before. You talk about the calculator example. You know, you want to push all of us on mental math with, you know, three-digit, you know, three-digit numbers. Um, if if you don't use it, you you are going to lose it.

22:57 And so, like if Matt, you have any comments and then Seth starting to think through like how are you thinking about this? How are you thinking about UI UX as you're thinking about development now? Now, now at Epic, >> I think this brings up a couple concepts. I think you know uh I mean this has been true in healthcare. So I guess the point of the paper which I think is valid which is that you know physicians are using this AI for colonoscopy for pol detection it's quite good they it shows that they do potentially better with the software and then you take the software away and you have them do the same tests and they do less good right and and it and it's it seems like it's close but you can you can see that this makes perfect s everyone has uh you know stopped using maps to get around. I'm old enough to have had to do that before GPS and I don't I don't I just use GPS and and then there's medical examples, right?

23:46 like I haven't tapped out a patient's plural eusion since a med school uh you know physical exam rotation because why we have x-rays and we have you know ultrasound and we have all kind so there are lots of examples um at the same time I mean uh it does raise a question I think there's the two points that I I think that that we still haven't solved one is what are the acceptables de what are acceptable areas in healthcare that we're willing to be be okay with the descaling like you know what are there are there just things that we're just going to say we seed to the the the technologies where it needs to be maturity wise we can use it reliably this is the way we do things now but the other part of that which I think is still unsolving that's related is the human computer interaction or probably some new term probably human AI interaction like what is that ultimate combo and and what is the how do you have to display it to me so that we're better together and we've had this narrative Justin through the several episodes which is I need to see that human plus AI is better than either alone and and if not a system where which I think there's been some papers on now how do we divvy up the work between the human AI so that again the in aggregate they're better and that those are just unsolved but very important questions and the papers like this are starting to tease at that a bit more >> I think you're highlighting this sort of if I look forward 20 years, I guess I'm curious. I think I imagine we're all in agreement on this, but if if not, we should have that discussion, but I feel like if we look forward 20 years, the quality of care we'll be able to provide, I think, is going to be unequivocally better. I think we're going to have more real world evidence at the point of care. It'll be more personalized. there will be a new set of tools that folks will be able to use and have at their disposal in addition to what they have today to do that. Um, if either of you disagree, I mean, >> well, no. Well, the only thing I disagree with is that we we never on this show we only go uh we only go as far as 5 years out because, you know, No, I'm just kidding.

26:00 >> No, I I I I agree. It's hard to I agree generally speaking and we could debate the time frame. I I don't have a a clear a clear definition there but I think that the question in my mind is how do we um move through this interim state there is things we need to learn and frankly I feel this in the same way uh in regards to the developers I work with at the company there's a similar question in regards to vibe coding tools and experiences and I see developers now walk in the door where it would have been a PowerPoint point presentation in regards to UX experience. They don't have a functioning prototype in the same amount of time and frankly it looks better than their poor PowerPoint skills did before. Um and so and we're be better able to understand it now. Not everybody's up to that state yet. And I think there's this sort of interim piece we need to move through and I think it applies equally to medicine as it does to development as an example.

27:07 Yeah, the the the vibe coding is such an interesting example. All of our all of our MBAs and I still teach a bunch. They are now expected instead of slides, they are now expected to show up with working prototypes, working demos. It's just a different way >> and it's so much more powerful, right, to to communicate, right, that when you can say not just, hey, here's the concept and these are the three bullet points I want. like no you can touch and feel this is what we're working at granted very very different than functioning software in the real world for all the kind of reasons we we we've talked about right but the communication is is is totally different and then Seth to your point on and yeah 20 years is far too long and I think that's where that that is where it does get really interesting though is really at what time scale are these changes going to start to happen and the fascinating thing we've been discussing is there's this balance between Patients are kind of going now. Again, for better or worse, patients are kind of going now to these consumer tools.

28:07 And for all of us who work closely with health systems who think about this on the other side, how do we as a field about balancing these imperfect tools to have people keep up or educate or change how patients are interacting with it? Um, and and that's just this this separate debate because I think you're Seth, I think your your 20-year answer is easy. IC answers undoubtedly yes. It's just is it fix 12 months 24 months but 5 years to Matt's point where we're how do we navigate as a field through it?

28:40 >> Well and this is so last week was our user group meeting and here here in Verona in Wisconsin. One of the things we spent time talking about was a collection of work we're doing both in my chart, the patient facing application. Most folks just have it on their iPhone. They're doing things like getting lab results or scheduling an appointment, those sorts of things, checking on their kids growth charts, that sort of thing. And it and then the physicianf facing experience in the office as an example. And part of what we were talking about there was that there's an opportunity in an integrated set of software to be able to kind of answer those consumer questions that are starting to be asked but grounded both in one their medical record, their family history, that information that is in the charts. So you can have more confidence in that piece. And then secondly, and I think this is I think this may be even more important bluntly, the opportunity to hand off from that experience to the care team, to engage a nurse, to engage a provider, and to escalate that conversation to somebody else that can intervene if necessary in that context and be able to do so seamlessly and go back and forth. And we tried to highlight the importance of that. I I think of it as two sides of the same coin, right? One side is facing the patient and the other side is facing the care team. And there needs to be a consistent dialogue between those two with grounding in this sort of rich history of the medical record. I I love this so much and and you know just again you there's always that classic thing in in in healthcare as our time to see our patients and discuss their care shrinks and the pressure on all the pressures around that but there's always that famous I mean every everybody's gone through med school has this like you know Marcus wellby kind of old old school doc that'll tell you the only question the patient actually came in to ask you will be asked once your hand touches the door to leave the next appointment And that's often the most gnarly complicated question, right? It's just it's go it's without fail. And and to me when you're talking about someone who's both getting uh more educated about their own data and they're asking the right questions, whatever, but then they're teeing that up for I mean that's such a beautiful connection to me because now I have a way better understanding of what are the things I really need to deal with in this visit that I, you know, without that information, I'm just going to assume refill a couple meds and and make sure that you're up on your screening and then all of a sudden it actually ends up being this whole other thing, right? And but but that communication is so key and I think that we talk about this a lot too, but this idea of information asymmetry between the physician and the patient that goes both directions by the way. Yes, physicians are knowledgeable about their field and the things that they do and the patients may not be as up to speed. But on the other side, I am I am at a disadvantage about, you know, all the data that I have about you as a patient and all the things that are troubling you when you're not in my office like that. Can we level the playing field on both sides? I think that's a phenomenal future vision >> for for for any listeners that happened to be at Tuesday morning of our user group meeting um and just heard what Matt said. He did not share that idea before what you saw on Tuesday happened.

32:22 We actually what we shared um was a my chart experience where the conversation with the patient through my chart previsit build out a visit agenda including double-checking for that last hand on the doororknob moment and then the the screen flipped and they were in the exam room and the vis part of what we're building towards is a real time experience driven by the conversation in the exam room that then walks through that visit agenda and brings insights including real world evidence from Cosmos into the exam room based on that conversation in the exam room and that kind of previsit experience that already teed up the visit agenda exactly to that point Matt um it's one example but I think it it helps highlight how there's s sort of so many new modalities that can come come into play with these different models when put into workflow in a user experience that makes sense and connects the patient and the care team.

33:30 >> I love it and and I I will confirm that I I unfortunately did not get to see uh the presentations there but but this but to me like this these are the kinds of things though that that I think if you grabbed you know Dr. Smith off the street in Anywhere USA, they'd be like, "Oh, that's that that seems like it's so far away." Honestly, and and this kind of goes back to just the general theme, which is that you almost have to just really say in a perfect world, a magic wand, what would you want the perfect visit to look like from both sides and and see how far really you are away from that technologically? And could you cuz I don't think again with that overhang that we're that far away from being able to do these things at scale today or in the shorter term without having to say sure guys but it's 5 to 10 years away. I think it's much more near-term than that. Right.

34:22 >> Well, it is. It certainly is. And I think an important aspect of this is that the incremental steps to get there each add both value in regards to time save for physicians as well as a better experience for patients and improve the quality of care. So you can step into this type of experience and incrementally get there because there are some of these are more complicated and we need to continue to both drive down latency in regards to how these model responses on a technical side and other aspects in regards to doing this at scale and there's but that works underway. it's going to happen to your point and so there's a real chance to quickly step into this type of experience.

35:09 >> Well, I I I don't even know how we how we top this. I mean, we just talked about an entirely new patient physician interaction. It's coming. I guess the the only debate we have or the difference of opinion is is is when um but I I I guess Seth wants was there anything else? Obviously, yeah, we we we glossed over it, but you know, Epic UGM, one of the biggest kind of events people eyes on watching in healthcare. Was there anything we missed that we that we should have have spoken about here as, you know, our listeners are thinking about, you know, the future of healthcare, AI, and and what's coming next? I I think one thing that's important not to lose sight of and I we've we've hit multiple times obviously on the importance of this patient physician experience and um helping everybody be better informed and and efficient in that context. At the same time there's also a lot of challenges facing the health systems. Um and I think you know that is a societal question as well.

36:10 there are not enough physicians. Um we have an aging demographic that is going to continue to put increased pressure um on society in this regard. And I think the types of things we're talking about help there. Help with access challenges, help um with you know even if it's as simple as helping answer billing questions in my chart with an agent, right? Um, I think we we need to continue to keep a beat on, and this was sort of the third theme that we touched on at UGM, making sure that continues to scale meaningfully as well. Um, it it's it's maybe not as directly applicable when Matt walks into the exam room. Um, and understandably so, but I think it is something that we all need to continue to keep a beat on together.

37:02 >> I love it. And and I I think that the you know we we I'm sure most of our listeners who are health have a healthcare background know this but but for those who don't the the epic UGM conference or you know event is think of like uh Apple you know worldwide developer. I mean it's literally like I I mean people lining up uh right I mean this is a this is a big moment. So kudos to you all for for pulling this community together. I think there's a lot of productive conversations that come out of just beyond what Epic specifically is doing.

37:32 And I think that the the the level of discourse has uh been you know accelerated forward I think a lot having to do with the way that you are positioning the technology uh in a very much a healthcare landscape thinking about each of the stakeholders and where and where we should be headed next. So uh thank you for all the great work uh Seth and really thanks such a pleasure to have you on. >> Oh I I've enjoyed it. Thanks for inviting me on and we'll have to do it again sometime. I wanted I want you to come with your turtleneck next time with the with holding up the the the next big device or something next time you're on.

38:07 That's that's all I ask. Something to do something to aim for.