Grand Challenges in Healthcare AI
37:40 Share

The beauty of technology is that I can unlock not just incremental improvement, but completely change the world we live in today. Just think one hundred years ago, we didn't have the internet, commercial flight or anything that resolve self driving cars. But equally in health care, a hundred years ago, we were still decades away from the double hillites discovery and biotics ivf behind so much more.

And perhaps there's no Better industry where we can make these truly monumental shifts in healthcare. And in today's episode, we explore these odda ious grand chAllenges over the next hundred years, or hopefully even less in healthy care AI. The immediate .

part is something that is so good, like not ten percent Better than what you have now, but like ten x Better than what you have now, that the adoption becomes natural .

from always on clinical trials.

A million clinical trials are just work, energy running in my population every day, and I have no idea how .

to harness IT spot market pricing .

on any given day. At any given hour, you might have a very different dynamic of supply.

more effective scheduling.

doctors designed their schedule, and in very protective, like almost a defensive way, because they felt wronged by the system.

Continuous monitoring.

How do you augment that with more continuous information, whether be self reported, whether be a patient monitor, da.

and even the holy .

grail of all holy girls, which is A I doctor.

and how far off that might be join us today are asic cases by on health general partners B. J. Panda and julio. This epsom also originally came from our sister podcast, raising health. So if you do like this, don't forget to catch more, just like IT, by searching, raising health wherever you get your podcast, or by clicking the link in our shots. All right, let's get start.

So, V, J, U, and I talk about the fact all the time that health care is an industry that used to be intractable with respect to the adoption of technology. But we are also super optimistic that health care is now potentially one of the biggest beneficiaries of technology in the form of A I. And so we want to have a conversation here today about busy, what you and I do on a daily basis, which is rifer about some of the grand chAllenges that we see fr builders in the heal care A I space. And so let's actually starch us with your sort of high liable thoughts on where do you think A I will make the biggest difference in health care immediately .

the immediately harper, the question, right? Because like if we talk about the one you are, I think a lot can happen. So the immediately part is a an issuable technology and also people, right? What will people accept? All people adopt.

And in many ways, I think when you think about the history of technology and health care, who will buy IT, who will incorporated? How will work into doctor's workless? So I think immediate part is something that is so good, like not ten percent Better than what you have now, but like ten x Better than what you have now that the adoption becomes natural or so easy to adopt, that even ten percent Better could work.

And so when you think about what could be ten x Better, IT has to be something where maybe is making decisions or it's it's helping doctors as a copilot and something that like a superpower they didn't have. Or if it's ten percent Better, but easy to adopt, maybe IT doesn't even look like soft, maybe IT looks like staffing or maybe IT looks like you're texting, uh, something and that's easy to incorporate. And even if that does a little bit, that could still be important because health care works at such great skills.

right? yeah. And I think it's timely in the sense that obviously, one of the number one crisis that are health corners is facing right now is a labor crisis and that we have of labor to do these kind of highly specialized jobs um that we have, whether be clinical or would there be administrative.

But also those individuals who are in those jobs today are extremely burned out because of, ironically, the technology burden that we put on them, whether be in the form of revenue cycle task or H, R. Works and things like that. So that's really something that we hear all the time. I think the other thing that you touched upon, you know, it's sort of the D M, humans that have gotten in the way in the past, not the technology process.

One of the hardest things in healthcare is behavior change and whether that be on the part of the patient um to adopt you know some sort of new behavior that helps them get Better or uh in the case of a condition, something that sort of changes the way that they do their job. And I think that to me, one of the biggest opportunities is how do you take things that constitute behavior change that have been proven in very niche populations and product died them, package them in a way that can of a sudden be sort of globally applied to the broader population that should benefit from. And so I mean, we have a number of examples in our port for that I think we will touch on along those lines, but those are a very good and big points to think about right now.

So given what you just described, and let's assume that we all finally, I believe in this optimistic view of what A I can do, what are the use cases for which A I can actually have utility in the near term. And we put fourth is of a two by two like good consult yeah exactly. And we said, okay, on the one access, you have B2B use cas es.

So historically, a lot of technology first gets adopted by the people on the inside. But then on the other side of the access, you have consumers or patients um and then the other access is uh things that are administrative in nature. So maybe more back office versus clinical in nature where you're actually delivering a clinical service to an patient.

And you've talked about this mistakes are very different in each of those quail rance. The area that has had the lowest hing fruit um so far has been really the administrative B2B set of tha t equ ation. How do you think um about of the healthcare administration in internal facing set of views cases? And let's talk about like what we've seen out there that we think has worked well.

This seems like a no brainer, right? Because like I think what everyone one's worried about is like being a doctor's really hard. And if you're having an A I do clinical recommendations or something like that, we haven't even figured out where the human goes in the loop and all these things, and we will figure out out that work to do.

Thing about today, thing about the back office, is that we've got computers there already. Yeah, we've got algorithms already. We also have tons of people. And you can ask yourself, like why do you have people doing our C M.

Or other tasks? Are these task that actually could be automatic and uh, that actually could really make huge an act on sort of the cost, but also actually IT also may be changes. How we think about this as we think of the back office as a data problem instead of staff in from yeah.

that's actually really interesting. You said algorithm and what that makes me think gove is the claim. So if you think about like the way the payments flow and health are, ninety percent of of payments and health care are reimbursed revenue where the provider has to submit literally a claim to the payer. That effectively and algorithm in many ways you could think of like the claim is a unit of a piece of logic that needs to be interpreted.

I don't know anyone ever, Tommy 的那 再去 看。

But then right now, the way that this process is like a very serialized workflow where first you have to interpret, okay, what kind of claim is this is a professional claim, is that impair patient claim? And and then which payer product is IT? There's like thousands of payer products in any given market.

So which one do we bump that up against? And then within that plan product, you have a hope venture rules about under what circumstances should this kind of claim be a valid and tea. So anyway, you have this whole value chain of decision making. Often times you have bring a nurse to that because there might be some clinical judged that's really prior authorization comes into play. So you can almost imagine a world in which like what if we were able to eliminate the claim and basically say, because we all the data, as you alluded to, as well as the automation of the set of decisions I need to be made around that, that adam, of data, that you could actually just eliminate the entire enter process and just have real time payments. So that to me is basically you could eliminate thirty percent in our system .

if you .

were able to do that. I N N two, there are companies that are out there that are doing this, but you do have to digitized. You know, like a lot of this is sitting in .

P D F documents and which are we are the very unstructured .

and unstructured and um yeah we have about like a company called her coys that is basically doing this for contracts and I mean the average payer provider contract which by the rate could represent like hundreds of millions of dollars of revenue of for any given pain identities, is typically like a two hundred page P D, F document that is completely monolithic.

But any one line in that contract might have huge implications for both the revenue to that provider as well as the cost structure for that payer. And yet those things don't get litigated, but for every two years when they come up for renegotiation, but no one's looking at that one line. They're looking at the whole kind of aggregate thing. So you know what, if you were to digitize structured data within that contract and be able to run scenarios on IT and say, like what if the Price for these ten services was this worse? Is that what would that implication be on the payments low through through that?

You may be even redo these things faster than .

every two years. yes. So the other really interesting thing about this concept of newly digitize streams of information and more launch total information about patients is what if we could have an always on clinical trial infrastructure in our country, such that on demand, you can slice and dice the population for exactly the characteristics of humans you want and produce a dat analysis, either retrospectives or prospectively. You know what you think of that idea and like, what are the barriers for us to achieve something like that?

Yeah, the fighting thing about that idea. I mean, it's a very exciting idea because we could get into much knowledge and we can improve health care so much. But like it's a ridiculous thing to imagine doing like without something like ai. And how you do that, how you pay, really, in me.

this becomes the data problem.

terms of sick and diccon. And then keeping track of, like this person got this drug and at this moment got had this response and serve understanding the causes, nature, right? So what we love of a clinical, get a little monkey is that like you have to have some sense of cause out like I took this drug, this happened.

And people say correlation doesn't an causation. That didn't mean we can't do caution. That's what trials about trials all about causing. So we have to understand the causes pathway. But with all this data, a is great, especially certain types of a like basic statistics are really good at casuality.

And so we could actually have caul understanding and even be complicated like, you know, you are supposed to take like great food juice with their birth control pill because the before fifties will not lifting this. I don't know how someone figure that out, but who knows what else is out there? Know that we just didn't find IT.

And if you could have an update like knows your diet and know those of basic things and knows your drugs and has an information that's kind of my log, that that how much we can just have no new drugs, no new, uh, treatments and just optimize. We are such an unopened ized nature because that optimizations almost impossible without A I. But then on top of that, won't you built out the infrastructure? Now, new drugs going to that infrastructure, and we can optimize.

And finally, we don't just optimize for health, but we can optimize. Join the optimize for health. Andy, crisis cost, like this drug is like ten x Better ten x more expensive than the other drug.

But maybe the outcome for me is going to no different. So um maybe I should get the other one exactly. And and and how do you think about that? That's such a complex data problem.

And logistics problem is also in the part of A I. And I think we could actually really find a tackle. So as you can have very .

excited yeah yeah no. I member, I conversation with like the cio of the B A many years ago where you know one of the ways you look at his population, he's like july, like a million clinical trials are just organically running in my population every day. And I have no idea how to honor that. You know, even with just E H R data alone, you can imagine the possibilities, let alone if you were also to layer on top of that, just your daily behavioral data, that kind of self, that's where the aleem s can come in, is just a conversational way to capture the data day journal of what you're doing, what you're eating, who you are acting with all that.

So dr. Behavior is like the whole system we can finally debug. And finding from coming from a tech background like, you know, if R, C, T is seems unfamiliar, this is basically a giant ab test, actually, you know.

And so this is deeply, deeply entangled in tech la. Even every pixel isn't even tested. I wish health are testing optimization .

is something that that would .

be just fantastic.

Yeah and that gets to the most the other sort of holy you hear people talk about this. Could you actually ever have almost a spot market for pricing and health are and on any given day, at any given hour of the day, you might have a very different dynamic of supply that's available for a given service. Why could you not press differently for the same way .

we do in other industry?

Why did that happen? I mean, today, it's probably deemed illegal onesta by many of the contracts because you're sort of bounded by these again, these moneth agreements that highly specify in the fact that you have this claim system. There is no really no notion of a real time a judicis of the actual Price that needs to be paid for that service.

And maybe doesn't this doesn't require the the fancy piva I press, but certainly um the notion of being able to to run machine learning on these things, say how many of these rules are just useless because they don't actually move the needle on cost or Price and but which are the ones that are most consequential that we are you should keep and therefore have some kind of a automated and systematic way to judicature. So huge opportunity and that again represents a very significant portion of work, its waste nar system. So that's a fun one to think about.

And then related dly, I need I mentioned at least my icarian health care like one of the sort of general themes of the problems that i'd like to go after or whether is like a fundamental miss match between supplying demand. And I think a lot of the companies import folio represent that problem space. And I had built to company that was in the scheduling space, which, you know, when you think about the phenomenon by which i'm sure you've experiences this, you know you is a patient or told to wait weeks for dr.

Appointment, and you assume that that's because every doctor is booked out solid. But IT actually turns out that a lot of the capacity in our system goes completely wasted and you could just simply get Better visibility into the underlying data streams than potential. You could really mitigate the way, times and the experience for the consumer while also helping the providers kind of best use their time.

Are there any examples that you have seen that you think are an interesting representation of that? I know we have a lot of companies that are trying to both increase transparency of supply companies are using coaching or provider groups and and then bringing that a bear in care models. Well.

here I think you're describing something which is even like maybe before even get to AI, which is like we got to get like off the Whiteboards and onto some more modern sort of computer approaches to system system record. And that actually we often think about like technology versus people. Where's the the weeks? But where is the problem? Maybe the most radical thing one could say is that maybe the .

way of doing medicine has to change .

ah so for instance, like just the workplace of a doctor, you think you if how a doctor goes through their day, how can we support the doctor to do what we all them to do and what they want to do, which is maximized patient a welfare yeah but and maybe view IT as something that takes IT off of them. So like I think about devoted their infrastructure and oo they address the system internally and to the they're not optimized that something that even they know and they can see, right?

Yeah and that's an example in the case of companies like devoted who actually did start with clean sheet, yes, right, and build their own scheduling system actually on the exact care model that they were going after. Yeah I remember from h some of the work that we did at my old company, you would see the way that doctors designed their schedule. And very much to your point, they would design the team lets of their schedules specifically in a very protective, like almost a defensive way.

because they felt wronged by the way, that the system sent patience or time.

exactly. We understand what this feels like when, let's say that we have ten pitches with new entrepreneurs, we've never met in a row in a given day and you're just on know for like straight and and you know how like physically and mentally exhausted that is. Um it's the same thing for for doctors when they say if you give me you know four new patients back to back in the morning, that is far more taxing to me than if you in the spurs repeat patients .

or other tasks and after .

is we even worse, my god, if we had to do that. But so so you can see how the unfortunate side effect of the way that the systems h had traditionally have been designed has now caused this sort of ribble effect and the of behavior. But if you were to just kind of start from a clean sheet, as as companies like devoted very able to do, could you actually design a much more logical system now that actually learns from historical data, right? And there could be almost like a reinforcement learning component to that where the doctors could provide, you can learn over time.

So we talk about one idea, which in the paper here is the only way what i'm .

hoping also happens as part of this wave of A I. That is really a forcing function for people to actually take advantage of the data that we have now digitize, right? So we actually only what ten, eleven years post meaningful use in health care, which was the act that instantly vize doctors to actually adopt electronic health records to begin with.

And it's kind of remarkable to think that, like even five years ago, less than seventy percent of doctors had any electronic health records. So we're actually relatively new into the era of even having digitize forms of information about patients over a lunch to period of time. And so in many ways, we haven't at all scratches surface of exporting those data sets.

And there really hasn't been that machines of two historically out argue and but certainly not the necessarily the technology capabilities on at of the middle le layer of the stack to be able to do anything meaningful and useful with that data. But I think that's where the advent of the tremendous technical ships that we've en on the AI side and what you can do with that information, how you can synthesize IT, how you can present IT to someone in a way that actually uses and friendly. That could be the tipping point that actually get people to unleash the data were also, by the way, in a period of time when provide organza hospitals are struggling financially.

And so many of them are looking at, okay, how can I SHE model ze my data assets, right? Like what a wall of our companies want. They want to eat data. And one of the ways by which they can do that is actually partnering with these provide org ization who have these systems of record that they haven't been able to explore IT and actually pay them and give them ref share or right or whatever IT is to get access, to prepare data, to train their models.

Interesting to ask, like of the various crisis, like the staffing in crisis versus issue hospitals are dealing with, like which crisis are gonna be catalysts and which will be impediments? Think we can have feel like the staff in crisis really is like tail .

ents for AI strAngely.

Ah yeah but maybe and cove IT I think is tell win for AI because we're so used to virtual yes. But like maybe not all these crises will be helpful. And I think I will be a title interest. And yeah, I think .

that's a great point. I think certainly there are many who would not one, but I think good for. Function because now people are at a breaking point where the the that is, quote, way of solving that problem, which again is how do we produce more doctors? How do we produce more nurses? We just we can do that physically. And so that is driving, I think, a lot of this adoption.

We were remarking in a team that at this last G P E Morgan conference, like a hundred percent of the incumbent payers and providers got up on stage and talked about not just what they want to do with A I, but how they're actually delaying ai in practice because they found no other way to be able to solve those of more fundamental problems. I think the other talent that some might call an impediment, but certainly builders in our universe call a tailwind, is the business model change in health care, right? Movement towards variables, care fundamental breaks, the kind of the schema of like how healthcare worked for decades and in incumbent are more likely to be on their heels.

With that, Daniel ic versus the upstart to themselves, can build their entire care model and an Operating model on the basis of those new payment domes. I actually don't envy organizations who have to have one foot in each world, right? Because having half of your shop in a fever service model and then the other half in a value based model is very, very chAllenging to do.

It's fun to think that in a fee for service world, A I is nice, but maybe actually doesn't go against what you want. And in a value based care world, actually A I is the catalyst, right? Because yeah, if you can do things Better.

actually fund the example that that reminds me of is how the A I are getting suit right now. So so there's a bunch of major national payers who are using A I algorithms to automate prior authorization. And so the doing is taking the rules that humans wrote, that humans were executing slowly and doing IT faster.

And so now of a student, everyone is complaining, oh, everything like the the denial rate is going up, but it's actually not the rate of denials is not going up. It's the speed with which the denials are happening that's going up. Don't blame the technology, blame the humans who actually wrote the rules.

And you're just seeing kind of an exacerbated version of IT going to deal with a lot of that of kind of the finger pointing at the technology where it's actually just implementing the broken system underneath IT. And that's why this kind of move to like new business models give you the opportunity to clean that up and start with logical ways to control spend. okay.

So we talked about health care administration. We talked about scheduling opportunities. Let's actually talk about the E H.

R. itself. So LLM as an E H R, what do you think about that?

Yeah well, so I think the think that's really underprepared. The LLM is like people think of IT as like this oracle or something. But I think it's maybe at least for us, I think of as A U I yes, right.

And I think because we start with like line interfaces, you know, for those who ever adopt with that. And I think we have goods because as Better than command line. But now back to text and typing things in except in set, like some weird command that you have to memorize. You just like, just tell me what you want.

right?

You speak english, you speak english and we're so optimize for speaking english to each other. I mean, that's like easy, doesn't required training in the same way. And so I think as A U I, now that makes sense.

And now when you're seeing as A E H R S and and then I get your kind of meaning that the data in there and IT can be clearly like this and may be synthesized. It's all very natural. I think obviously, you want to be very clear about partitioning things.

And so maybe you're doing worth like rag or something like that. Where is getting information coming back? H, but maybe the question turnaround is like, uh, again, the technology sounds very plausible.

You imagine a hacker that would puts some pieces together and get that done. But I think you need more than just like connecting to GPT four or or german or something like that. You need something medical specific.

Yeah yeah. absolutely. This is where I think one of the prime examples where we should they believe that a specialist you model is necessary to understand the specific nuances of how to interpret medical information versus general internet information. And I mean that certain in a big area of development for builders, as we see IT company, is that are building tools that can do everything from summarized existing medical record data.

How do you told the story of vj on day before he walks in the doors so that you understand like his journey, and not just look at a bunch of numerical records and concert of sporadic formation about different visits and what not, but really, truly the story of him, including things like your social determines and what happens in the home. And outside of the clinical setting, we see a lot of companies kind of building that. The other obvious use case is the scribing use case, where have a conversation with your doctor as he looked them in the eye and rather than them sitting at a keyboard during your entire visit and um that also gets written as a story. It's a story that then gets added to your medical record IT can create that fly will effect of continuing to add to the narrative of your journey.

One way to sort of think about A I and health care, take the existing jobs and then see which ones can go in. And that makes a lot of sense.

Also curious if you slice and ice to a different way because we don't have people with A I, how would we do different differently? Because like people are science specific jobs because the way humans work, right? But like maybe if when we're finally said and done, when I can do everything, maybe the resident isn't the world that I would take. Yes.

but if you were like unbundle the job of the doctor, what importance could you yeah bundle into a different thing? Yes, there was actually a time where this concept of like a data was sort of popularized where like a made like a baseline component of busy what every job in healthcare is doing as some degree of data.

And so if you were to unbundle that component and create almost like a horse zonal job, that was just doing this interpretation and to think, and maybe that's actually the Better analogy, what I described earlier is like, what if there was a data role that effectively is an L. M. That is synthesizing all this information? I think the thing that you know missing right now to make this a reality is today's information, uh, architecture is very sporadic.

So you pretty healthy person, and you probably see your doctor maybe once, twice a year. And so how do you augment that with more continuous information, whether be self reported, whether IT be remote patient monitoring data, whether be just other information sources, to create that more holistic picture? But I like that notion of flipping the jobs on their head and thinking about the components of different way.

All the fun thing about the data is I think the you eyes is pretty important, right? Because we're time about the team of people and the medicine. After done by a team, there might be a nurse, pa or or doctor or specialist, all these people. And where the comes in, one idea is a copilot, which is like each one of the t members has a co pilot.

But what's interesting about the data st, this is like the the A I is on a Peter contribute yeah for the team and has its role that if everyone feels pretty good about yeah and you think about you don't put a person to do the data is job, I mean, in principle with a calculator and a lot of time that maybe you could do what's necessary but you'd never having him being do that and that might be a very easy first entry. Where is like they're good at. They're doing what they're at.

Yeah, that actually reminds me of a company that we saw that what if every nurse in the impatient word, because the impatient setting is very chaotic, very active, things like surprises happen the time. And a lot of nursing teams have sort of either live like rocky talky type devices just on their shoulder, or they have some real time communication mechanism with the rest of their care team.

And this company was saying, why not put an LLM into the same I I talk, I signal ah and actually literally just have a big almost of mini crichtons being like i'm in x pattern by virtue of listening to your let's all remember that this is happening with this patient over here, and I think there could be a safety issue over there. It's almost like the way that I talk about being max all the time. So I could everyone just kind of a big max companion, you know, hang out in their care team and be sort of the steward of all the information flow, synthesized IT and read IT back when they find something that probably warrants and alert within that group.

I, if you spent much time in the E. D, I like.

break this and come. And I have last .

scars and stitches, and so I don't like so I remember I was a few years ago, actually right around here, cut myself with a chef knife was not play showing after the kids and and so so IT wasn't looking good like I get stitches and I go the E D. I'm therefore like two hours wo and and I look around and just waiting and I look around and i'm like, this could be another four hours and like i'm doing my math and and then maybe for the first hour i'm bugging the nurse and the invite. But like everyone, I just leave. And like there is a various situations where I just wanted talk to somebody, but you can have what I talk to somebody because I can be overloaded if I could just be texting somebody A I just want know where things are and if it's busy, that's fine or maybe I don't even need to beat there right now. But like that trios ing too could be huge.

absolutely. yeah. And this gets to OK. So going back to the concept of unbundling, the role of a condition, there is one part which is actually the treatment part. So that's a part, or maybe we can necessarily today building l am, that will stitch your finger. Yeah, but the notion of trios getting you to the right size of care.

So should I stay home? Should I go to an urgent care clinic? Should I go immediate to the E, D or mk, just going to my pcp? That is actually a critical role. You we were a piece about this where we said my version of that today is I call my doctor cousin as all my family members do.

The poor lady, she's cardiologist, but he gets every single call about every single special under the sun and you'll tell me little like you sure you take your son to the urgent care or does this need to go meet to the E. D. That is like one of the rules that L M construct could play, which actually would also do a huge service to doctors that they don't have to be the ones who are feeling those questions.

dr. Google ah the mature is deciding like you .

know and especially with .

my friends are problems through this like with like your kid is sick and and you're like I probably don't need to go in but like it's my kid so it's like even percent ago yes, and that's a just a drain on everybody. P doctors and patients.

Yeah yeah. So we talked about H H, R. And this notion of the the patient story we're now getting into, this notion of OK if were to take almost like the front door experience to help care.

And one of the big opportunities for A I to make an impact there, one assembly instead of going to google, you know, going to a specialized tool or whenever might be that trained in this way, what what are the one of the questions that always comes up is what are like the regulatory roles on this? So like at what point do you sort of cross the line into actually clinical decision making? And how should I think about this is a builder? Um I know you've obviously done a ton of thinking on this and a ton of work including like talking to the regulators and understand what they think you what what kind of advice would you give to to entrepreneurs who are trying to figure out where that line is and whether or not they should cross IT?

yes. So a couple of things is I think lions are more clear than others, but in the cases where there is any gray zone, I think the regulators are, I think, eager to chat with startups, especially maybe on the soft side that might be and see but you know like but like to try to figure out where you are. And you know we see a lot of successful founders.

We've done that type of collaboration. And I think generally, that's a pretty strong approach because then there is no surprises on either side. The tRicky part is when nobody knows, you know.

And so I think it's not just about the consultation, but it's also leading and so are taking the framework of the philosophy for how we regulate things right now really understanding is this software rs advice. Is that the right framework? And really think of being a leader in terms of how we should be thinking about this. And I think there is actually a welcoming of that as well um because it's new for everybody .

as you're leading to. We are actually industry that has a regulatory framework when IT comes to A I specifically. And so in many, these are like some of the rare cases where health care actually ahead of the curve as as technology goes.

What's your sense of you? Does generative vi specifically, do L, L, ams constitute enough of a sea change relative to historical waves of A I, that I should weren't an entirely different regulatory framework? Or do we think we should try to make know the current system work .

for for those new technologies, ala lation? And so in in that case, you have to now ask what's the specific use case where more regulation is helpful for patients? I don't see people .

talking about that. Yeah, the broader point of the necessarily focus on regulating the technology, but rather the thing for which .

the technology will be used.

Okay, let's go to the holy grail, abot holy grill, which is A I doctor. Yeah, I like. How far off in the horizon do you think that concept is where you could fully embody the the full stack role of a clinician making diagnostic decisions and treatment decisions and and what not and what needs to be true do you think and just like the broader ecosystem for for that to be the case?

yeah. So I think when the two by toes, I like you know, is sort of trying to understand which decisions are complex in which are simple, and then which answers are robust to mistake, which are not robust of mistakes.

And so things that are simple and actually robust of mistakes, those can already be done by machine learning and someone, things that are simple, but actually have major consequences of mistakes, like driving a car, like government can drive a car, but like if you do IT wrong to people. So that once actually tRicky, but you see people working on that was soft and cars and is a lot of work. I think the word medicine is hard is that it's something is complex and mistakes can have huge impacts.

And so maybe what we could do, we should work our way up. And it's not even a question of like should we but we kind of have to if you think about some, he's crisis are coming. And so maybe you start with nursing and and we've seen this a with hypocrite and that makes sense. You're not doing diagnosing. You're doing no harm.

you little .

yeah and so that's I been very clever. Then maybe you could work your way to P A, A physician system. Maybe you could work away from there into A G P.

And I think the general practitioner, a concious doctor, that tear is kind of a really interesting teer because largely your trading and setting off the special. So the ad didn't have to be an, and so that tear actually alone, kind of really interesting. Sn, so much of medicines done at that tier and so much of act issues of access or access to that tier.

Two, if everybody had the A I concious doctor in their pocket, I think that would actually be drained in terms of the impact on health. So even just get to that year, i'll be pretty excited. And once you're at that year, then you can imagine sort of going to specialist world, but that might be that later part might be bit off.

yeah. And I mean, to your point, this is it's a inevitability that will have to figure out a way to create leverage on the post side of of this portion of our our labor base. I guess what are your thoughts on copilot for doctors? And is is at a more near term tractable version of this that you think could happen?

I think so the whole problem copilot is can you work IT in a way that goes on the doctor workflow where they view as benefit, not a new yeah no, it's not some alert or whatever. It's it's like something where they are going to IT if you can do that and maybe ve seen this with like scribes and so on, like something we're doctors like like, hey, I want this. This is great if we can create that and maybe that is the chAllenge.

And the cult arms for founders like create something some product that like people are climbing for IT. And that's obviously knowing that space really well and knowing your customers and and knowing your people, you're going to use IT. I think if you can get in the workflow.

then I think, well, beijing, the ultimate, which is like this just embedded into the hr workplace is just there when they open and up and there's not really any need for individual physical bian. Yeah, given what we just talked about and all these grand chAllenges, what are some of the types of of startups that you know we wish that would walk through the door that we just haven't seen .

yet one year that i've been waiting for? And I think it's maybe a little early main us maybe just right at the right time is something where clinical trials uh, can be addressed with A I and and this is where it's a confluence. A couple of things.

One is, I think of so important. We talked about real world and like the ongoing clinical trials as a part of health care. But then finally, clinical trials because so much money flows for IT ah you could improve them five percent or ten percent. It's not like you have to do something heroic right to be able to do not have to ten x or hundred exit. I remember like um a decade or go equator who was um working for google and they're optimizing various ills and this and that they made IT like five vex Better.

Uh when the add filters and basically five vex was like a hundred million 到 yes， five percent was hundred million doors yeah and so I may be really job like i'm working on like something and drugs or whatever to make big leaps and bounds and small things for big cash flows can have a huge impact. So something for clinically trials could be huge or even just picking like the order of rank ordering of clinical trials to sort do Better job there. Anything in that space, I think, would have a huge impact. And we haven't seen very much part of IT is like it's maybe not where if you're outside that space, you mean not think to go? I think that i'll be my pic.

pic might would be I mean, kind of comparable in the sense of the nature of the opportunity, which is if you were to design an AI native health plan from scratch and basically be the way by which healthcare payments flow, like all the problems that we talked about earlier, what are the components of a health planets? Um a payment and mechanism and claims IT is a underwriting chasa in terms of how you score risk within a population. And then it's a network of where the providers that you would actually steer patients too on the basis of understanding, you know what inds of services they need and um the way that those are built today.

You know you see huge opportunities to both leverage data and AI in the sense of exactly what you just talked about, where a one percent impact on the cost structure of a health plan or the way that you under right risk in a certain health plan could literally mean hundreds, millions of dollars uh of uh either cost savings or Better economics to the providers who are part of those networks so that to me, kind of this notion of a full stack AI native health plan that take full risk on populations and exploits all of these data sites that we're talking about to really understand. I have almost at an individual level, and you can most sort of imagine like an individual hot plan that is like purposeful for you on the base of your behaviors and your medical history and and things like that, that is Priced entirely different than all of your employee peers who are in the same group plan versus what IT is today. Words so least comment dominate and like word of everyone loses because you're trying to design for for everyone in in the same of force fashion.

So that would be man.

be one yeah what those are some very uh, big audition ous grand chAllenges that we hope many builders go off and pursue and. With obvious ly love, talk to anyone who is working on problems of the silk.

Yeah, absolutely.

Thank you for listening to raising health. Raising health is hosted and produced me Christa atos ian and me Olivia web with the help of the bio and health team at age sixteen. The show is edited by file haga.

If you wanted suggest topics for future shows, you can reach us at raising health at sixteen z 点 com。 Finally, please write and subscribe to our self. The content here is for informational purposes only, should not be taken as legal, business, tax or investment advice, or be used to evaluate any investment or security, and is not directed at any investors or potential investors. In any sixth, please note that a six teensy and its civilians may maintain investments in the companies discussed in this podcast. For more details, including a link to our investments, please see a extensive 点 com slashed discloses。