Using AI in Healthcare: Diagnostics, Patient Records, Research, Revenue Management, with JT Perry

Episode 280 · Thu, 30 May 2024 16:00:22 +0200 · 1:02:26 · SRT · archive.org · all Tanzu Talk

When I have questions about what's going on in healthcare, especially with regards to IT and things like that, I think of you. I think maybe we spoke a couple years ago now. I lose track of time, what with life and everything. But here, this episode, we have J.T. Perry, who works here with me at Tanzu. You were on the Tanzu Value Advisor team. You work just kind of general itinerant. Remember that, you're the subject matter errant when it comes to healthcare and things like that. Talking with our customers and people we talk with about kind of applying our whole Tanzu way of thinking about getting better at software, data management, all that stuff in the healthcare world. So thanks for being on again. Pleasure to always speak with you, sir. And yeah, I'll let my ego take that introduction. It's been here about three years. And before that, involved in several companies. Primera Blue Cross here in the United States is probably the best known one. Yeah, yeah. So it's always good to talk with someone who's not just a pure thought leader like myself, who's huffing the good ideas as they go. You've actually worked on these things. You've huffed bad ideas in addition to the good ones. Absolutely. Well, so the first thing I want to talk about, like I thought it would be, let's call it update on, you'll have to tell me when we get to it, what people call this, but digital healthcare in general, right? There's always a lot going on. But what I've been really interested in recently, just like everyone, of course, I'm interested in generative AI. And I'm just going to say AI. No need to get all persnickety. You can if you want, but that's fine. But I've really been trying to figure out like what it means in enterprises beyond, well, not even beyond, but just like I use it all the time. I don't know about you. Like I have, I'm probably spending too much money on chat GPT and Gemini and other things,

but whatever. So I use it all the time for all sorts of things. And I've almost reached like, I feel like the limit of things I can do with it in my life. Like there's a lot of like work things I can't do with it because I'm a good corporate citizen and I don't like cut and paste all the work stuff in there. So like I only use it for my personal stuff and my podcasts and like public things, if that makes sense. So I don't know if it's good for summarizing meetings or like all this kind of stuff. But what I'm, what I'm mostly, what I'm also really interested in is if we look at various industries, like what people are thinking about doing with it and like the potential for all the AI stuff to help out. So that's the first like, like question is like, what are, what are people, well, let's, let's do first, like, what are they actually doing with AI right now? Like, and then we'll get to the more speculative things, like what it might be used to, but what are some applications that people are doing? So in a lot of ways, I think, uh, uh, in some of this could to be own it my own bias, cause that's where I pay the most attention, but it feels like healthcare has had kind of a head start or a significant amount of, of interest with the, the, the career kind of just explosion of, of activity and investment, uh, and even production in the AI space. And the biggest one that we are seeing like things already happen is in the radiological space. And by that, I mean, all the diagnostic stuff, right? Um, x-rays, cat scans, MRIs, nuclear, all of those kinds of things. Um, and a lot of that is because it was happening. The, the interest, uh, uh, and the experimentation was happening before kind of chat GPT and open AI and all of that, that took out, right? I think we would have called it machine learning, right? Before we decided, right, right. Two major providers in this space, both Siemens and, and GE healthcare, um, and without having

direct knowledge of their stuff, just even what they've talked about publicly and, and, and, you know, kind of dinner barroom type conversations. We know that they have been using it for a period of time to help with diagnostic. And I'll give an example, um, an MRI, you know, you go, you're, they put some body part in the tube, take, you know, thousands of pictures and put them together. And it is important to realize those are a thousand pictures, right? Well, that radiologist. So the doctor that looks at those and tries to kind of come up with a, so what about that has to look at each one of those pictures, right? You could think of it as a slide show now, there's all sorts of tools that can scroll their, their, their mouse or whatever, and, and make it look like a movie and all of that. But, but human nature happens, right? So if I was going in for, uh, an MRI to say, look at my left lung, because I, you know, somebody thought something was going on in there. You go on and say, I'm going to get an MRI. Uh, they tell the radiologist, Hey, we're looking at something in the left lung. Well, meanwhile, you know, the MRI was of your whole body of your whole chest, right? As a doctor, as I'm scrolling through that and I'm looking at your left lung, I might not see that other spot over on your right side, right? Cause, cause I'm looking at how many of these a day. And I just need to crunch through those. So that's where we're starting to see a lot of these tools happen, right? Is, is using AI to go, Hey doc, go look at that. So if I were to generalize that, cause I think you're making me think there is a generalization of let's call it enterprise AI, which is we'll use tech, tech terms. We'll use like a, some, some tech terms, which is you can scale a human's effort. And, and in that instance, like I think that example is great because well, great. It's interesting because you're constrained by the expert's time by the human's time. And so one of the behaviors that they have to do by necessity is focus on just one thing. So in a sense, like if they're looking at your left lung, there's throwing out all this

data, all this opportunity to also look at other things. So it actually makes probably, I mean, tell me if I'm wrong, but it definitely makes economic sense and probably even diagnostic sense because then you can, I don't know the ethics of radiology, but you can look at more people to help more people instead of focusing all of your time there. So the scaling is sort of like, well, what if we had like, instead of just one entity, one human looking at it, now we can kind of simulate like five people looking at this one thing to catch other stuff that's going on. Yeah, it is scale. And I would say it's both sides of the ethical box, so to speak too, right? That are really positive one, right? Let's, let's use these computers to check everything and be that backstop, right? And say, oh no, hey, look, here's something here. Maybe it's smaller than frankly, a human eye would, would pick up, right? It's a, but then there is the other side where in general, and again, not to speak for anybody that's actually doing this, they feel like it's a great assistive technology right now. It's not something that should be used to diagnose by itself. Meaning it shouldn't be the doctor. Right. It should go, hey doc, I think he broke his arm. And the doctor should go, well, yes, you broke your arm, right? That's maybe too obvious of one, but, but something like that. We do know that there are countries that don't have enough radiologists out there that are kind of basically going, eh, good enough. We'll just use this as diagnostic. Right, right, right. There is some ethical question, right? I mean, it's not, nobody's even tried to put it through like a FDA trial here in the United States, um, for, to be the sole diagnostic, uh, it goes through lots of trials to be this assistive technology. So, well, I mean, that, that brings up a side note I wanted to ask later, but we'll, we'll, we'll side note here, which is, uh, like what, like I would imagine, I mean, I mean, this gets in a weird squishy space, but like, is there like medical innovation that happens in less

regulated regions out of necessity that kind of comes back to more highly regulated? Cause, cause I could see kind of in this instance, you know, if you were in some remote place that didn't have a lot of radiologists, I mean, you could almost have like some rudimentary stuff and get like a Starlink satellite uplink, just like uploaded to somewhere. And I don't know what, right. But there's all this kind of weird stuff and, and, you know, remote health seems like fraught, except that you have to like back in COVID, as we talked about last time, like, it's just like all sorts of crazy stuff happens, but sometimes like necessity might like cause interesting innovation to happen. Well, I think that's, that's been true in healthcare all along, right? Nothing brings out innovation in healthcare, like a good crisis. Look at the MRNA technology around vaccines and all of that, right? It looked good on paper. People were really excited about it. And then COVID happened. And then now it's become the silver bullet, right? Yeah. Aspect. Yeah. We'll go too far down a tangent, but the obesity crisis and bringing in Ozepic and now us looking at the semi-glutide that makes Ozepic go and realizing all these other cool things that might do too, right? But again, let's fall down a big rabbit hole if we go too far there. But I would say in general, yes, in that I don't know that necessarily innovation in those areas is bringing it back, but it's, it gives the opportunity for a wide scale testing. And, and again, there's, there's an ethical issue about that, right? Do people even know they're being a test case, so to speak. So I don't think any of those companies would consciously honestly go, Hey, let's go give it to that rural area that doesn't have great healthcare and let's see what happens. It's more of a toll, right? Those areas, those governments go and grab this technology and use it. And, and we all look at, at kind of the outputs of that, but it is, it's a, it's an ethical mess and it's a, you know, you can have a lot of conversations over a lot of beverages on should we, or shouldn't we, um, you know, we learned a lot of good information out of this horrible scenario. Can we use it?

Should we use it? Uh, uh, big philosophical books have been written in that space in the healthcare world. Well, back, back to the AI stuff. So yeah, I can see that that was, that was, that is one of the things that it wouldn't occur to me, uh, as an enterprise AI application, because it is basically images, analyzing images that, that brings up one sort of structural issue with, with AI stuff, which, you know, just to confirm with you, it's a lot of, at least my thinking, but probably our thinking of how to apply enterprise AI is limited by the fact that we consume it as a chat application essentially. Right. Whereas that's just one way of doing AI. And as, as, as I'm thinking more about this, and just as you brought up, it seems like there's a lot of applications beyond that, that are kind of hard to see, uh, essentially. Um, so anyways, I, I mean, that kind of opens up the aperture for like all sorts of other things, but what, like, like, what are some other things beyond radiology? A lot of drug research, a lot of solutions, a lot of kind of think about the, what if type type scenario of expert, you know, what if, you know, today I have chemical A and chemical B and it gives me Tylenol. Well, what if I took chemical B and chemical C, would it give me like something to cure cancer? Right. I, I, and, and I don't even mean to be flippant about that. Like the research around these diseases that we've never been able to like crack open and solve, um, ALS, right. Those, those big ones, Alzheimer's one super, super close to me. And what can AI help us? Uh, can it give us ideas, things we haven't thought of? I don't know that anybody realistically expects to go into chat GPT and go, here's a thousand case records. Help me solve this. Right. But could it down to just, you know, almost an idea generator. Hey, this might work. Yeah. Go look down with that area or the common, well, I'm going to bring it back to a geek example that I see in our enterprises today. Right.

SRE is taking, uh, problem reports, outages, reports, test failures, uploading it into AI and seeing patterns they never saw before. Right. Oh, the core weakness of our system is this widget over here. We'll try to do that with the system. That is the human body. Right. Look across all these thousands of test diagnostics of, uh, I'll use Alzheimer's, you know, cat scans, even just reports from doctor interviews and stuff like that. Is there a pattern we're missing that that's the, the, the hope, the thing that like personally gets me excited, um, that, you know, can we make advancements where we're not, it's AI isn't going to cure all our ills. I'd be a little worried if it magically did, but, uh, you know, does it help us to start to make real advancements in some of those? That's, that's the cool, exciting stuff, frankly. Um, and, and the part of it that gets me excited. Yeah, like, like a lot of what you're talking about is, I don't really know the right word for it, but it's sort of like analysis analysis is too fancy of a word, but it's, it's almost like more of what I have in my head. It's like the, the, the person who brings a stack of papers home with them and just keeps looking through it, like kind of the, uh, the, the, the red string crazy board on, on their wall to try to look through something. And you know, that, I mean, tell, tell, tell me if, tell me if I'm wrong to limit it to this, but like in my experience and my very basic understanding of like generative AI stuff is that very rarely comes up with something new, right? Like there, there's not this like creativity so much to it as much as, as you're saying, pattern matching and like finding things that are there. And of course, you know, anticipating what the logical next thing would be. I mean, that's kind of the core of how it works, which for the scenarios you were describing, like if I were to, I guess, I guess a lot of diagnosing is pattern matching, right? See, seeing what's there. And then also a lot of just any gigantic process like healthcare is, is also paperwork, right?

Like both, both paperwork as far as running in the positive way that you're obviously, but also like just the paperwork of just like looking through data, like you were talking about looking through all this, this knowledge. So much of healthcare is I had this treatment for this thing and we accidentally discovered it also helped this thing. Right. And not to get into like uncomfortable or giggly of a subject, but, but right. The whole, uh, Viagra erectile dysfunction was a blood pressure medicine that helped this. And I just bring it up because it's kind of the one everybody knows. Oh, right. And then you also, I think, I think you brought it up earlier, but you have the, uh, that diabetes thing that helps out with, with obesity, right? That seems like an accidental discovery. Yeah. Yeah. I mean, again, we could be a whole nother topic. We're now discovering it may help with, uh, addictions, whether alcohol or, or, or, uh, opioid, um, just a crazy number of studies going on that are adjacent for, yeah, this drug that started off as, Hey, maybe we could help folks with type two diabetes. Uh, and so get, give it, given that kind of analysis and diagnostic stuff. I mean, it seems like there's a lot that you would apply, uh, like enterprise AI to and healthcare. And so like what, so then this gets to more of like what our business in the Tanzu world is, but like, what are, so, so when, when you either think to yourself or you're talking with people who are doing this, what is like the plan, right? Like they need the data, they need like the governance, they need some kind of like what I would call a product management cycle, like just discovering what to do and experimenting with it. But like, what's kind of like the toolbox that people need to just like start doing stuff with it. The folks that are doing the, the, the research stuff, right. And we, we should at some point talk about the healthcare industry and how it's, it's helping in similar ways to just like every other industry, to be honest, but, but, uh, the healthcare world, these folks are really, really good today.

And already, how do I take an idea, take or take my hypothesis, go through various trials, try things in a very safe, ethical, regulated way. That structure is there. AI is not going to, uh, nor should it frankly go around or change that process. Um, you know, yeah, yeah, I, I guess, I guess of there's, there's other industries, but of, of, of very few other industries, healthcare is very scientific in what it does. Um, for us, for, for just frankly, our day jobs in the Tanzu world, um, there's a lot of geeky stuff that becomes really interesting. And to use that radiological example, um, I can't upload an MRI to the cloud, right. Despite all our advantage advances today, hospitals don't have enough bandwidth, right. Right. I mean, MRIs are just gigabytes and gigabytes. So how do I do this dramatic AI helping in kind of a, uh, to use another buzzword of the day edge scenario, right. And how do I have a thin little office in a hospital that does that? Um, there's a lot of innovation happening in the healthcare space there. Um, that, that, you know, how do I, if I find this great idea, how do I make it practical to, uh, the hospital in, you know, a very rural area around the world, um, that may have Starlink may not have Starlink. How do I do it? So there's a, I don't think that's quite what you were asking, but I would say there's a ton of innovation and things as an industry we need to solve to bring some of this healthcare stuff to life, as opposed to the research stuff, which is, Hey, give me all your GPUs and I'm just going to crunch this to see how to solve Alzheimer's. That bit's already there because they've been using a lot of that for jump. Yeah. Stuff like that. Yeah. I think, I think that's, that's, that's, that's a, that's a good, helpful answer because let, you know, as the first two things you said, one there, I mean, the first two things are in, in healthcare, people are very good at experimenting and, and just, I guess, to make a pun, generative experimenting, where you're trying things out to find out what works and you understand how a

scientific experiment works. And then, and then two, they're also very good at regulation and governance. Now, obviously there's, there's, you know, that creates some friction and things like that, but in general, they're probably kind of like banks, one of the better industries when it comes to dealing with regulation and making it not shut everything down, which, which, which, which, which would, which would seem like it'd be great for like AI stuff, because then you don't have to worry about that stuff so much. You just kind of fit it in to, to, to those processes. And I think super importantly, medicine and healthcare research has an ethical framework too, where I think, you know, when we talk about some of the AI stuff that's going out there, whether it's the, the Hollywood, right. Making videos, making music, all of that, like what's the ethical impact of doing X, Y, or Z the healthcare world. There is a whole structure around that and process. And I'm not going to pretend that the healthcare world always gets that right. Like, like it's not perfect, but the structure is there and it gets it right much, much more than it gets it wrong. When it comes to, should we try this? Should we experiment? All of that's already there. I actually believe that'll be an accelerant for AI in the healthcare space because it'll be the word where I think a lot of other industries are going to go forward, hit a wall, have to take steps back, whether court mandated or, or whatever, uh, and then kind of go forward. Yeah. I mean, I guess, so, so that's, that's what we're talking about next is the, the, the, the risk or the, or the doom and gloom things, the FUD, right. And, you know, one of the main things people don't like about AI, which certainly, you know, if, if, if it were to affect me, I would not like it is instead of augmenting people being co-pilots as Microsoft would say, just replacing them. But it, it seems to me that in healthcare, there is such a lack of supply of skills that that would be an impossible scenario that, that basically like, we're always looking for, for new, you know, more nurses and doctors and radiologists and everyone's skilled at it that like any help you can give is like

needed and people would be very excited to, uh, have, have, uh, augmentation there. I think that's true. Right. I, I, I, the, the doom and gloom around AI and healthcare is somebody using it before it's ready. Right. We can, I can write hours about the FDA regulatory process and how long it takes, but there is a tremendous amount of safety behind it. Right. Conspiracy theorists decide it's a damn good system. It really is as far as making sure treatments are safe before they're widely used. Um, and, and, and then how about, how about like, like privacy and leaking concerns and things like that? Like what, what, um, I, I mean, you know, to, to, to, as I like to do, ask a question and, and talk some more about it. Like I was thinking recently about privacy and leaking things. And it seems like if, if you don't let what you have leak back into the basic model, then you're fine. Like, I don't really know how this stuff works. Uh, technically as far as like writing data back that other people then use, but it seems like you can just not allow things to happen. I'm not quite sure, but like when it comes to like, like healthcare data, what are people thinking about there? My, my personal risk around it, isn't that this concept of AIs and, and using right GPT-4 as a base model and augmenting it with my local medical records. I totally contrived that scenario. Um, I think as an industry, we've figured out, right. Using RAG, using these various other approaches, how do I add my local data into the mix to, you know, to the base model, but yet keep them separate. I think I don't want to say it's solved. We're on the large part. What, what I worry about from a privacy perspective is in the urgency to, to use all of this is we're just uploading, right? Again, even if I'm going to use this adjacent database, right. Using VAG and vector, RAG and vector and all these other things. Um, I, I still probably am going to upload it somewhere. And then my fear is

just frankly, somebody leaving the S3 bucket exposed, right? Just standard data security. I think that's, that's probably the bigger risk because we're all in a hurry. Right. And because I'll use Alzheimer's, right? Look, if somebody could like forget all the regulation and solve this for my mother tomorrow, I'd be really appreciative. And I might even be okay. If you leak some of her data in the process, I hate to be that flippant, but the around healthcare, a risk and why I'm really glad we have these ethical structures, we do desperation comes into play, right? Very, very much so. And so, uh, but anyway, I'm rambling. I think the issue I worry more about is folks now doing more with big data or these large data sets, uploading them, you know, or frankly, new AI providers. I, I think open AI would be great at this. So, so excuse me for using them as an example, but, but, you know, I upload my medical record database next to open AI and open AI just somehow leaves a back door open, right? I, again, I think open AI is really good. I shouldn't even use a brand, but, but I think just general data security to your point, um, folks playing in areas that they aren't used to, right? Um, uh, around this is the bigger issue. I don't think there's a leaking, you know, I could see a startup in the AI space that, you know, Hey, I can review medical records. I can review your, your Epic data as a startup, right? Uh, you know, give me access to your Epic data and I can find these patterns to, you know, doctor help you treat these patients better, which, which gets into a financial piece, right? And we can talk about revenue management in a minute for providers, um, you know, in them just being new to the space and leaking data. But again, that has nothing. I don't think there's anything inherently an AI that's going to leak data, not anything. Yeah. I know of, or I'm aware of today, but just suddenly this many more people playing with data and uploading probably large data

sets, you know, just good old, to your point earlier, a good old fashioned data security is probably the big risk. Right. Right. It's the, uh, in a good way, it's the same problems we've had forever. Like, like eventually the files have to be copied somewhere and be unencrypted for you to use them. And so you have to make sure that, that the wrong people are, or, you know, the wrong bit, the wrong thing doesn't happen to your unencrypted data. Right. But you gotta, it's, it's, uh, but there's, but there's not something inherent in the generative AI thing that requires you to leak information or it makes that a lot easier. Like it's, it's easy enough to, uh, to protect things. Well, well you brought up like revenue management, right? And I think, I mean, there's obviously, well, not obviously, but many more scenarios when it comes to the, uh, well, you have to tell me the terms, but the, uh, the content, the actual practice of healthcare. Uh, but then there's the, the, uh, running all of this, the backend, if you will. Uh, and that seems like there's probably also a lot, uh, of things you could, you could apply generative AI to, uh, in that area. So what, what are, what are, you know, what are people doing there now? And what are they thinking about? I think the big key here is it's we're now, you know, so, so we just got done talking about, I think what's unique about AI and healthcare. And now the other place that we're seeing it like in production and making a difference now is like everybody else, right? That those business process type stuff, you know, revenue management, uh, almost seems like a euphemism, but it's, it's the backend of, of providers. And when I say provider, I mean, doctor, hospital, clinic lab, the, the people doing the healthcare stuff, the back office of that is generally referred to as revenue. That's the dealing with the insurance companies. And that's the part I would, I would, you know, I'll talk about here in a second is, is the biggest example. And this is a U S uh, uh, thing for, for those of, of you in the international markets, uh, feel free to turn this off for the next five minutes and laugh at the U S healthcare system. But this is the prior authorization space, right? Where as a provider, I have to ask the insurance company for almost anything that isn't an emergent. Can I, right?

Can I please let J do this back surgery on JT? Can I please do this MRI on JT? Right. That, that kind of stuff. Um, you know, we always hear the, the horrible story of prior authorization conceptually, it's there for fraud protection. Um, it is also there, uh, conceptually to make sure you get the care you need and don't get the care you don't need. Right. Uh, uh, right. There are downsides to endless cat scans that there really are right. Right. Right. Radiation is a problem. Um, so, you know, it's things like that, but that process is incredibly, uh, manual today. And I will right. Highly skilled. Right. So as a provider, if I'm going to ask the insurance company for, uh, uh, permission to do a back surgery, I have to know all these codes called ICD codes and let's not go into that, but just know I've got to get the code, right. I've got to get the text, right. Kind of the justification, the evidence, why I want to do this. Um, right. And a doctor, a certified doctor has to be the one to do this. Right. So I can't, I can't be a doctor and put into the, the, my medical record system. I think JT needs a back surgery and then have, right. Right. In a legal system, we've got the concept of the lawyer and the paralegal. We, we don't really have that for, for prior authorization. Right. So there's entire management companies that have doctors that never see patients. So when you talk about capacity, right, we've got a huge number of doctors, both working on the insurance company side and the back office revenue management side that are doing this. So a lot of prior authorizations get denied because they use the wrong code. They forgot to attach the, the, the, the, the test result or, or they didn't frankly use the magic words in the, in the, in the justification, right. For, for all of that. So the generative AI, the AI you are talking about is really coming to bear there and, and helping clean that up. Um, R1 RCM. Yay. What an acronym. It's actually the name of a company that, that does a lot of this

revenue management stuff. They've really brought AI to bear in this and they've, they've been able to, uh, greatly reduce the rejection rates that are due to just paperwork, which is 90% of the rejections by the way. And I think that number is actually, right. We always hear about the horror they won't let my grandma have the cancer medicine she needs. That's a different story for a different day, but most prior authorizations, time and, and delay and rejections are, are paperwork errors, right? This is where just making a tremendous difference on accuracy, speed, and reducing the number of doctors that are doing this. So I, this is a little bit of a thought experiment, but maybe those doctors can get back in the caregiving game, right? Like it's really hard to get a physical, an appointment to get a physical here in the United States. You know, it would be good if some of those doctors went back and did that. I don't know if they want to, anyway, a lot of, a lot of stories, but that is a space where, right. We're already seeing a material change using generative AI, creating the, the, the letter, so to speak, it's all electronic, but, but letter, uh, getting the right codes and getting that prior authorization. Tremendous, like real material advantages there already seen now. Yeah. It's like a, another example of, uh, having the co-pilot, having the augmentation of what you're doing. And I, I was, I was watching a, uh, to, to try to be a little smart on this topic. I watched a few videos and there was one from some UC Davis people talking about this topic. And I forget the person's name, unfortunately, but they brought us a great sort of, uh, I love these kinds of like scenes where they're like, you know, anytime you go see a doctor, there's always a keyboard involved and they're like typing things and like mousing around. So there's, he didn't quite put it like this, but there's, there's almost like there's a keyboard between the patient and the doctor, like time-wise and sometimes physically. Right. And like that seems to use a layman's term, like a bunch of nonsense that doesn't need to exist. And it also seems like, so, so it seems like a good example where you've got, uh, I don't know the terminology,

but I think like training and reinforcement on both sides of the transaction where the, the, the generative AI can like try to use the right words. And then on the other end, you can verify that like, these are the right words and you can do some kind of like random checking in and to make sure that like, what would you call it? Not the right, right words, but they, it didn't make up that these are the right words because it wanted to try to do this, even though it was unjustified. And if you allow this kind of like going back and forth, like it was, you're, you were making me think of this scene from, uh, you know, I don't know if you were a big Mad Max fan, but it was a TV show and, and, and they, they, uh, of course it's the dystopia. So I'm going to ruin the coolness of what we're talking about, but there's a scene where someone has to go to court and the two lawyers just put like a little to their own discs into a computer and, and then the disc, like, you know, argue the case on either side, but a positive version of that kind of, it's almost a way of reducing like, Oh, this was a bunch of nonsense we were doing. Uh, and because I would assume, because a lot of it has to do with trust is unfortunately an overloaded word, but there's not a way to establish trust between the insurance for a provider and the doctor where they would just trust the doctor would say like, yeah, they need a back surgery. And the insurance provider would be like, great. You, you actually, so I'm going to say, right. And it's funny to use that Mad Max analogy of the two discs kind of arguing against each other. So the payer side, so the insurance, we talked about the revenue management side, the insurance company side of this prior authorization is also using AI to try to automate many more of these decisions, right? A lot of the were automated before that. And when you talk about trust, I would say, yeah, it was, it was automated trust, right? This doctor repeatedly has asked for this surgery. Every time we've approved it, we've never detected any indications of fraud, right? There isn't this guy that's required, right? You know, you start to pre AI, right? All of this will go back to pattern matching, right? Well, just the same way that, you know, banks did credit card

fraud for the, for, for years, right? Um, is this one person constantly like requested like 90% of the back surgeries we see, uh, we might want to ask some questions, right? Right. Is there, is there another back specialist that boy only asked for surgery for one out of a hundred of his patients? The rest of the times we always see him walk the protocol, what we talk about, right? Hey, let's, let's try the muscle relaxer first. Let's, let's take an x-ray. Oh, let's, let's try physical therapy, right? If somebody always follows that. And generally, by the way, it gets good outcomes. Cause as an insurance company, I know whether or not you got healed or not. Um, then you know what? I am just going to automatically. Yes. There's just a human's never even going to see it on the insurance side. The ones that humans see are the odd cases are the ones that get kicked out of like, wait a minute. Right. Um, or, you know, part surgeon requests to do a back surgery. You know what? We should probably ask some questions, right? Right. About this, that kind of thing. But seeing AI and be able to do that at a much bigger scale. Now, this is a place where it's not in production yet that I know of. I should put that big caveat in there, but talking to friends back in the insurance world, and especially in the blues world, seeing a lot of experimentation where they're, uh, uh, having the, the AI system make the, the, the first decisions on this and then having humans look at them and go, Hey, when I made the same decision, that that's kind of where we're at right now. Um, I don't know whether to call it experimentation or training, uh, you know, the model and things like that, but it is setting up this case, right? I do know of a case of where somebody is using this on the insurance side as an experimentation and they're getting a lot of R1 or CM requests. So in my head, it is that exact scenario you talked about, right? These two AI bots fighting over and deciding whether or not you should get the back surgery. I, and that could be creepy. That can feel very uncomfortable. Like, wait a minute, why are the computers making these? Yeah. Well, right. That's, but, well, I think, I think in the, in that scenario, what you would want instead of like the max headroom thing is, you know, to make it, to stack it, to be absurd, you would

almost want a third AI or, or a third disc you put in there that was like, all right, now you were going to standardize how this is done. So you don't have to do that in the first place. Right. It's just like, we're going to come up with a better scheme because, you know, uh, I, I, you know, as, as a U S person, like you would kind of know this intuitively, but maybe other people is like the problem you just described assumes a one-to-one relationship between provider and a health insurance company, but we have a uncountable to uncountable relationship. I guess you could count the number of insurance providers in the, in the States, but like, there's a lot of different things. So not only, I mean, that's where things get crazy is there just so many combinations of, of that single discussion that happens that you do need, you do need some augmentation to deal with all of that, that those con is a permutation or a combination. I was bad at math, but all of those various, uh, mix. Yeah. I mean, look, sign me up to burn the current U S healthcare system down and, and, and come up with a new one. Um, but we, we've got to make do with the one we have. Um, and yeah, a lot of those problems that you just talked about, right. Are so many tech firms or, or large retail firms have thought they could just crack this healthcare nut and they haven't. It's been a train wreck lately of, of, of, you know, folks try new things in the healthcare world from a business model perspective and absolutely nothing stuck the stickier. So, so as a transition to like general digital healthcare stuff, you know, as, as we're talking through this, one of the things I was wondering is like, who, who does this innovation in healthcare? Like, I don't, I mean, maybe I'm totally wrong, but I don't, I don't think I wouldn't imagine that an even kind of a medium, large size hospital system. They have a lot of programmers who are like writing applications and doing this. And so like a lot of, in my mind, a lot of what we talked about is just like, here's, here's like, I mean, generic, and here's a general purpose, generative AI engine.

And now we need to make a bunch of, we need to write a bunch of applications. And that might be like, you know, not the most beautiful AI, beautiful UIs, but like some kind of application development you're doing. But where is that? Where's that done in the healthcare system? So most of the healthcare back office stuff is, is outsourced these days, right? Or vendor sourced outsourced as a weird connotation to me. Um, right, right, right. Like, like this revenue management thing that I was just talking about that providers and hospitals do, they generally contract with firms to do that. Now there are exceptions, the very large, uh, uh, healthcare systems in the world are in the U S right. Uh, I'm thinking of the HCA is, which kind of covers the whole Southeast United States, uh, with, with hospitals, excuse me, with hospitals and clinics. They're big enough. They have a lot of their own developers. Uh, like, frankly, they're doing a lot of development for stuff that we talk about, right. Backstage development and things like that, even. Um, so they're actually themselves doing, uh, uh, you know, a lot of this, but most, right. When you think of hospitals, back offices, these business processes, there's vendors that are, that are doing this innovation that are, are right. And they're doing it for their own selfish needs of, Hey, if you're still going to pay me a dollar per prior authorization, it's not that, by the way, I just make that up. Um, you know, and I can do that without having an adopter in the mix. I can make a lot of money. Um, that, that kind of thing. So who like, like in the, at least in the States, like who are those vendors? Like I know Epic is one, but that's it. That's, that's my novice understanding, but who like kind of what's the set of those vendors? So there's the EMR providers that, that, you know, we, we, you talk to in the hallways of conferences and stuff like that. And they're trying to experiment with it. Epic, Cerner, Meditech, those kinds of people. Um, R1 RCM is, is a revenue management. I hate to keep using that. Uh, good friends, that of architecture there, um, is why I know them so well. Um, you know, things like that, the insurance companies, they're doing this themselves. They, they, they have traditionally

done a lot of software development and experimentation. Right. That would make sense. Yeah. But then we're also seeing like the big SIs, right. The cognizance, the Accenture's, Deloitte's, all of those folks are doing a lot of this and trying a lot of this and then, uh, bringing it back to their customers. Um, that's where we're seeing a lot of this happen right now. And, and then like, uh, again, in the scope of, of the, the, of the U S like, do do things like the VA and, uh, I always get Medicare and Medicaid mixed up unless I think about the little poem about them or whatever, but like, like, if you look at those big, I, I wouldn't call them like shadow healthcare, but like people don't always think about those as huge monolithic healthcare institutions. So they have like their own, yeah. I mean, innovation staff, their, their own development staff, or do they also just work with these other providers? The VA works a lot with providers. Now, I think that what I don't know, and actually it's your, your question just made me want to go ask questions and research over the next couple of weeks. Right. There's been in the U S the whole move to, uh, a lot of divisional development by the business, by the government itself. Right. Whether it's stuff we do at VMware with some of the military branches, um, you know, in Tanzu world, right? Like software factories, right? Software factories and things like that, but also just in other parts of the, of the government as well. Um, I don't know if they're making an impact on, on the VA, but the VA generally big exception here has always been very vendor sourced. Um, yeah. Cerner being the one that that's, you know, having a little bit of a bump in the road, trying to roll out a new EMR there. As a matter of fact, a little bit is a, a big bump in the road, trying to roll that out. Um, but, uh, so I don't see a lot of government activity. Medicare, right. Is a more of a set of regulations than it is an insurance company. Uh, uh, news to medic, I guess outsource to various insurance companies under the guise of Medicare advantage. So, yeah, yeah, I guess, I guess the, the Medicare and Medicaid, you're, I could imagine there are a lot

more of the, uh, revenue management is, as you say, just sort of like managing that back end of stuff. There are as much regulatory frameworks as they are actually at. And so most of that work, the thing I just talked about, right. The revenue management and the insurance company, that's the same for Medicare and Medicaid or commercial that you get through your employer, or maybe even individual, individual insurance, you buy yourself that struck those structures and how that money flows is all the same. There's just a regulatory framework and some price controls around it for Medicare and Medicaid. Frankly. Okay. Yeah. I'll have to, uh, have to, it'd be interesting to see like what in the UK, like what the NHS does and how they're structured. And I think, you know, I think where I live in the Netherlands, I'll have to ask about that too, because the Netherlands has this weird, well, I I'm sure the Dutch love it, but you know, if compared to the system, we just went over this kind of weird in a good way system of, I don't know, there's like five or six insurance providers. And as far as I can tell, they're so regular, they're basically the same, uh, but whatever, you know, they've got that interesting hybrid system, but anyhow. So, so the other thing that I, you know, I was curious to ask about is just like, in addition to AI, like other, whether it's like, well, let's, let's call this business innovation made possible by digital stuff. Like, like that's, that's interesting to go over. And, and the thing that brought this up is I, you know, I was just looking through my, um, whatever, whatever, I guess you, is it called ham? Whatever it is between email you like and spam, like promotions that you have. And I got one, of course, well, of course I got one from the Amazon prime going over like their prime healthcare. And I was looking through it and it seemed like, I don't know, it seemed interesting, but then, but then also I know that like, you know, we even talked about this a couple of years ago, there, there have been some experiments like at Walmart and other places to have, and I think CVS had some of these to have sort of like, you'll have to tell me the term for this, but like mid tier, you know, the middle of healthcare stuff done more affordably. So like,

well, first of all, what do you call that whole area? And then like, what's going on in that area? Like, is it working out? Like, what are the, what's up with it at the moment? You know, six questions in one there. Um, the, the, uh, key answer to what you're looking for is dock in the box. Um, and that's where the, you, right, right, right. Carve out a corner of your Walmart, frankly, not frankly, but, but carve out a corner of your Walmart and set up a general practice clinic. So it's something more than urgent care. Although some of those were set up as urgent care too. Like, you know, I, I, I cut my hand or, or, uh, hit my head. Right. Right. Some of those kinds of things. Um, but also to be more general practice. So the one thing we know in the United States is there is not, and you and I briefly touched on this earlier capacity, right. Especially in the rural areas to just get folks in, to get a physical, to, uh, you know, checkup, or even when they don't feel good to actually go see a doctor. Um, because if I can see you early on and treat you, I can keep you from getting worse. Right. So, so it's actually, even if I'm only going to look at it from a system dollars perspective, I would rather much rather you get into the doctor more often, frankly, like that's okay. Um, I'm going to, I'm going to watch how much of the big fancy procedures that are really expensive you get, but I, I really want you to go to the doctor fairly often. And that's not very possible. Like it's even in a very metro area that I am right. Getting an appointment to get a physical is very hard. So the idea that, uh, um, the Walmarts in the world, the Walgreens that already have footprints in these rural areas, uh, in a large number, if they could stand up these clinics, the concept was, boy, we could increase capacity. But then if I'm Walmart, if I get big enough, I can then negotiate with insurance companies like I do with the maker of paper towels, right? Like Walmart has a huge advantage of being able

to go to their suppliers and go, no, you're going to, I'm only going to pay 50 cents. I'm not going to pay a dollar and you're going to like it. Right. Because I'm so big and I'm so key to the market and things like that. Kind of inverting that to the insurance company and going, no, no insurance company. I'm so vital to you being able to sell your insurance because I'm so much a percentage of your doctors. You're going to pay me a dollar instead of 50 cents. Like that was the quick example of how this would work. And it just, and I don't know why I, I, I mean, I can be snarky, right. You know, they just couldn't get in there. So, but Walmart just in the last month announced they're shutting down this, this entire concept and that's huge, right? And it's having some secondary for some big vendors, by the way, to think EMR providers, cloud providers and stuff like that. Um, but, but that Walmart's doing this Walgreens tried the same thing. Uh, they bought village MD, which was these, these doctors that would be in shopping centers and we're going to bring them into wall there, their Walgreens clinics. They've had to walk that back just at, it hasn't been financially, uh, uh, really good. It hasn't worked. Um, and so they're shutting that down. There's been a lot of these retail companies that I'm going to call Amazon a retail company for the purpose of my discussion here. Um, I know there are many things, right, but our retail companies trying to solve this problem, I think because a little bit out of arrogance, they thought they could do it. Nobody else would. Some of it's their scale. We're big enough. We can really do this. It just hasn't worked. We've seen over the last six months, especially so many of these things being walked back and, and shut down. And just, it's not very profitable. You could look at CVS, Walgreens, uh, uh, Amazon, uh, in the past with, with several iterations of this and go back to Amazon and Haven, right? That was their, uh, partnership with Berkshire Hathaway. And was it JP Morgan? Excuse me,

JP Morgan? Yeah. Where they were going to solve insurance and that just didn't work either. It's such a regulatory mess. I mean, uh, you know, I, I joke that we have to burn it all down. You're, if you're going to make a material change to the U S healthcare system, it's probably political and regulatory. That's going to do it. And we don't have the environment for that to, to come back and talk about Amazon health, right? Which is what they just have been, been, uh, not spamming, but hamming, to use your term everybody about, uh, um, of late. That's kind of an iteration on Teladoc, something that existed before COVID, but really exploded during COVID and became normal during COVID. The idea that I could chat with a doctor on my phone, um, and get some basic medical or, or maybe do a video call, not, not different from what you and I are doing here with a doctor for many things. The doctor doesn't have to touch and feel, um, from, from that perspective, maybe it's cold and flu. In fact, if I'm the doctor, I don't want you in my office. I do not. Right. Sure. Spread it. Yeah. So I'm interested to see if this works. Um, and it works just by the meaning, can you get enough people to use it? Um, I will say, I assumed Amazon was going to require you to use their pharmacy and, and, and deliver their drugs and they aren't. That is very, very interesting. Um, I'm, I'm sure you can use Amazon pharmacy. Um, that has been something that's been pretty successful for Amazon by the way. And the healthcare is they bought pill pack five years ago now. Boy, don't quote me on that timeline. Um, and I've really done a good job of, you know, sending you your blood pressure medicine every month and all of that kind of stuff and making some good money at it. But, but trying to see if this, this, this Teladoc method works, uh, I hope it does. I will say I use a variety of it. I, a different vendor. Um, if you've ever heard, we talk about healthcare. I've, I've tried to go this whole conversation without mentioning zoom care, but I'm going to mention it once. Cause I'm a fan, not an investor. Um, that's a really kind of a hybrid of the two. Um, and, and so it works. And again,

if we can get folks that don't have regular access to doctors to use that, it can be fairly inexpensive. The magic bullet for Amazon will be, can they get the insurance reimbursement? So will the insurance company pay Amazon for that visit at a rate, not the same as you going into a doctor's office, but enough to make money. That always seems to be the crux of this. Yeah. Well, you did a good job of responding with 10 topic ideas to my six questions. So, so that's, we, we, we, we matched up there. And I think like, like, uh, of, of many of the things there, what's, what's, what's interesting is to try to figure out, I'll, I'll say constraints, but the constraints to deal with there, uh, because, you know, the first one you brought up is, I mean, it's one of the, um, the old micro Porter five forces of strategy is the relationship between supplier and supplied to, right. And, and like, if you, if you as a company or an organization control your suppliers, then, you know, that's kind of, you, you know, these, as you say, dock in a box we're going for is that if we're a giant company, our suppliers, um, we can command a lower price from our suppliers. And, you know, the positive version is also, we can maybe drive some efficiency and standardization, right? Like the suppliers sort of kind of would like to sell to less people to have more of a guaranteed revenue stream. I don't know. I mean, that all depends right on, on, on, on, and again, that dynamic of control between the supplier, the seller and the buyer. But so there's, there's like that one constraint, but then as, as you bring up, it didn't work. It wasn't like a magic fantasy, which could mean all sorts of things, but you know, that, that is really interesting that you can't get the efficiencies of paper towel buying for cold diagnostics or something like that. Right. And so, and yes, then the other thing

that you bring up that is interesting is, you know, you especially see this when you have kids and you learn this when you've had like a second or third kid is really, you just need to call and say like, my kid has 102 fever and they're like, call me back tomorrow. If they have 102 fever and it's having a third day, then actually bring them in. Right. Whereas, you know, I don't know about you, but we had your first kid just like, Oh my God, he seems, they seem sluggish. I'm going to the doctor. Right. Uh, and so there is kind of that screening that you would hope like in telemedicine and doc in a box stuff that you could just, I mean, especially in healthcare triage means something a little more dramatic, but you could kind of like filter out like, like things that are easier to solve that don't consume your constrained resource of the healthcare people. Yeah, for sure. And the idea that, look, I can prescribe you this and that's probably going to solve it like with, with a 90% confidence. And it doesn't have a downside of trying, right. That's always the other side of I'm just going to like, have you go take that medicine and try it. And, and that's probably going to solve you. And it saved you from driving and it saved you from standing in another crowd of people and giving it to other people. Right. But it also frankly gets people to come and talk about things that they may not want to. So the idea I can, I can say things to a chat bot that I might not want to say to you in a room. Right. And again, this is where we get into things like sexual health and all of that. And while they're, they're, they're tongue in cheek and people like to giggle about them, they're really important. Right. But there are other medical issues too, right. That, that people are just frankly embarrassed about their bodies or whatever. And if you can get them to open up in a remote sense, we do see people say things in this environment, like you and I are that they wouldn't necessarily say in person or being comfortable with, with a person. And as a doctor, that's a very positive thing. Right. You might say something that has, Hey, I know you're uncomfortable, but I really need you to get in here. And that's a viable outcome to that. But a lot of times, if I can get people to

talk about them, I can get them happier and healthier, right. Mental health is just as important as physical health. Right. And so. Right. I'm hopeful, right. This is the case where I, I, well, I generally want them to, I don't want to spend Amazon and we more normalized. Yeah. You know, that's, that's an interesting one maybe that to, to, to, to bridge to other interesting things going on in healthcare is the, the, the mental health angle seems almost like, I mean, I'm, I'm, I'm, I'm well past the thing of, of doing the, uh, the, the anti-pattern of some outsider amateur trying to diagnose a complicated system. But that said, as a consumer of the complicated system, like, you know, uh, until COVID stuff, like the therapist I used to see, uh, back in Texas, like wouldn't see people online. And then they had to, and they realized it's totally fine or, or, you know, or, or at least it's better than a lot better than zero, uh, so to speak. And so I don't know if this is true, but it seems like the supply of therapists is maybe less constrained than the supply of doctors or, you know, I don't know, maybe it could be, maybe that's just like a supply and demand problem. Like not as many people see therapists as, as would be helpful. So there, there's a better equilibrium there, but it all, that, that, that seems like a good experimentation area that you could get a lot of learning about how to do, I guess it's remote stuff, but also just the efficiency of it, because I think it's also, I mean, the amount of money that I pay my therapist per session is, is relatively low to back surgery, right? So the, the, the, the backend part seems like it's a lot easier than, uh, than, than others. Yeah. I, I, I, you know, there were more therapists than, than, than medical doctors and I would strongly disagree. I would say, uh, one of the top five biggest problems we have right now in the U S anyway, I don't know, uh, how this plays in Europe or, or, or Asia or other places is folks that want mental health treatments and, and whether

that's just a therapist, just a therapist or even something more right. Uh, versus the, the supply is really off. Like it is really hard to find a therapist, to find a therapist that your insurance covers. That's a very U S thing. But, um, and then get an appointment when somebody is in like distress, right? Because it's one thing to say, well, you know, what would really help me is some therapy. Okay. It's going to take me three weeks to get an appointment versus a teenager or, or somebody else who's been really some distressed environment, right? Having some very ideation type thoughts and stuff like that. And then it takes three weeks to get and find somebody. Um, and it's all right, right, right. You need like urgent care for therapy basically, which is, uh, yeah, yeah. Yeah. Okay. Okay. Now that's, that's, that's good to know. And I mean, I guess it gets back, you know, the, the, the question I was having about constraints is it seems like the number one constraint is just supply. Like there's just not enough people, uh, who can supply the healthcare that you, and so, you know, just to repeat it, it goes back to what we're saying. It's any way you can augment that supply, there's going to be huge advantages that you have. And then you also have the, the, I don't know if it's compounding, but you also have all of the, uh, the bottlenecks and friction of regulation in that very many to many complicated system we were talking about that just kind of piles things on top of it, but maybe to be a little more optimistic, like what, what are, what are other areas in healthcare that like, you know, if you look at like applying it or just like better software or whatever that, that you see people, uh, being successful with. Yeah. I'll come back to some of the stuff we already talked about that really does have me excited, right? The research side of this, that is starting to be really cool. Um, I think the training aspects of this getting doctors to be able to do more intricate, harder things and giving them experiments. Uh, uh, I don't know how super

public it is, so I'm not going to use their name, but there is a vendor, uh, that does a lot of robotic type surgery kind of things like, like tools that can help a doctor that is kind of now like using technology, all the buzzwords, right? AI and all sorts of stuff, but to, to allow kind of simulations of that, like how can I, as a doctor don't necessarily get the touch and feel, but walk through that intricate cardiac surgery, removing a blockage in a very touchy space and things like that, and allow me to do it or allow complications to pop up in the middle of that. Right. Um, uh, VR. Oh, right. Right. That's the acronym. I couldn't pull out a second ago. Um, using some of that. There's a lot of cool things. I think, uh, personally, we're going to make some real advances in some really tough areas. We haven't been able to crack before or, or, or fast enough, whether it's some cancers out there, I'll use the one that I'm passionate about right now, Alzheimer's, right. ALS, some of those things, I really bet in the next, you know, four to five years, we see some real, but no, I don't want to say cures that that's but, but some treatments and being able to really make some steps forward where we haven't before and give people some real hope and, and positive change in their life environments. And then I'll come back to as epic, right. To talk about once. Well, somebody's being the type of drugs that is that's the diabetic drug that, that became everybody's using for weight loss. And now we're finding all these secondary and tertiary, uh, uh, benefits too. But if we can really start to solve some of the obesity issues in the U S and some of those are genetics, some of those are behavioral, all of that kind of stuff, we can really start to impact the aging curve. So all of that takes technology behind the scenes. Right. And, and, um, I'm super excited to, to kind of start to see some impacts there in the next four, you know, like I said, kind of five-year horizon, which is what I think about. Um, it could be really, really cool. Yeah. You know, there is, I don't know if this is

necessarily the case, but at least as an industry, it seems like, you know, like going all the way back to the beginning, we were talking about how, when it comes to learning how to apply generative AI healthcare is pretty good, pretty well set up because they already have an experimental mindset in place. And then there's all, it's also already, you know, they know how to work within a governed framework and, and how to do things carefully like that, which, which, you know, little startups don't, you know, like, like, like all those things. Then, you know, there's to think about, so I was asking where the innovation happens and it can be a drawback. If you have a handful of people doing that innovation, if they're not innovating, if they're not changing, however, if they are working on new ways of doing things, there's a benefit because they, if you have, I'm just making the numbers up. You have five people supplying the software, quote unquote, everyone uses. That means once, once those five people figure it out, they can quickly get it out to everyone. You don't have to rely on independently doing that. So that has a huge advantage to it as well. And, and, you know, I think the other, the other advantage is just like, this is overstating it a little bit, but a lot of those tools are just ready to be used and people can just start experimenting with how to apply like generative AI or other things. They just sort of like have to will it. They have to decide to do it and start to deploy it, which, which I think is handy. And, you know, to, to be a little crass commercially, right? Like that is not some of the, some of the people you mentioned, I know they use the, uh, you know, our Tanzu, uh, application service or Tanzu platform. Uh, and so it's a good example again of that, that limited supplier thing where generative AI stuff and managing all your application workflows already built into that. You know, you can just start using it. And about this enterprise AI stuff who are using that in, in, in the Tanzu platform, other ones, it's just in there. So you can decide to start doing it. You don't have to, uh, you know, download one of the models and figure out how to do everything with it, which, which is handy. It's, it's unlike some other

innovations in the past where you really do have to do it all from, uh, from scratch, uh, which, which is helpful. Well, as, as always, this has been great. Uh, you know, you, I, I know you've got a, uh, where do you put your stuff nowadays? Probably in LinkedIn, like, like, like everywhere else, where, where would you point people? I did. I still hang out on the X occasionally, um, or, or what are the, the site formerly known as Twitter, uh, uh, and blue sky as well, but circle JTP will find me on most social media. Uh, and then JT Perry under LinkedIn is, is where I'm doing a lot of stuff as well. Yeah. I, you know, I think for most, like most people, Twitter is now your ex social media location instead of your, your current one. Well, with that clever turn of phrase, uh, as always, this has been a Tom's we talk. If you want to find show notes, uh, you can either look at the video where you might've watched this, or you can go to Tom's talk.com and look this episode up. And, uh, you know, it has been a while. If you're listening, uh, in serial, it has been a while since we published something, but I've got several more things lined up and in the works. So, uh, there'll, there'll be more return to what we're doing here. And as always, thanks for listening. We'll see everyone next time. Bye-bye.

Tanzu Talk · 🍕 Get more pizza.