The Failure of Ultra-Stability: Robert Pearl on Why American Healthcare is Quietly Rationing Us to Death

00:00:31 Andrew Keen: Hello, everybody. As artificial intelligence continues to revolutionize the world, one of the sectors it seems to be having the most profound impact on is the medical profession — the American health care sector. And our guy, who I've always used as a reference on American health care, is my guest today, Robert Pearl. He ran Kaiser Permanente for eighteen years. He is the author of Uncaring: How the Culture of Medicine Kills Doctors and Patients — a big hit. He's a podcaster, a writer. He's written some interesting things in Forbes recently. He's talking to us from his home in San Diego. Robert, as AI begins to change everything, you've been pretty outspoken in your argument that health care is an ideal sector for AI to revolutionize. How has the American health care system changed over the last few months with generative AI — with ChatGPT and Anthropic and Gemini and all the other platforms?

00:01:47 Robert Pearl: In my recent book, ChatGPT MD: How AI-Empowered Patients and Doctors Can Take Back Control of American Medicine, I highlight that the tools are mainly being used for administrative tasks. I wrote that several years ago, and it's just as true today. I think there's a massive potential for generative AI to completely transform and revolutionize health care. I'll give you a very simple example of what's possible. We know that chronic disease affects seventy percent of Americans, and we know that if we could effectively control chronic disease as well as the best medical groups do today — and they do, with a lot of people, though those people are very expensive, which is why the cost savings are minimal despite improvements in clinical outcomes — but according to CDC data, if we can effectively control chronic disease, we would avoid thirty to fifty percent of heart attacks, strokes, and kidney failures, probably cancers as well, and we'd save over a trillion dollars. To me, that would be transformative. Right now, what we're seeing is that the cost of medical care is rising twice as fast as our ability to pay. Whether we measure GDP, wages, or overall inflation, health care costs are going up between 7 and 9% every year. I don't see anything on the horizon that's going to change that until we shift how health care is delivered. And that's why I am so bullish on generative AI and its opportunities for the future. What I see is a lot of potential, but I'm not yet seeing movement, which is why I've written a series of pieces recently about why generative AI tools need to be implemented as soon as possible. Because if not, what we're seeing already is restrictions and rationing. We saw people on the exchanges see their premiums rise. Medicaid costs are doubling. We know that in the future we're going to see Medicaid losing millions and millions of members, and no one should fool themselves — Medicare and private insurance are right behind it.

00:04:10 Andrew Keen: So, Robert, you've been saying this for a few years now. Ever since we started talking, you've been very bullish. You understand AI as well as anyone, certainly in the health care industry. As I said, you ran Kaiser Permanente, so you're all too familiar with insurance companies and systems. But from what you're saying, it sounds to me like your opinion is pretty pessimistic — that for all the new technology, the health care system is mostly failing to use it to address its fundamental problems. Is that fair?

00:04:50 Robert Pearl: I think it's half fair, if that's a phrase one can use. Let me start with what the data says: 40% of Americans are using this technology every month to help them understand and manage their medical problems. The data also says that 70 to 80% of physicians are using it every week to help them provide better medical care. So it's not that it's not being used — it's that at the systemic level, it's not happening. Now, you and I are both familiar with the business world, and we know that when industries fail to embrace the potential of technology, we can look at Kodak. It had the first filmless camera. It could have dominated the industry, but its model was to make money by selling film. So it didn't do it, and basically it's out of existence today, although there's a tiny remnant still sitting in place.

00:06:01 Andrew Keen: And I actually wrote a book about that. I even went up to Rochester, New York to look at the old Kodak offices. It's a very depressing experience. So I'm all too familiar with them.

00:06:10 Robert Pearl: So what I see is — it's going to happen, because disruption can't be stopped. It can be delayed for a while. It was probably twenty years between Kodak having the technology and finally being disrupted. And I think that's what we're going to see in medicine. Pessimistic is the wrong word for me. I'm pained by the fact that it's going to create massive dislocation in the transition, as it always does with disruptive technology. I fear that patients are going to get caught in the middle, clinicians are going to get caught in the middle, and people are going to get harmed in the process.

00:06:53 Andrew Keen: So maybe I'll change the word — rather than saying you're depressed, you're disappointed. What you're saying, Robert — correct me if I'm wrong — is that on the one hand you've got this stealth revolution, with large proportions of doctors and patients using AI to try to figure out what's wrong with them and improve their conditions. But the system itself is not adapting. The old system of hospitals, health care, insurance, and pharma is all staying in place. Is that what you're saying?

00:07:31 Robert Pearl: Absolutely. I call it ultra stable. Health care doesn't move. Hospitals have stayed at between 30 and 35% of total health care expenses. They've gone up at 7 to 9%. Doctors are charging 7 to 9% more. Drug companies are charging 7 to 15% more depending on the medication. You see little tweaks here and there, but it's ultra stable. If you could take a biopsy — and clinicians love a biopsy, to see changes in a lesion — and you biopsied American health care in 2010 and then again in 2026, I don't think anyone could figure out which slide was which. Both were overpriced. Both were underperforming. Life expectancy hasn't changed. There are four hundred thousand people dying from misdiagnoses every year. Burnout is sitting at around fifty percent amongst clinicians. I can go down an entire list. In fact, my first book, Mistreated, published a decade ago, outlined the problems. If you read them today, they're exactly what they were a decade ago. We have made, Andrew, no progress that I can see in making health care higher quality.

00:08:53 Andrew Keen: It's a —

00:08:54 Robert Pearl: — rhetorical —

00:08:55 Andrew Keen: Yeah. Robert, what you're really doing, I think, is observing the nakedness of the emperor. Most people are not smart enough or brave enough to actually tell the truth about what's happening. And it reminds me of the same thing happening within the university system, which is also profoundly overpriced and underperforming, and yet no one's willing to acknowledge it. AI is being used by students, probably by faculty — though a lot of them won't admit it. But the universities themselves don't use it. So in a sense, this stealth revolution on the one hand and this ultra stability in an archaic system are a feature broadly of the American economy. It's not just health care, although health care seems a particularly troubling example. Is that fair, do you think?

00:09:50 Robert Pearl: I think that's very fair, but I want to point something else out to viewers and listeners: this technology is three years old. OpenAI released ChatGPT on November 30, 2022. It's basically three years old, and we're still trying to figure out what it's going to become. This is a technology that's doubling in power every year. If you look at the internet, there was a good twenty years between it becoming available and fully becoming functional. The amount of progress that's happened with generative AI in three years — I can't think of an example anywhere close to that. Everything else has taken decades, and in some cases centuries. So this is where "disappointed" is a great word, because I'd like clinicians to lead the way. If they lead the way, they'll put the best system in place. If they don't, what we're going to see is that the tools being created will be driven by the economic needs of the manufacturers, the technology companies, the drug companies, and the hospital systems — not by what's best for the doctor-patient relationship.

00:11:13 Andrew Keen: So what has to happen? We've been talking on a number of shows recently about revolutions of one kind or another — socioeconomic and political. At certain points in history, systems just collapse. They break. What would a broken health care system look like? Is it already broken? We always hear these dire predictions of the collapse of American health care, and yet it stumbles onwards, limps onwards. Is there a moment where it literally falls to pieces?

00:11:53 Robert Pearl: I don't think so, if you're looking at it from an industry perspective, because what it's done is what a lot of other parts of the K-shaped economy have accomplished. It's still making a lot of money — in fact, a massive amount of money. It's just making it off a smaller proportion of the entire population. What's missed in that equation is that unlike a lot of consumer goods, health care is essential. When half the country can't afford it, you can still make profit off the other half — like the airlines. They're all moving from economy into first class. That's okay if you want to drive instead of fly.

00:12:37 Andrew Keen: Yeah. Now they're moving into super first class where you get your own cabin.

00:12:41 Robert Pearl: Of course they are. And hotels are going to do the same. But those are opportunities you can turn down. You can't turn down health care. So what are we going to see? We're going to see rationing. Increasingly, it's going to be harder and harder for a larger segment of the population to get the care they need. We're going to see people dying earlier. A lot less health, a lot more disease. And as long as we want to ignore it, it won't be in our faces.

00:13:14 Andrew Keen: It's a very chilling prognosis, Dr. Pearl. But will people know it? Will it be obvious? Will it be self-evident? It's not going to be like a pandemic where people die on the streets or in their homes. It's going to be a drip-drip kind of process, isn't it?

00:13:33 Robert Pearl: It will only happen when something bad occurs. Doctors are pretty good at resuscitating people at times, but that's just going to create a vicious cycle. As people can't get the preventive care they need, they'll have more chronic disease. If that chronic disease is poorly controlled, over the course of a decade or two, we're going to see more heart attacks, strokes, and kidney failures. At some point, people are going to have to depend upon the largest providers to treat them without the dollars. They can't even mortgage their homes, because they won't have enough value in those places — and that's what people are doing now. We won't see it because it won't be in our day-to-day lives — assuming, like you and me, we have health care coverage and can afford the rising costs. But for the families that are just getting by, which is the majority of our nation — living paycheck to paycheck — it's going to break at some point where they simply say, we're not going to do it. In fact, we're already seeing survey data showing that 25% of Americans made a major sacrifice last year in order to pay their health care bills. And when that starts to reach 50%, maybe we'll see it at the polling places — different people elected. But I'm even doubtful of that, because that's why I say it's so ultra stable: when you have literally ten lobbyists for every single elected person, and your agenda is backed up by campaign contributions, we're not going to see it coming out of Congress. And that is where, if we're going to avoid some type of major disruption, it's going to have to happen at the legislative level. There's been a lot of talk — look at GLP-1 drugs. There's all this talk about how the prices are coming down. And yes, they're coming down. But how many people earning $60,000 a year can spend $400 a month? Less than the $1,200 it was in the past, but that's $5,000 post-tax. You just can't afford that if you've got a couple of kids and a mortgage or rent to pay and you're trying to buy food and clothes for your children. We're getting to a financial breaking point.

00:16:08 Andrew Keen: But aren't GLP-1s a luxury product, or are they essential, Robert? If people can't afford these weight-loss drugs — and my understanding is the jury is still out on their long-term impact, whether you have to stay on them for life and what the other health consequences are — does the fact that people can't afford those drugs mean the system is going to collapse?

00:16:35 Robert Pearl: No. What's going to happen is what's already occurring: probably the number one cause of the growing disease burden in the United States today is the proliferation of chronic disease. And the main driver of that — whether we're talking about diabetes or, as an article recently noted, cancers — is obesity. We have an obesity epidemic that is significantly harming the nation. And we have these GLP-1 drugs. According to Yale, the cost of creating a GLP-1 drug is $5 a month. They're being sold now at a discounted price of $400. If my math is right, that's 80 times more than it costs to make. How do we justify that? If we could cut obesity in half, we'd see health care costs plummeting. We have an imbalance between the demand for health care and the cost of providing it. Across the United States today, the cost of the average American is $15,000 a year. The average employer is spending over $25,000 to provide coverage to a family of four. These are astronomical numbers, particularly when you consider that Switzerland is at $10,000 and Germany is at $9,000, and everyone else is half of us. And that's why I'm so excited about the generative AI tools. One thing I've been writing about lately that I'm sure your listeners and viewers will find interesting is vibe coding.

00:18:23 Andrew Keen: I know. I want to come to vibe coding, but before we get there — you're a mine of essential information, Robert. You wrote an interesting Forbes piece about three health care problems that are "too big to solve." You may have already touched on them. What are these threats that are, to quote the headline of your recent Forbes piece, too big to solve?

00:18:54 Robert Pearl: The point of that article was that problems which could be addressed in simple ways, if caught early, become so big over time that the solutions no longer work. A great example in life: if you educate children when they're young, you can make a relatively small investment. Wait until they're much older to start their education and you're so far behind it won't work. Take a disease — find cancer early and you can cure it. Not always, but a lot of the time. Wait until it spreads and you can't do anything like that. So take the cost of health care. Right now it's $5.6 trillion. It's projected to go to over $7 trillion. Right now, we probably could solve it if we could control chronic disease. But if you wait and give people ten more years of becoming ever sicker, of having their blood vessels become ever more damaged, then even if you decide at that point to start treating chronic disease, there are so many people with advanced problems that it won't be adequate. And so you end up having to ration because you don't have the resources. If you wait an entire decade to teach doctors how to use generative AI tools, you lose the opportunity to have the next generation leap us ahead rather than fall farther behind. But if you start using generative AI today and create tools to allow patients to improve their health, we will change the percentage of the population with advanced chronic disease. We'll change the health of the nation in a positive way. If we wait, the tools are still good — it's just that it becomes the opposite of too big to fail. The problem becomes too big to succeed, because the size of the hill you have to get over is just too great.

00:21:22 Andrew Keen: Robert, you wrote another interesting piece for Forbes — and you've been writing these sorts of pieces for a while, and you've written books on them as well — about how gen AI will replace much of what clinicians do. You've noted that you want clinicians to lead the way, but you also brought up the example of Kodak. As I said, I wrote a book about Kodak and the shift from Kodak to Instagram. Just as Kodak went bust and 20,000 to 30,000 workers lost their jobs over a number of years, Instagram went public with 15 people working for them and got sold to Facebook for about a billion dollars. So the real question is: to what extent will gen AI replace clinicians? How much of a fear is that? Where is the value of a clinician or a doctor in our age of AI? How are we going to guarantee that they lead the way?

00:22:35 Robert Pearl: We can't guarantee anything to anyone. This is the evolution that has to happen. The only way you can guarantee something is with enough money, and we'll never have enough money to do it if we don't plan for it. But it's not just medicine — there's going to be a lot of changes everywhere. You have to put it in the context of the value that's going to be created. Generative AI tools, effectively utilized, will create massive value. As you know, Sam Altman wrote about that this week — that we need to create a fund to support people, and we have to recognize that we may end up with a four-day work week rather than a five-day work week. And for a lot of people, that would be a very positive outcome.

00:23:22 Andrew Keen: Although not everyone trusts Sam Altman. There was a big piece in the New Yorker this week about how profoundly untrustworthy he is within OpenAI. So I'm not sure we want to use Sam Altman as the authority on trust.

00:23:39 Robert Pearl: No, I agree with that. But the point I was making is we have to understand that if these tools allow us to be so much more productive, there will be a lot more money in society, and we have to decide how it should be distributed. That is a political decision. To simply have one individual worth a trillion dollars in a nation with a total GDP of $17 trillion — it just stops making sense. Whether we have the political will to address that is probably a question for another guest. I'm pretty pessimistic. But let me go back to your question, which is that right now we need a lot more continuity of medical care. You see a physician, you have elevated blood pressure, you're started on a medication, you're seen again in four months. What happens in the interim? We don't know. With chronic heart failure, which sometimes becomes acute, you go to the ER and get admitted to the ICU. It takes three days typically to develop — retaining fluid, creating the problems. If we had someone checking in on you every day at home, we could tell what was happening and intervene. It's too expensive. We don't have enough people. A generative AI tool can accomplish all of that. You see a mental health worker, you have depression, and they're worried — not enough to commit you or force you into ongoing observation, but they want a tool to evaluate you. The opportunities are so massive. In the end, medicine is an intimate industry. You have a doctor-patient relationship. You tell doctors things you wouldn't tell anyone else. You take your clothes off. There's something about the human touch that has been there for five millennia — you can trace it all the way back to India five thousand years ago. If I'm at the end of my life, I want to talk to a human being. If I have a cancer that's going to be problematic, I want to talk to a human being. But I don't need to talk to a human being to get the care that is algorithmic. I think the first place generative AI will have a systemic impact will be in primary care. So much of primary care is chronic disease management and it's algorithmic. If 30% of that can be done by technology, primary care can spend more time with the people who are very sick — the things that are uniquely human. When there are too many variables, the generative AI tool isn't reliable. Primary care physicians can also start to do some of the things that only specialists do today, because they now have more technological assistance and the time to accomplish it. And specialists can shift to doing more procedures, freed up from the office. You train for all those years as a cardiologist to pass catheters, not to rip blood vessels; as a surgeon to perform procedures, remove a cancer. This is the evolution. Will some people get lost along the way? Probably. That's just the nature of change. You can't point to a single event from the printing press to the steam engine where that didn't happen. But society comes out further ahead. Did we get universities because we had printed books rather than handwritten manuscripts? Did we get an infinite number of applications of steam engine technology — cars, planes, construction? This is just the process that happens. Right now, medicine is a very stable industry in the sense that it has an overall deficiency of people. What will happen in fifty years? I have no idea.

00:28:08 Andrew Keen: Well, maybe your next book should be called Ultra Stable. Robert, let's be a bit more specific. You mentioned the price of GLP-1 drugs. My understanding was that Trump Rx was supposed to address that — that at least you could go there and get cut-price GLP-style drugs. Is that wrong?

00:28:38 Robert Pearl: It's not wrong, but it's wrong for most people because you have to pay cash, and most people have insurance. You can't use your insurance to get drugs at a discounted price there — or for that matter on many other similar websites. The site itself has some drugs that are less expensive and some that are more. Mark Cuban has his Cost Plus operation. You have GoodRx. There are a lot of ways to buy less expensively. But even if you could, it's still $400 a month. I have done the analysis, the mathematics. If you want to make a GLP-1 drug cost-effective — meaning the improvements in health lead to less medical expense, so it's cost-neutral — the price has to be under $200. So we're still massively overpaying for drugs that the human body makes naturally and therefore can't be patented. Only the processes of bringing them to market can be, which is why we're still paying so much more than the cost to create them. And you were right earlier, by the way — we don't know the long-term impact because they're new. And you're absolutely right that when you go off them, you regain your weight. So this is a process that needs to be done for life. Would I like everyone to lose weight by eating differently, exercising, and making other lifestyle changes? Absolutely. But if they can't, it is far better to get the weight off and prevent the diabetes, the heart disease, and the kidney failure. I think you're going to see almost no impact for ninety percent of the population from Trump Rx —

00:30:41 Andrew Keen: More disappointment. You're in the business, Robert, of making Americans healthy again. Another man — RFK Jr. — is also in that business. Last time we talked, you suggested he hadn't been particularly successful, though you were open to the opportunity he represented. There's a recent piece in Politico about how he has launched a midterm travel push to shore up support for Make America Healthy Again. Has he continued to flail? Has he done anything of any positive benefit for the health of Americans?

00:31:22 Robert Pearl: Not that I can find. I think he's done a lot more destructive things, particularly around vaccines. It's been blunted by some court actions and decisions, but he's coming back at it. I always try to give people the benefit of the doubt. But it's fool me once, shame on you — fool me twice, shame on me. I think he has shown his true colors, and I'm very concerned. But I'll tell you what I'm even more concerned about coming out of him and the whole of HHS, which is that we are dramatically undermining our entire investment in science. We're losing scientists. We're losing the students who want to become scientists. We're losing the innovation. If you want to talk about a battle of a nation against China, this is it. And right now, we are losing. Everything I'm saying is not just my opinion — it's what I'm reading out there. You may have seen the recent White House budget with an additional 15% reduction. People don't realize that most drug development doesn't happen in drug companies. They only take something that's 90% of the way there and carry it across the finish line for a significant reimbursement and profit. Most of it begins in the NIH — or in academic medical centers, where cutting the NIH budget cuts the university budget. When you start going after those institutions, it's almost like not educating children for a decade. You end up with such problems when they become adults, when they have children, in the next generation. Listening to myself say these things, I feel like a doomsayer.

00:33:23 Andrew Keen: Some people would suggest we're already seeing those problems. What's the first political crisis we're going to see? We've noted that RFK isn't doing a great job. Is it going to be a crisis over Medicare? An attempt to resurrect or undermine Obamacare? What is happening politically? It's such a complicated and, frankly, boring subject, Robert — as you know, when most people hear debates on health care, they switch off.

00:33:53 Robert Pearl: I would say we're not going to see a sudden collapse because the system itself has strength, particularly right now with the courts holding up their end of the constitutional expectation of three equal branches. Where are we going to see it? The midterms. I don't know what's going to happen there. I wrote about this at the turn of the year — this was going to be a driver across the first ten months of 2026. If the outcome is as most of the pundits are predicting, that's when we're going to start to see massive crashes, collapses, and battles. I have some fears for our nation and how we're going to respond in that environment. Health care will be only one of the major parts — it's 20% of the US economy that's impacted. I think we're going to see the end of a lot of what I'll call the errors that RFK Jr. is doing. I think we're going to see a Congress that becomes much more activist. But whether it can get its act together and finally stand up and make good decisions for the country, your guess is as good as mine.

00:35:29 Andrew Keen: Is there a solution out there, Robert? If you got a call from a leading Democrat — from Gavin Newsom, or even from JD Vance, or Marco Rubio — and they said, "Robert, you've been around, you've written books, you've run a major insurance system — what's the answer?" What are politicians talking about coherently? Is it just to leave Obamacare alone? Return to Obamacare? Overhaul Obamacare?

00:36:00 Robert Pearl: Obamacare — basically, the president at the time looked at three things: quality, cost, and access, by which I really mean coverage. He realized he couldn't do all three. Obamacare was about coverage, and it dropped the uninsured from around 16 million to around 8 million — basically in half. I don't think that's either the answer or the problem. To me, it's straightforward. The first thing we have to do is change the way we reimburse care. Right now, the incentives in a fee-for-service, pay-for-volume system are to do more and more and more, raise prices higher and higher and higher. Instead, we need a pay-for-value, capitated system — and I would say not through insurance companies, because that's the problem right now. Yes, it's capitated insurance companies, but they pay doctors and hospitals on a fee-for-service basis, and you're right back where you started. We need groups of doctors and hospitals to come together in sufficient size and be paid on a capitated basis. Because as soon as you do that, you align the incentives. In a capitated world, the way you make the most money is by avoiding heart attacks, strokes, kidney failures, and cancers. To do that, you minimize chronic disease. To do that, you actually change people's dietary habits, their exercise patterns, their lifestyle medicine. Everything aligns. I was the CEO for eighteen years, and what I learned in that role is that any CEO who thinks they can send an email and tell people what to do is deceiving themselves. All you can do is create the structure in which people understand the right thing to do and have incentives to accomplish it. That's the first thing I would say. The second thing is you can't make that shift by simply deciding on January 1 to change how you pay people. You need a transition plan. I would continue to pay people for five years at the same rate they would otherwise have received, so they're not being hurt. But I would also let them know that I'd share 50% of the savings with them at the end of five years if they were successful. Doctors are very smart and very intuitive — they would come up with the ways to make these changes where currently little happens. The third thing I'd recognize is that you need to make capital investments in order to make operational changes occur. Particularly around generative AI, I'd be investing a lot of money upfront: ways to have patients use it better, doctors use it better, and to support clinicians in helping patients accomplish it. Those are the three things I would focus on, because all of them drive a process where people come to work every day and ask themselves the right question: How do I keep people healthy? How do I avoid a complication? How do I provide access sooner, because if I don't provide it now I'll pay a bigger price down the line? That would be the start of the framework.

00:39:19 Andrew Keen: You've been very patient with me, Robert. You've answered all my political questions. So finally I'm going to let you talk about vibe coding. You had an interesting piece about how you believe vibe coding will reshape medical practice. Not everyone in our audience will know what vibe coding is — we have a lot of tech viewers and listeners, but not all of them. So you might define it and then explain why vibe coding might be the answer.

00:39:48 Robert Pearl: I think vibe coding will be — I'll call it a powerful tool, though [unclear] the answer. First, what is vibe coding? It barely existed a year ago. If it did, it at least wasn't being publicized. So it's only one year old — not even close to three years old. What vibe coding does is allow individuals, even ones with no background in computers, to create software applications. So if you want a computer to do something for you — take something very simple: I want a spreadsheet. Well, you'd have to learn how to use Excel. No. With vibe coding and applications like Codex on ChatGPT or Claude Code, Anthropic's product, you can tell it in plain English what you want. Now why is that so vital? Because you've gone from a small number of companies creating products to every clinician being able to accomplish this — to personalize the care they provide to their patients. This is where vibe coding will radically change things, assuming people want to do it. If you read everything in the papers or see the talk shows, you can see that the computer industry has changed — they're able to accomplish things with a third or a tenth of the people, or achieve a lot more with fewer people. Dramatic change — that's vibe coding. One person in a weekend can accomplish what a team would have taken months to do. Any clinician who didn't want to create the applications themselves could hire someone, and it's relatively inexpensive because that person would use vibe coding to build the application. So what would you do? Let's take the example I mentioned earlier — heart failure. Someone with chronic disease has had a heart attack in the past. Their heart has lost a lot of muscle because the heart attack kills muscle, and they've been given medication which keeps them in relative homeostasis. Now they have a problem — they forget to take their medication, they get a viral infection, and they start to decompensate by retaining fluid. The fluid backs up into their body, affecting their lungs and their breathing, and that's why they end up in the ER and then being admitted to the ICU. But as I said, that's a three-day process. So I say to you: you're a clinician, Andrew — what would you do if you could go to the patient's home every day? And you say to me: well, Robbie, that's pretty straightforward. I'd put them on the scale to see if they gained two or three pounds, because that's a sign of fluid retention. I'd listen to their lungs to see if I could hear fluid. I'd look at their ankles, because gravity pulls fluid down towards the floor, and if they start swelling that's a good sign their heart is failing. I'd check their blood oxygen. I'd probably have them climb some stairs to see if they could still exercise, and I'd have them lie flat because when the fluid backs up it floods the lungs more as fluid mobilizes from the legs. Now, I say to you, Andrew — that's exactly what you would do, and you can create an application that does exactly that. With Bluetooth it can connect to an electronic scale. With Bluetooth it could have an electronic stethoscope tell you how the breathing sounds. With Bluetooth it's able to connect to a smartwatch or wearable device to check blood oxygen. And it now has something my book three years ago didn't have — multimodal capability. It can look at video of what happens when you lie down and when you sit up. With all that information, it can tell you as soon as that patient has a problem. When I was CEO of Kaiser Permanente, we knew all this. We were sending nurses out to people's homes. It was making a big difference, but the cost of doing that was as much as the cost savings. A vibe coding application could be built in a weekend — though it has to be tested and revised; it's not as simple as I'm saying today, one year in. This is going to be the personalization. And why I think vibe coding is so powerful is that up to now, if we're going to make the transition from fee-for-service into capitation, you need a tool that allows you to take that risk because you have a high probability of achieving a better result. Up to now, we haven't had it. Tell the patients to eat better. Tell the patients to lose weight. It doesn't work. GLP-1 drugs are too expensive. Everything we had, we just couldn't translate theory into practice. Now a generative AI tool can do that, and it can do that for a cost of $20 a month.

00:44:59 Andrew Keen: Well, on that note — vibe coding might be the answer, might not be. You've convinced me, Robert. You've shamed me into going to the gym. As always, it's a real honor to talk to you. You don't mince your words. I think your notion of ultra stability is a very interesting observation. I hope you'll write more about it, and I hope you'll come on the show in the next two or three months to tell us if AI has indeed begun to reform the health care industry. As always, Robert Pearl, MD — real honor. Thank you so much.

00:45:33 Robert Pearl: Always a pleasure to be with you. You're a great host. Anyone who wants more information, go to my website, robertpearlmd.com, and you'll find a lot of information on all these topics. I look forward to the next time, Andrew. Thank you so much.