Transcript
Intro
Stephan Shipe: Welcome back to the Scholar Wealth Podcast. This week we open with a listener whose brokerage is pitching a fully paid lending program. He has a concentrated position and is looking at the offer as a way to earn a little more from the shares he already owns. He wants to know whether it’s a good idea, a good deal on its own terms, and whether it does anything useful for the concentration problem itself. Then we hear from a listener who ran a Monte Carlo with Claude to test whether he can retire at the end of next year. He’s calling it a green light. His wife is not convinced, though, and he wants to know what else should go into that decision. And in From the Field, we are joined by Dr. Nate Ruch from the Princeton Longevity Center for a conversation about proactive preventative medicine and where the highest-return investments in your own health sit. Let’s start off with question number one.
Question 1 – Vanguard’s Fully Paid Lending Program: Free Money on a Concentrated Position, or a Distraction From the Real Problem?
Stephan Shipe: Vanguard is pitching me their fully paid lending program. I have a concentrated position in a name that gets shorted a lot, and it sounds like I can earn extra income by letting them lend out the shares. It sounds like free money on shares I already own. And does it do anything for the concentration problem itself?
So, really cool question. The concept of lending and short selling — when you short anything, when you short a stock, you’re betting on the stock going down. So you borrow those shares, immediately sell them, and then you have to buy them back later. So the hope is you borrow some shares at a hundred bucks, sell them to somebody for a hundred, wait, stock goes down to eighty, and then you buy back the shares at eighty and you made twenty bucks. So it’s the reverse of going long a stock, when you’re buying low and selling high. So in this case, what you’d be doing is providing those shares. So if somebody wants to short it, they can short the shares, and they would be your shares that they’re shorting, and you get paid to do that. Typically not you — broker gets paid. They split some of that with you.
So there are a few things that come into play, though, that are important caveats. One is that the amount of money that you can earn for lending out securities varies widely based on the type of security that you’re lending out. If you were to try to lend out like a VOO or a VTI or one of the major index funds, there’s not a lot of short interest in that. So if there’s not a lot of short interest, or a lot of people who are demanding shares to short, that’s a lot harder to deal with. You’re talking about a really low amount — on 100 grand, you’re probably looking at $50 a year to lend those shares out.
Where things start to get interesting is if there is a demand for an individual stock to be shorted. In your case, typically concentrated positions, single-name stocks, are the ones that people want to short. They’re targeted. People can look at that and say, I don’t want to bet against the whole market, I want to bet against that stock right there. So the fact that you’re in a single-name stock gives some credibility to you being able to earn a little bit more on this. Let’s call it a few hundred dollars for this stock, depending on what it is, on a hundred thousand dollar position on average.
My bigger concern, though — and this gets very philosophical about the whole idea — for you to earn money on this, you need to have an assumption that there are a lot of people out there who expect this stock to drop because they want to short it. So you have to look back and say, I understand I’m going to earn some money, but I’m so glad that I own the stock that a lot of people are betting against, so that way I can allow them to bet against my stock and watch it drop and then earn some money on it. That should set off some red flags on what you’re dealing with.
And it doesn’t solve the concentration issue. Some people argue that, well, you’re getting income on it, so that helps deal with some of the concentration because you’re offsetting some of the downside that you’d have. That is very true, in the smallest extent, though. Sure, if you make $100 loaning out your shares, that offsets some of the downside you could have. But the downside is substantially larger than any money that you’re going to get in income from loaning out the shares.
Not to mention, if this is a company that pays you dividends, usually dividends are qualified and you’ll end up having that taxed as long-term capital gains. But in your case, if you loan out these shares, the people who are borrowing them from you have to pay you the dividend. But they pay you that in a payment that you get taxed as ordinary income on. So you give up dividends, which are taxed favorably, for payments, which are now taxed as ordinary income. So depending on where you’re at from a tax bracket, that could be problematic. You give up voting rights as well. That’s typically not too much of an issue, because I doubt you vote your shares. Very few people do on individuals.
So we’re not helping the concentration risk. I think if really the concentration risk is what’s driving this, then we need to just exit out of the concentration risk and think about that as a stronger argument, as opposed to, I’m going to lend out my shares to earn a little bit more. There are other paths if you’re truly wanting to generate income. So let’s say you look at it and say, I’m going to keep the shares, I’m not selling them, regardless of what you say. Then I would say, if you’re going to keep them, instead of just lending them out to be shorted where you’re going to earn a little bit, your best bet would probably be to go and sell calls. And do a covered call strategy on it where you own the stock, you sell calls above the market — that would be a substantial enough income for you to argue that you’re offsetting at least some of the downside, not a ton of it. And then if you were going to try to mute that position completely in your portfolio, you’d put a collar around it. You’d take the income you’re getting from selling those calls and you’d buy some puts below the market. And that would lock in the range of movement that this stock would have.
But now we’re adding a lot of complexity. So if you’re wanting to go that direction — it sounds pretty easy to jump into the fully paid lending program. I don’t think it’s worth the kind of pennies that you’re going to get on that. And I think it avoids the bigger issue of having this concentrated position in the portfolio.
Question 2 – A Monte Carlo Gives 92.67% Success — But Is That Actually a Green Light to Retire?
Stephan Shipe: I’m 61, my wife is 58, and we’re trying to decide whether I can retire at the end of next year. I ran a Monte Carlo with Claude using our portfolio numbers and spending target. It ran 70,000 scenarios and gave me a 92.67% success rate. That feels like a green light, but my wife is saying it’s not enough to base our decision on. Any other considerations?
So anytime there’s a Monte Carlo — this is an area that’s extremely important. I think the AI usage around finances is really putting this into proportions that are getting a little dangerous on how people are interpreting Monte Carlos. But Monte Carlos, by definition, are just a fancy way to say it’s a what-if analysis. But it sounds really good if you say, I ran a Monte Carlo. Really all you did was run a thousand different scenarios, or 70,000 different scenarios, allowing different inputs to fluctuate in a range that you tell them they’re able to fluctuate in.
So in other words, if you do a really simple Monte Carlo that says, what is an account going to look like in 20 years or 30 years, ignoring all the other retirement-type math and calculations — all that would say is, I want to take a portfolio, let’s say it’s $10 million today, and I want it to go and be random every year, randomly select a return from a distribution that would be set by you as what would be normal on the market. So you would put an average, you’d put a standard deviation, we’d assume a normally distributed distribution, and we’d say, randomly pick every year from that distribution. So one year it’s going to be 8%, another year it’s going to be 15%, another year it’s going to be negative 10%, another year it’s going to be positive 20%. And every year it does that, and your portfolio is allowed to go up and down with that change. And then it goes all the way to the end and says, were you able to make it through this portfolio value without running out of money? And then it goes back to the beginning and says, now run it all again. And continue to pick from that distribution.
So the nice thing about that is it allows numbers to be random. So it allows something that we are uncertain about, something that is stochastic, like a market return, to do its thing, to be random within a certain range. The problem with that is that it assumes that you’re going through this world where last year’s market returns don’t affect this year’s market returns. And there’s some truth to that. You have kind of this random assumption of returns not being affected by the previous year. There’s also some knowledge of that with market values being higher at certain points and PE ratios. So we have to be careful — that’s not included in Monte Carlo.
But the bigger issue that I see in Monte Carlos in general — and this is coming from someone who uses Monte Carlos regularly — is they’re very easy to game, and they’re very easy to mess up. And what I mean by that is — I used to teach quantitative finance and Monte Carlo simulations all the time. And one of the issues was, I can make any portfolio look really good and any portfolio look really bad on a Monte Carlo, because as long as you know what assumptions you can change, I can make things look good. I can go and have a tighter distribution. I can change the distribution. I can allow for things that don’t have volatility that’s accurate. We see this with individual stocks all the time. It’s very difficult, if not impossible, to put individual stocks into a Monte Carlo, because by definition, we don’t really know what the idiosyncratic risk of that stock is.
So we have all of these issues. When you go into a retirement Monte Carlo, you’re allowing multiple things to change. So you’re allowing not only the return on a portfolio to change, but you’re allowing spending to change. You’re allowing timelines to change. You’re allowing things like Social Security to change. You’re allowing portfolio withdrawals to be variable not only year to year, but from decade to decade, where you have different goals, liquidity events that are coming in. There are a lot of different inputs. Every time you add an input into a Monte Carlo, each one of those inputs has to have its own assumption of distribution. And when you start to mess with those, you can make anything shift one way or another.
And the issue that I’ve always had with Monte Carlos is it throws out these percentages, just like you’re seeing — the 92.67%. And it’ll say you have an 85% probability of success or a 90% probability of success. And I was looking to say, well, what does that actually mean? And what that’s telling you is that 85% of the time you don’t run out of money. Well, no one runs out of money, right, in that scenario, because you wouldn’t just watch your portfolio go to zero. You would change something. You would end up having a decrease in spending. You would change your travel plans. You would be dynamic. So Monte Carlos assume that you’re sitting in a point in time and you’re not going to be dynamic going forward. And all the assumptions that you have are not related to each other.
And this is a big factor of where we have issues. Take a Monte Carlo where you have inflation that’s increasing. Say, well, that’s okay, I want inflation to move over time. In a year where inflation’s increasing, we’d expect the market probably to be up a little bit, because equity markets would go up during that time. But those are separate. We have your spending increase early in retirement, and we could possibly have a market dropping 30%. Well, that’s not going to be true. Market’s dropping 30%, you’re probably not going to spend as much as you would. So it doesn’t take into account the correlation of these.
The bigger issue is this illusion of precision that starts to show up with Monte Carlos. That somebody says, I want a Monte Carlo, this is what I want to do, and it pumps out a number like you’re saying of 92.67%, so you’re good to go. When really the issue with a Monte Carlo is not the output, it’s the inputs. It’s understanding what assumptions are being made. And in the past, that’s always been relatively transparent, because you could look at all the individual assumptions. You could build a Monte Carlo model in Excel, or Monte Carlo software, or some planning software, and you’d have all of this information to go pull every dial and pull every lever that exists, so that way you can see truly what the assumptions being made are, and which one of those breaks the model.
Whenever you’re just saying, run a Monte Carlo, here are some general assumptions, you miss out on what are the things that you can’t input into the model, or you may not have thought to input in the model — things like tax law changes, hedges against changes in health care costs, long-term care, helping out family members, longevity risk. All those things you can’t put into a model, because there’s no assumption of a distribution around what the probability of a tax law change is, or what the probability of you needing long-term care is in a model. So you end up in a lot of situations where there’s a Monte Carlo that, while it has all these assumptions and seems like you’re covering everything, you’re really only covering everything that has a distribution that you can follow. Everything else ends up being a judgment call around it that needs a lot more analysis to see where the plan would break and what would happen if one of these events happened.
So the stress tests on a Monte Carlo end up being a lot more important than the Monte Carlo itself. The Monte Carlo, in other words, is not an end result. It should be the starting point for you to start looking at different advice around how your finances in your Monte Carlo and the model would react to different types of shocks to the plan. So on this one, I’ve got to side with your wife on it. I think that’s a good start to where you need to be, but you need a lot more steps in here. And I wouldn’t let the 70,000 scenarios change anything either. Usually Monte Carlos around 1,000, 2,000 scenarios are sufficient for enough degrees of freedom in a model. Once you start adding above that, it just again makes it sound more precise when really you’re not getting any added benefit of all of those different scenarios.
From the Field – Proactive Prevention and the Highest-Return Investments in Your Health
In our From the Field segment, we’re joined by Dr. Nate Ruch, a physician at the Princeton Longevity Center in Princeton. Originally trained in emergency medicine at the Cleveland Clinic, Nate now works in executive health and preventative medicine, where he helps patients build bespoke, evidence-driven plans for early detection and long-term health. He joins us to talk about proactive prevention, where the highest-return investments in your health sit, and how new diagnostics and AI are reshaping the field.
Stephan Shipe: Nate, welcome to the Scholar Wealth Podcast. To start off today, why don’t you give us a little bit of your background, what got you into this work?
Nate Ruch: Sure. So I’m a physician who originally trained at the Cleveland Clinic in emergency medicine. And almost from my first hours working as an emergency physician, I felt like I was seeing people 10 or 15 years too late, and that there had been a missed opportunity for some proactive prevention. And that led me over time to where I am today, which is working in executive health and preventative medicine at the Princeton Longevity Center. Talking to you from my office here in Princeton, New Jersey.
Stephan Shipe: So has there always been such an interest in longevity in general? I know you see this over time, but it seems like recently this has been the hot topic.
Nate Ruch: Yeah, I think that Princeton Longevity’s been around for a little over twenty years and was definitely a pioneer in the space. Although I think thinking about longevity isn’t necessarily new. Almost everyone who reaches a certain point in life and resources realizes that all the resources in the world really aren’t worth anything if you don’t have health, and that ultimately it’s time and health that are non-renewable resources that are hard to recharge as you go through life. It’s definitely true that in terms of social media and YouTube and kind of popular discussion, longevity’s definitely had a moment over the last three years or so. So it is becoming more a topic of discussion for sure. But it’s been in the background for a long time. And, like a lot of things, some of the future is already here, it’s just not evenly distributed yet. And that’s changing, as some of the technologies used for evidence-backed proactive prevention become easier and more accessible to folks.
Stephan Shipe: Do you think the research is catching up to that, or having a spike as well? Is the increase in interest due to increased breakthroughs, or is it just an increase of interest that hopefully will lead to more developments in research?
Nate Ruch: That’s a great question. I’m not sure what the answer is. I think that there are definitely places where there’s irrational enthusiasm for things that the evidence just doesn’t support, or isn’t really mature enough to give a full recommendation to. And then there are places where that’s balanced by the fact that medicine, particularly at its highest levels at the most prestigious institutions, is very slow to change. Like a lot of fields in science, it kind of changes one retirement or funeral at a time. And guidelines are often behind where the evidence is. Now, some of social media and some of folks’ algorithmic feeds are way ahead of the evidence. And there are definitely people that talk with inappropriate levels of conviction about trendy topics. And there’s also, like a lot of things you see, there’s profit motive behind some of that as well. So I think here at Princeton Longevity, that’s one of the services we offer — kind of separating signal from noise and determining, what are the high-yield things you should focus on? And there isn’t really a one-size-fits-all answer to that question.
Where Supplements Fall Short — and What Actually Moves the Needle
Stephan Shipe: And that’s exactly where — if you could juxtapose those two, what are the ones that people come into you and you just shut down fast? Like, I’m glad that was on social media, but that is just not true, or it’s not there yet and we’ve got a long way to go.
Nate Ruch: Yeah, I think supplements in general is a category where sometimes people don’t fully consider the magnitude of effect. If you look at changes in level of cardiovascular fitness and you compare that to the impact of any supplement — if you look at people fifty to sixty years old and you compare the least fit twenty percent to the top fittest two and a half percent in that cohort of people, there’s a five hundred percent difference in ten-year mortality. It’s just massive. It’s almost impossible to exaggerate the effect that improving your cardiorespiratory fitness and maintaining a high level of fitness has on overall health. And there are people that spend a lot of time learning about, sourcing, and taking supplements that ultimately are a rounding error in their health, when there’s something they have ready access to that’s a couple of orders of magnitude more effective. And it is true that there are some supplements that can make a meaningful difference, but by and large, if people took the same resources, time, and energy that they put toward selecting and taking supplements toward exercise, sleep, and nutrition, they’d be better off.
Stephan Shipe: So I like the high-ROI options. You would throw exercise in general, sleep, just general nutrition. But a lot of those are not necessarily financially constrained, right? In the sense that anyone can go for a run, right? So you have those opportunities. Are there other things that you would add to that list that have the higher ROI that money can buy?
Nate Ruch: Yeah, sure. I think people that have more resources than average have a couple of problems that they face. Many times they get too much care that leads to downstream consequences, or they get involved with charismatic folks that claim to have a solution that no one else has, and it’s very expensive, so you need to pay me for this special thing that nobody else has. I think both those things are dynamics to be aware of.
If you said to me, well, where’s a good place to deploy my resources if I want to have better-than-average long-term health? I think a good place to focus — there are some places in health where your subjective sense of how great you feel on a given day is totally disconnected from what’s happening under the hood. And cardiovascular disease, heart disease in particular, would be an example of that. And there are some newer technologies using techniques like CT coronary angiography, where we can do a really thoughtful evaluation of the degree of atherosclerosis you have, the degree of blockage you have in your heart vessels. And unfortunately, a question that medical students get asked is, what’s the most common presenting symptom of a heart attack? And it’s kind of a trick question. Sometimes you’ll hear people say shortness of breath or chest pain. Well, the most common presenting symptom of a heart attack is sudden death, right? So many people — about a third of people that have a heart attack — they don’t get any warning symptoms. And there just really isn’t a good coupling or correlation between how you feel on a given day, and how well you’re doing in your workouts, and your overall level of vitality, and what’s progressing under the hood.
And there are other conditions like that. Bone health is another example. For reasons that are mysterious to me, DEXA scanning is not used in a very widespread manner. And there are lots of people walking around with big opportunity on bone mineral that they don’t really know about. A third place where you could deploy some resources would be around cancer screening. National cancer screening guidelines are designed at the population level with resource constraints. They use calculations like dollars per cost-adjusted year of life. If you’re a more well-resourced person, you could look through your genes, your family history, and your environmental exposures, and you could craft a bespoke cancer screening program that is more effective than what the national guidelines would be.
Stephan Shipe: Where does someone go to find that level of screening? Because I know you mentioned some like executive medicine programs and everything else. Is that where they should go?
Nate Ruch: Yeah, I would say it’s a tricky thing. And unfortunately, there’s no gold standard training in longevity. And there are a lot of different avenues you can take to get evidence-backed, bespoke healthcare planning that’s more designed toward chronic risk management and mitigation of onset of chronic disease, rather than the kind of reactive sick-care approach that occurs in a lot of traditional medicine. The Princeton Longevity Center where I work is a great example of a place that offers that approach, where we take a comprehensive look at the whole person and develop a bespoke plan to help folks live the longest, healthiest life possible. We’re definitely not the only solution. And it can be a tricky thing, because many times well-resourced people will buy very, very expensive programs that, unfortunately, are driven by marketing and the charisma of the person presenting, rather than by evidence.
Stephan Shipe: Where does that line get drawn, then, that you get too much care? Because you mentioned that was one of the issues that you run into.
Nate Ruch: Yeah. I will sometimes encounter — I see a lot of very well-resourced patients in my practice, and I’ll sometimes encounter people that have had really impressive levels of diagnostic testing, just kind of way beyond what’s useful or appropriate. And whenever you get any medical test, you always want to go into it with the idea of, what am I going to do differently based on the result? And there are some folks that just end up doing testing for testing’s sake, without a clear plan of what’s going to be different driven by the results. I think the answer’s different for each person. I think it’s challenging to come up with a one-size-fits-all. But I think in general, finding somebody who takes an evidence-informed approach and gets to know you and doesn’t have a lot of time pressure when you talk to them, and can really deeply understand your history, your family history, maybe do a little bit of genetic testing, and meet and see where you stand and come up with a plan.
Long-Term Consistency vs. Short-Term Intensity — and What Actually Drives Longevity
Stephan Shipe: How long does that take? And I know that’s an odd question — we’re talking about longevity, right, of how long does longevity take? But when you hear about things like, look where you live, or your environmental factors, or the blue zones across the world — can I go and pick up and move my family to a blue zone area, and that’s going to increase life expectancy?
Nate Ruch: I think the longer you’re healthy, the better your outcomes are going to be. And I think a good guiding principle is that long-term consistency is far, far more important than short-term intensity. And it’s way better health-wise to be eighty percent good a hundred percent of the time than it is to be a hundred percent good fifteen percent of the time. And many people fall into that latter category, where they have bursts of time where they’re — I’m going to do everything perfectly starting at 5:45 tomorrow morning. And then they revert back to their kind of mean behavior. And it’s much better to drive your health planning and your behavior through the lens of long-term consistency.
And blue zones are super interesting. I think how much is demographics, and the advent of more systematic record keeping, has had a big influence on some supposed blue zones, right? Once birth certificate use became widespread, the percentage of superagers, the percentage of those folks in some of the blue zones, has gone down pretty dramatically. So it’s kind of an interesting thing. But to answer your question more directly, I think the longer that you have a healthy set of behaviors, the higher the probability you’re going to live the longest, healthiest life possible. And I think another really good framework along those lines is that you’re not going to find some hack or perfect supplement, right? The place to focus your energy is on early detection of preventable chronic disease, fitness, sleep, stress management, and kind of mindset and nutrition. Those five buckets are the place to focus your efforts.
Stephan Shipe: And from your research and being in this area, is that what you think drives more of the blue zone phenomenon? Like a good combination of those? Or maybe, for those listening who are not familiar, a little bit of background of what’s driving that?
Nate Ruch: Yeah, I think it’s hard to say. I think it’s complicated to research. There’s social connectedness and daily movement, and there are many, many factors that influence lifespan. And there are pockets like Loma Linda, California, and Okinawa in Japan, where on average folks live longer than in other geographies. And how much of that is social, how much of that is environmental, how much of that is behavior — it’s a tricky thing to study. And some component of it is definitely poor record keeping, right, in some parts of the world.
Stephan Shipe: So have they redone those studies? From someone who’s coming from a background of academic studies that are numbers-based, right — this is super interesting to find out that you could even attempt to put numbers associated with this type of longevity compared to all of these other factors. If you had to nail down one of those things across all of those different countries or locations, what do you think it is from your experience?
Nate Ruch: One what? One behavior?
Stephan Shipe: Yeah, what behavior or environmental factor? What do you think is driving most, or most compelling to you based on the research or your experience in it?
Nate Ruch: I don’t think you can distill it to one factor. My sense — and I don’t have super robust evidence behind this — is that if you move more than average, if you eat better than average, if you’re more socially connected than average, you’re highly likely to have better health outcomes in the long term. And sometimes it’s more about what you’re not doing than what you are doing, right? If you’re living in an environment where you’re inhaling a lot of low-quality air, where you’re eating a really processed diet, you’re not moving around a whole lot — those are things that are going to drive, no matter what supplements you take, what healthcare you get, your probability of having the best life possible is going to be lower.
Stephan Shipe: Do you think the research that’s going on now is localized into any certain countries, or is it pretty widespread? Is this a US interest, or is this kind of everywhere right now?
Nate Ruch: Well, I think every year it becomes broader and broader. And one of the natural things that happens as you get out of survival mode and accumulate more resources is you start to think about things like that. So as the standard of living increases in more places, you’re going to see more interest in longevity for sure. And you’re going to see more diversity in terms of research and approaches. And that’s definitely happening.
What’s Coming Next — and How AI Is Already Changing the Field
Stephan Shipe: Yeah, what are you most excited about on the research front? For someone who’s not in the area, and you hear about all these — you hear Yamanaka factors and all these different types of things being thrown around. What makes you excited about the research in this area in the next five to ten years, let’s say?
Nate Ruch: Yeah, I think one of the things I’m really excited about is evidence-driven prevention of heart disease. And for many years, cardiology and medicine in general has treated heart disease based on this idea of predicting ten-year risk using risk calculators, where they’ll put in — do you smoke? What’s your family history? What’s your cholesterol? And that approach is really sunsetting. And we have diagnostics that can measure, like, well, how much heart disease do you have compared to an average person? And my view is that every person should have that done in midlife. So you could see, well, where do you stand? And then you can use that information on where you stand to power the intensity with which you pursue modifying what you can control. And that’s a big shift from what’s been done before. I’ve always thought, from being in medical school and learning about these risk calculators, it always seemed to me that if you’re a forty-five-year-old and your ten-year risk of having a heart attack is low, I mean, that’s really great. But why do you care about ten-year risk? How about forty-year risk, or lifetime risk? Like, how about never having a heart attack? And in order to power that, we now have the tools to do that. So that’s really exciting. I think that’s something concrete that’s available today, if you see the right folks.
In terms of something that’s coming down the pike that I’m excited about, there’s all this technology around omics. So proteomics and metabolomics. And essentially what that is is there are all these intermediate things that happen between your genes and the output into your body. And if you go get a hundred lab tests at LabCorp, you go get function health labs done at Quest, you get a lot of information about yourself, but boy, you’re just measuring the tip of the iceberg. There’s a gigantic amount of information that’s not being measured. And right now there’s a huge amount of science developing around this idea of omics and measuring all these intermediate things. And I think that’s going to power much more accurate customization of care, because there’s a lot more individual variability between folks than people appreciate. And many times medicine gives advice to patients based on what works on average in a big population. And what works on average in a big population doesn’t always distill down to the individual. And sometimes people get either bad advice, or they adhere to a lifestyle practice that really degrades their quality of life, and sadly, it doesn’t move the needle on improving their health. So yeah, I’m super excited about that. You mentioned Yamanaka factors and cellular reprogramming. Definitely super interesting, but a lot of work needs to be done there before that has utility in humans.
Stephan Shipe: Are you using or able to leverage AI at all for all this? I’m imagining if someone comes in, they have all of these different diagnostic tests — I mean, they’ve got the whole portfolio to bring you. Are you able to leverage that now, or have you used any of those tools to see if you can distill that?
Nate Ruch: Yeah, for both the things that I mentioned, AI plays a big role. So when we do CT coronary angiography on patients, we take the resulting images and we send them to a cloud-based machine learning algorithm, or a cloud-based AI, that has a superhuman ability to categorize the plaque into different subtypes and to help us tell where the patient sits in the continuum of heart disease. And also, with serial measurement, it helps us track over time. And it can track over time in a way that’s beyond what a human reader can do. So the more context you give AI, the better the output you get. So we definitely use AI tools where we put in every scrap of information we know about the patient, and we get insights that would be challenging to get.
The other place AI is really useful is combing through very large amounts of medical information and summarizing it in an accurate way. Or if somebody gives you a thousand pages of information out of their Epic MyChart, it’s challenging to digest all that. But if you have an AI agent take a look at that, you can easily summarize things. You can find very specific pieces of information quickly. You can trend things over time. It’s really useful.
Stephan Shipe: Are you seeing a lot of pushback on that in the field through your career? Because you just made a very compelling argument in my mind of, why wouldn’t you want to have all of that, especially when you’re talking about trends and historical tests as well, not just that one test?
Nate Ruch: Yeah, I don’t really understand folks that are against it. I’m a huge proponent. I learn about and use as many of the tools as I can. I think it’s a really interesting thing. If you look at the places where AI doom has been forecasted — probably radiology and medicine was the first place that I’m aware of where credible folks said, well, I would definitely not tell my kids to go into radiology, they’re going to be replaced by AI. And some of those forecasts go back 10 years now. And if you fast forward to today and you look at what’s happening in the employment market for radiologists, there’s a big shortage. There are more exams being done than ever. And it is true that AI penetration is high, but the AI is augmenting human readers, not replacing them.
And the AI is doing things like — we just got a brand new MRI machine a couple months ago that has an AI built into it that helps separate what’s signal and what’s noise and helps curate the images even before the radiologist sees them. And I couldn’t be more bullish on it. I think that the human component of medicine — of sitting across from a patient at the bedside — is not going to be replaced by AI anytime soon. But a lot of the kind of low-value cognitive load, and the summarization of huge amounts of information, and the looking through big data sets for patterns — those are tasks that are much more suited to AI than to humans. And the faster the better. I used to have a research assistant that I would work with, where if I had a clinical question, this person would — years ago, they would go to the medical library and print off journal articles. And then a lot of it kind of became over the internet, so they would be able to get a whole bunch of articles from PubMed or from the medical library and summarize them. Now I don’t need any of that. If I have any clinical question, I can summarize the world literature on it in seconds using an AI model. I mean, it’s fantastic.
Stephan Shipe: That’s great. That’s really exciting, I think, on the combination of not only where the research in this area is going and all the AI. So that’s wonderful. Well, Nate, I really appreciate you coming on today and sharing some of this with us. It’s an exciting area, and I appreciate you taking the time.
Nate Ruch: My pleasure. Thanks for having me.
Outro
Stephan Shipe: That’s our show. Thanks for listening, and we’ll see you next week.
Disclaimer: The information provided in this podcast is for general informational and educational purposes only, and is not intended to constitute financial, investment, or other professional advice. The opinions expressed are those of the hosts and guests and do not necessarily reflect the views of any affiliated organizations. Investing in financial markets involves risk, including the potential loss of principal. Past performance is not indicative of future results. Before making any investment decisions, you should consult with a qualified financial advisor who can assess your individual financial situation, objectives, and risk tolerance.