The Exchanges

Q provide a solution, um, we are seeing them start to come and, and take part. And in fact, both of you wrote another article that said, Hey tech, it's time to build in healthcare. And we're seeing founders from Instacart, Spotify, and Coinbase. Probably among many others that have started to build health care companies. So why do you think that is? What's the why now of these people participating?

A Yeah, I think there are a couple different elements, but one of the primary ones is that much of healthcare and life sciences is starting to become much more of an engineering discipline. I think that's what really attracts us to tech. I mean, the thing about computer science is that it's almost like the ultimate engineering discipline, and software is just easy to engineer. But I think as we can engineer these other areas, it attracts that mindset. But also, I think they see the huge TAM that Daisy was just talking about. But then finally, I think it's more than just The means and, and the, the, the money. I think the, the mission actually also is, is very, very appealing. And the fact that you can actually change the world's healthcare, that itself draws a ton as well.

AI assessment note: “healthcare and life sciences is starting to become much more of an engineering discipline”

Q in healthcare, but that's a really broad brush. And we've maybe touched on like consumer, but that's also a broad brush. So we all know there's a few technologies that seem to really matter. AI. Some people are looking at like augmented reality is another potential interesting entry into healthcare. Um, what is catching your eye, um, in terms of the kind of technologies that can really move the needle?

A I can get off, you know, I love to talk about AI and I think that's a natural one. I think there's two sides of the coin that's really interesting. One is that, um, you could argue healthcare really needs AI in that the cost of healthcare is rising exponentially. There's a huge amount of data. There's all of these services that could maybe become compute. And it's not going to start with like brain surgery. It's going to start with a simple sort of maybe subclinical things first. And we've seen AI for nurses and AI for maybe therapists and other things like that. And that's going to be a huge change in the cost curve. And the funny thing about, or the amazing thing about technology is that what AI will do is by making this technology, we'll all get the best. We'll all get the best nurse, the best therapist. Kind of like the Spotify in my pockets, the same Spotify, um, Elon Musk has, you know, the same Spotify billionaires have. Technology is a great democratizer. So we'll see that. I think the things that people don't realize though is there's a flip. I was talking about why healthcare needs AI. I would make an argument that AI needs healthcare. And the argument there is that actually, where is a place that AI could be applied to where the cost is so high that justify the cost of AI right now? Because AI is really expensive. These models cost, you know, billions of dollars, the…

AI assessment note: “I love to talk about AI and I think that's a natural one.”

Q provide a solution, um, we are seeing them start to come and, and take part. And in fact, both of you wrote another article that said, Hey tech, it's time to build in healthcare. And we're seeing founders from Instacart, Spotify, and Coinbase. Probably among many others that have started to build health care companies. So why do you think that is? What's the why now of these people participating?

A Yeah, I think there are a couple different elements, but one of the primary ones is that much of healthcare and life sciences is starting to become much more of an engineering discipline. I think that's what really attracts us to tech. I mean, the thing about computer science is that it's almost like the ultimate engineering discipline, and software is just easy to engineer. But I think as we can engineer these other areas, it attracts that mindset. But also, I think they see the huge TAM that Daisy was just talking about. But then finally, I think it's more than just The means and, and the, the, the money. I think the, the mission actually also is, is very, very appealing. And the fact that you can actually change the world's healthcare, that itself draws a ton as well.

AI assessment note: “healthcare and life sciences is starting to become much more of an engineering discipline”

Q like the apple of healthcare, or you talked about payments and, you know, kind of replicating in some way what Visa's been able to achieve. So why wouldn't it come from them, right? We, we actually have heard that quite a few of these large fang-like companies are, you know, trying to enter healthcare because it is such a big market, to your point. So could it come from them?

A It could, but I think this is the traditional innovator's dilemma dilemma. Right. And that it's really hard for incumbents. I think occasionally incumbents can disrupt themselves, which is pretty rare. And I think this is a case that would be particularly hard to because it's just so technical and difficult. And I think if you're an incumbent, your, your temptation is to stay away from healthcare. You know, like you, if you can be this massive horizontal and AI, we probably want to do everything except the hard clinical AI stuff. You leave that to something else because that's a whole different beast. And so I suspect that, uh, for those reasons, it's something that The big tech companies will try, and we've seen Apple and Amazon and Google all dabble, but I think those experiments probably won't fit the way they want to build their company, and so I think they'll stay as experiments.

AI assessment note: “It could, but I think this is the traditional innovator's dilemma”

Q it comes to more of the system, it's still breaking down. And then I'll just read out Dave's, uh, comment directly. Um, He says, the U.S. is an outlier in terms of health spend as a percentage of GDP. In terms of a measure of that service, life expectancy in the U.S. is worse. What technology has the rest of the world kept from the U.S.? Why is technology necessary?

A Yeah. So let's take that in two parts. So in terms of technology, I think there is technology in healthcare, but it's largely in a user interface. So like MRI is really the same MRI from like the fifties just now with on a computer display. And so that's not really tech. Uh, that's actually, I would call it the absence of tech. And a lot of the tech that doctors use are things like electronic medical records, which again is kind of a, it's a record keeper. It's not sort of the high technology that we're talking about right now. And so what is different today is that we have This huge revolution in artificial intelligence that actually can do amazing things with data that actually we could never really even dream about, and so that's a huge difference in terms of why now. Now, in terms of what's actually different about the rest of the world is, well, those non-US healthcare systems don't have to deal with Americans, and American culture is different, right? So it's apples and oranges to say, well, we have a country A's healthcare system dealing with country B's people. And so the issues Americans have are different, and we have issues in terms of diet, in terms of lifestyle, and those things are independent of technology or not. Actually, you can imagine what, ah, and so, you know, we actually love doing clinical trials. The real thing to do would be to apply one system to anot…

AI assessment note: “those non-US healthcare systems don't have to deal with Americans, and American culture is different”

Q in healthcare, but that's a really broad brush. And we've maybe touched on like consumer, but that's also a broad brush. So we all know there's a few technologies that seem to really matter. AI. Some people are looking at like augmented reality is another potential interesting entry into healthcare. Um, what is catching your eye, um, in terms of the kind of technologies that can really move the needle?

A I can get off, you know, I love to talk about AI and I think that's a natural one. I think there's two sides of the coin that's really interesting. One is that, um, you could argue healthcare really needs AI in that the cost of healthcare is rising exponentially. There's a huge amount of data. There's all of these services that could maybe become compute. And it's not going to start with like brain surgery. It's going to start with a simple sort of maybe subclinical things first. And we've seen AI for nurses and AI for maybe therapists and other things like that. And that's going to be a huge change in the cost curve. And the funny thing about, or the amazing thing about technology is that what AI will do is by making this technology, we'll all get the best. We'll all get the best nurse, the best therapist. Kind of like the Spotify in my pockets, the same Spotify, um, Elon Musk has, you know, the same Spotify billionaires have. Technology is a great democratizer. So we'll see that. I think the things that people don't realize though is there's a flip. I was talking about why healthcare needs AI. I would make an argument that AI needs healthcare. And the argument there is that actually, where is a place that AI could be applied to where the cost is so high that justify the cost of AI right now? Because AI is really expensive. These models cost, you know, billions of dollars, the…

AI assessment note: “I love to talk about AI and I think that's a natural one.”

Q So take me up a level then beyond the individual Patient experience and talk about this at a structural healthcare system, systemic level. What does this mean for insurers, for hospitals, for researchers, for clinics, drug testing, we don't have to boil the ocean here, but how does this play into that?

A There's so many different ways that data can inform us. I, you know, so just one off the top of my head is that if you're a pharmaceutical company, you want to be able to understand how things are going for patients, maybe in a clinical trial, maybe in a baseline, and you need to be able to get a large number of patients that are the right ones. So this is not necessarily a hundred million people. This is maybe thousands of the right ones that can give you the data you need. And what I think we'll start to see is, especially as new statistical methods come online, that real world evidence will be very useful, in some ways will be more useful than what you can do in a clinical trial, both due to power and due to the fact of real life is different than a clinical trial.

AI assessment note: “if you're a pharmaceutical company, you want to be able to understand”

Q So that tells me how it informs public health and people thinking about this, but how can something like what we're talking about where this dark data unveiling Actually solve. I'm not, again, trying to advocate for solutionistic view. It's a larger, bigger problem beyond technology, but how can it help?

A Yeah, I mean, there's different issues with the opioid crisis. One is that often the first intervention is opioids while you wait to see the back surgeon or, you know, musculoskeletal surgeon. That's an accessibility issue. And then, then once that starts, then the problem is that the opioids are more accessible than the other solutions. And so now you have people sort of doing a doctor or opioid arbitrage between places. And so hopefully we can understand first why this is happening, and then why it starts to sprout, and then what's facilitating it. And having the records in one place would do it. I think it's a little challenging because if it's driven by the patient, the patient's going to want to have to...

AI assessment note: “hopefully we can understand first why this is happening... having the records in one place”

Q it comes to more of the system, it's still breaking down. And then I'll just read out Dave's, uh, comment directly. Um, He says, the U.S. is an outlier in terms of health spend as a percentage of GDP. In terms of a measure of that service, life expectancy in the U.S. is worse. What technology has the rest of the world kept from the U.S.? Why is technology necessary?

A Yeah. So let's take that in two parts. So in terms of technology, I think there is technology in healthcare, but it's largely in a user interface. So like MRI is really the same MRI from like the fifties just now with on a computer display. And so that's not really tech. Uh, that's actually, I would call it the absence of tech. And a lot of the tech that doctors use are things like electronic medical records, which again is kind of a, it's a record keeper. It's not sort of the high technology that we're talking about right now. And so what is different today is that we have This huge revolution in artificial intelligence that actually can do amazing things with data that actually we could never really even dream about, and so that's a huge difference in terms of why now. Now, in terms of what's actually different about the rest of the world is, well, those non-US healthcare systems don't have to deal with Americans, and American culture is different, right? So it's apples and oranges to say, well, we have a country A's healthcare system dealing with country B's people. And so the issues Americans have are different, and we have issues in terms of diet, in terms of lifestyle, and those things are independent of technology or not. Actually, you can imagine what, ah, and so, you know, we actually love doing clinical trials. The real thing to do would be to apply one system to anot…

AI assessment note: “those non-US healthcare systems don't have to deal with Americans, and American culture is different”

Q So that tells me how it informs public health and people thinking about this, but how can something like what we're talking about where this dark data unveiling Actually solve. I'm not, again, trying to advocate for solutionistic view. It's a larger, bigger problem beyond technology, but how can it help?

A Yeah, I mean, there's different issues with the opioid crisis. One is that often the first intervention is opioids while you wait to see the back surgeon or, you know, musculoskeletal surgeon. That's an accessibility issue. And then, then once that starts, then the problem is that the opioids are more accessible than the other solutions. And so now you have people sort of doing a doctor or opioid arbitrage between places. And so hopefully we can understand first why this is happening, and then why it starts to sprout, and then what's facilitating it. And having the records in one place would do it. I think it's a little challenging because if it's driven by the patient, the patient's going to want to have to...

AI assessment note: “having the records in one place would do it”

Q brings up the point around these technologists and it's like, what are technologies uniquely good at? And I'm thinking about, about some of these consumer engagement strategies like, you know, retention or like, um, gamification. So I'd love to hear from both of you. Where do you see some of these engagement strategies that maybe tech has previously flourished being applied to healthcare that seems like an obvious potential fit?

A I think it's going to be all of it. I mean, you mentioned two in terms of, uh, retention gamification. I think it starts with just even, uh, there's the product and that's the gamification and other types of things. And then there's the mindset for how to run the business. And like often, uh, to Daisy's earlier point, when we're not paying where people don't worry about retention, right? They don't worry about because you want your healthcare or don't want, you don't want your healthcare, right? I think those are not really choices. And so when you actually have choices, then people actually have to worry about retention. And that will actually feed into product. So it's that whole ecosystem shifting where we are much more in the driver's seat in terms of choice will lead to better products, as we've seen in many cases.

AI assessment note: “I think it's going to be all of it. I mean, you mentioned two”

Q they're like, yes, yes, yes, I agree that I want this to change, but maybe I don't necessarily agree that technologists are the people right now who are going to be able to make this shift. Um, I guess, what would you, what would you say to those people, and also maybe what would you say to the technologists who maybe are on the fence who are thinking of participating?

A Let me take the first one. So, uh, so in terms of technologies, I think part of it is that if you're a medical professional, I think the key thing is all these technologies are not necessarily complete outsiders. You know, there are people that also know the healthcare system in and out. And a lot of our founders, I think are in that category. So they're both. And so this is something where it's, it's, it's because it does feel a little odd that like some person who just graduates from an undergrad degree with a undergrad CS degree is going to solve all the healthcare. That's probably not the case. Uh, and, and especially you have to understand all of it, but I think part of it is, I mean, the first thing I'll tell them is that these people are, um, are deeply rooted in the medical system, but are also technologists, and that's something that just didn't exist before, and that's kind of, ah, another why now.

AI assessment note: “these people are, um, are deeply rooted in the medical system, but are also technologists”

Q emergence of now, today those options are very consumer-driven, right? I'm buying my Eight Sleep, I'm buying my Oura Ring. Do you think that all of those devices, the holistic healthcare system outside of maybe the hospital where you get your serious surgery, are those all consumer-driven, where I'm purchasing those, or how do you see the overall, like, large incumbent model Evolve to that picture. Does that make sense?

A So yeah, I can take it for, I can start. Yeah. So one thing is, I think that's helpful is actually healthcare is a term that's applied to a lot of different things. And so in the most extreme case, it's like, you know, I need surgery because I, I, I hit my head and my brain is hemorrhaging or something like that. That's like this extreme kind of thing. That's, uh, hospitals. That's, you know, that's not what we're talking about. Then, uh, then the second one is like, oh, I've gotten ill with something and I need some sort of treatment. In both of those cases, those are examples of sick care. Where it's something that, um, those are procedures that just have to be done. There's a third area that actually isn't always associated with healthcare, which is the preventative side. And that's something that technology has a very natural role. With preventative, you can avoid getting into that second or to that third bucket. And when you start to think of healthcare that way, just even from a financial point of view, the, the most extreme thing is kind of like if you were a car, that's like, um, collision insurance or something like that. And then the middle bucket is, Well, I'm gonna, like, just do routine maintenance and stuff like that. But the first bucket is, I'm not gonna be a crazy driver. You know, I'm not gonna slam into things. I'm not gonna go super fast. I'm gonna do all th…

AI assessment note: “healthcare is gonna be that combination of all those three... payment of these things will reflect those”

Q How do you, in this particular example, know?

A Yeah, I mean, that's the billion dollar or trillion dollar question, right? Yeah, literally. And so we're, you know, when it comes down to when, at least for me, what I'm looking for is both In Safi's language, that there is a destination, that we see where you would land, and that at each step of the way, that there is a way to, to do it through engineering-like execution. There's all these different places where bio plays a role, either with consumer, or with healthcare, or, um, other industries. And in each one, it's just not enough that there's a technology and a product that there really, there's product-market fit, and that's always unavoidable, and product-market fit is one of the hardest things for anyone to sort of just You know, because you can very easily be wrong, you can, or you can be wrong, could be three years too early.

AI assessment note: “what I'm looking for is both In Safi's language, that there is a destination”

Q brings up the point around these technologists and it's like, what are technologies uniquely good at? And I'm thinking about, about some of these consumer engagement strategies like, you know, retention or like, um, gamification. So I'd love to hear from both of you. Where do you see some of these engagement strategies that maybe tech has previously flourished being applied to healthcare that seems like an obvious potential fit?

A I think it's going to be all of it. I mean, you mentioned two in terms of, uh, retention gamification. I think it starts with just even, uh, there's the product and that's the gamification and other types of things. And then there's the mindset for how to run the business. And like often, uh, to Daisy's earlier point, when we're not paying where people don't worry about retention, right? They don't worry about because you want your healthcare or don't want, you don't want your healthcare, right? I think those are not really choices. And so when you actually have choices, then people actually have to worry about retention. And that will actually feed into product. So it's that whole ecosystem shifting where we are much more in the driver's seat in terms of choice will lead to better products, as we've seen in many cases.

AI assessment note: “I think it's going to be all of it. I mean, you mentioned two”

Q emergence of now, today those options are very consumer-driven, right? I'm buying my Eight Sleep, I'm buying my Oura Ring. Do you think that all of those devices, the holistic healthcare system outside of maybe the hospital where you get your serious surgery, are those all consumer-driven, where I'm purchasing those, or how do you see the overall, like, large incumbent model Evolve to that picture. Does that make sense?

A So yeah, I can take it for, I can start. Yeah. So one thing is, I think that's helpful is actually healthcare is a term that's applied to a lot of different things. And so in the most extreme case, it's like, you know, I need surgery because I, I, I hit my head and my brain is hemorrhaging or something like that. That's like this extreme kind of thing. That's, uh, hospitals. That's, you know, that's not what we're talking about. Then, uh, then the second one is like, oh, I've gotten ill with something and I need some sort of treatment. In both of those cases, those are examples of sick care. Where it's something that, um, those are procedures that just have to be done. There's a third area that actually isn't always associated with healthcare, which is the preventative side. And that's something that technology has a very natural role. With preventative, you can avoid getting into that second or to that third bucket. And when you start to think of healthcare that way, just even from a financial point of view, the, the most extreme thing is kind of like if you were a car, that's like, um, collision insurance or something like that. And then the middle bucket is, Well, I'm gonna, like, just do routine maintenance and stuff like that. But the first bucket is, I'm not gonna be a crazy driver. You know, I'm not gonna slam into things. I'm not gonna go super fast. I'm gonna do all th…

AI assessment note: “And how we think about payment of these things will reflect those.”

Q And isn't there also actually stuff you cannot even do in classical computers today with, for computational chemistry?

A Uh, well, it depends on the number of atoms. So, like, you can do this full calculation for, like, tens to maybe a hundred atoms, but you probably couldn't, it'd be very difficult, I think, to do it for a thousand atoms or 10,000 atoms. I mean, because we, all, chemistry is everywhere, and you think about energy, uh, when you drive a car, it's going through chemical reactions. When you, your proteins in your body are working, those are chemical reactions. I mean, your, uh, plants are growing and using, uh, fertilizer and all this stuff. It's all chemistry. There could be one that gets into market. And that we start seeing use and uptake, and that's where it starts getting interesting, because once there is an application and these things become cheaper and more ubiquitous, I think we're going to see the second, the third, the fourth, and fifth, and it's going to roll from there.

AI assessment note: “depends on the number of atoms... very difficult, I think, to do it for a thousand”

Q effects are really important for software-based companies, Especially in this age, as you mentioned in machine learning, deep learning, AI, all the things kind of trends coming together. So what concretely can entrepreneurs do to a build data network effects or think more strategically about it early on versus by accident? And secondly, what do you want to see in pitches from entrepreneurs when they talk about data network effects?

A Yeah. So, you know, in terms of a startup, usually startups do well when they focus on one area. And so the challenge here is that how can the data network effect really accelerate what they're doing? I think too often that, um, what happens is the data network effect almost suggests a side business or something like that. And, and so one challenge is how to think about what is the real go to market? Is the data network effect really germane and central and key to the focus? And then how can you monetize it? How can you take advantage of it? Um, What often happens is I think there's the aspiration for taking advantage of the data network effect or the assumption that it will just come. But, uh, often we see, uh, situations where maybe that plan hasn't been well thought out yet.

AI assessment note: “Is the data network effect really germane and central and key to the focus?”

Q I'm going to ask a really obvious question though, which is why is cancer so hard? I get that it's a complex disease.

A Yeah, I think there's a couple answers. One, it's not a disease. It's really many diseases. Number two, actually defining is when I was studying biology, uh, many years ago, you know, my knee-jerk reaction immediately after reading this stuff is like, I'm amazed it works at all. So the fact that apoptosis breaks down, or there's some somatic mutations, that mutations happen, that's actually not that shocking. I think the fact that actually our body does such a good job Of maintaining things to me often sounds even more surprising, but, you know, we can do more to help it. In addition to machine learning, be able to come up with the equivalent of new types of biomarkers that people couldn't come up with, there's a learning aspect here of machine learning, which is intriguing, that as you get more data, you get better, and that's something that's really not like any other test, where, you know, a lipid blood test doesn't get better as you have more patients there, and that's, I think, a really intriguing aspect, especially As we want to detect cancer earlier and earlier, where there's fainter and fainter signs, and especially for a wide range of cancers.

AI assessment note: “One, it's not a disease. It's really many diseases.”

Q How do you, in this particular example, know?

A Yeah, I mean, that's the billion dollar or trillion dollar question, right? Yeah, literally. And so we're, you know, when it comes down to when, at least for me, what I'm looking for is both In Safi's language, that there is a destination, that we see where you would land, and that at each step of the way, that there is a way to, to do it through engineering-like execution. There's all these different places where bio plays a role, either with consumer, or with healthcare, or, um, other industries. And in each one, it's just not enough that there's a technology and a product that there really, there's product-market fit, and that's always unavoidable, and product-market fit is one of the hardest things for anyone to sort of just You know, because you can very easily be wrong, you can, or you can be wrong, could be three years too early.

AI assessment note: “what I'm looking for is both In Safi's language, that there is a destination”

Q effects are really important for software-based companies, Especially in this age, as you mentioned in machine learning, deep learning, AI, all the things kind of trends coming together. So what concretely can entrepreneurs do to a build data network effects or think more strategically about it early on versus by accident? And secondly, what do you want to see in pitches from entrepreneurs when they talk about data network effects?

A Yeah. So, you know, in terms of a startup, usually startups do well when they focus on one area. And so the challenge here is that how can the data network effect really accelerate what they're doing? I think too often that, um, what happens is the data network effect almost suggests a side business or something like that. And, and so one challenge is how to think about what is the real go to market? Is the data network effect really germane and central and key to the focus? And then how can you monetize it? How can you take advantage of it? Um, What often happens is I think there's the aspiration for taking advantage of the data network effect or the assumption that it will just come. But, uh, often we see, uh, situations where maybe that plan hasn't been well thought out yet.

AI assessment note: “often we see situations where maybe that plan hasn't been well thought out yet”

Q I'm going to ask a really obvious question though, which is why is cancer so hard? I get that it's a complex disease.

A Yeah, I think there's a couple answers. One, it's not a disease. It's really many diseases. Number two, actually defining is when I was studying biology, uh, many years ago, you know, my knee-jerk reaction immediately after reading this stuff is like, I'm amazed it works at all. So the fact that apoptosis breaks down, or there's some somatic mutations, that mutations happen, that's actually not that shocking. I think the fact that actually our body does such a good job Of maintaining things to me often sounds even more surprising, but, you know, we can do more to help it. In addition to machine learning, be able to come up with the equivalent of new types of biomarkers that people couldn't come up with, there's a learning aspect here of machine learning, which is intriguing, that as you get more data, you get better, and that's something that's really not like any other test, where, you know, a lipid blood test doesn't get better as you have more patients there, and that's, I think, a really intriguing aspect, especially As we want to detect cancer earlier and earlier, where there's fainter and fainter signs, and especially for a wide range of cancers.

AI assessment note: “In addition to machine learning, be able to come up with”