309 ‒ AI in medicine: its potential to revolutionize disease prediction, diagnosis, and outcomes

1. 0:00 – 1:08

   Intro

   1. IKIsaac "Zak" Kohane0:00

      Let me quickly say, uh, the old very bad joke. What do you call the, uh, doct- the medical student who graduates at the bottom of his class?

   2. PAPeter Attia0:07

      A doctor. [laughs]

   3. IKIsaac "Zak" Kohane0:09

      Doctor. And so if you could just merely get the t- bottom fifty percent of doctors to be as good as the top fifty percent, that'd be transformative for healthcare. Now, there are other superhuman capabilities that we will-- that we can go towards, and we can talk about if we want, that do require the next generation of algorithms, NVIDIA, um, architectures, and data sets. But there's so much. If everything stopped now, we could already transform medicine. It's just a matter of the sweat equity to create the models, figure out how to include them in the workflow, how to pay for them, um, how to create a reimbursement system that, uh, and a business model that, uh, works for our society. But there's no technological, um, barrier.

   4. PAPeter Attia1:04

      [upbeat music] Hey, everyone. Welcome to the Drive podcast.

2. 1:08 – 7:56

   Zak’s unconventional journey to becoming a pioneering physician-scientist, & his early interactions with GPT-4

   1. PAPeter Attia1:08

      I'm your host, Peter Attia. [upbeat music] Well, Zak, thank you so much for joining me today. This is a, a topic that's highly relevant, um, and, and, and one that I've wanted to talk about for, for some time, um, but, you know, wasn't sure who to speak with. And so, uh, we eventually kinda found our way to you. So, um, again, thanks for, for making the time and sharing your expertise. Um, l- give folks a little bit of a, a sense of your, your background. Um, what, uh, what was your path through medical school and training? It was, it was not a very typical path.

   2. IKIsaac "Zak" Kohane1:44

      No. So, uh, what happened was I grew up in Switzerland. Uh, nobody in my family was a doctor. Uh, come to United States, decide to major in med-- in, uh, biology, and then, uh, I get nerd sniped by computing back in the '70s, and, uh, in the late '70s. And so I minor in computer science, but I still complete my, uh, degree in biology and I go to medical school. And then in the middle of medical school's first year, I realized, holy smokes, this is not what I expected. It's a noble profession. It's-- But it's not a science. It's an art. It's not a science, and I thought I was going into science.

   3. PAPeter Attia2:27

      [laughs]

   4. IKIsaac "Zak" Kohane2:27

      And so, um, I bail out for a while to do a PhD in computer science. And this is during the 1980s now, early 1980s, and it's the heyday of AI. It's actually second heyday. This is now the-- we're going through the third heyday, and it was a time of great promise and, uh, with, uh, retrospective scope, very clear that it was not, uh, going to be successful, that, uh, there was a lot of overpromising. There is today. But unlike, uh, today, we had not released it to the public. It was not actually working in the way that we thought it was going to work, and certainly didn't scale. And so, uh, it was a very interesting period, and my thesis advisor, Peter Szolovits, a professor at MIT, said, uh, "Zak, you should finish your clinical training because I'm not getting a lot of respect from, uh, clinicians. And so to bring rational decision-making to the clinic, uh, you really wanna finish your clinical training." And so I finished medical school, did a residency in pediatrics and then pediatric endocrinology, which was actually extremely enjoyable. Uh, but when I was done, I restarted my research in, uh, computing, started a lab at Children's Hospital in Boston, and then, uh, a center, um, of biomedical informatics at the medical school. Like in every other, almost every other endeavor, getting, uh, money gets attention from the powers that be, and so I was getting a lot of grants. And so they asked me to start the center, and then eventually a new Department of Biomedical Informatics that I'm the chair of, and that we have now, uh, sixteen professors or assistant professors of biomedical informatics. And [laughs] then, um, I have been involved in a lot of machine learning projects, but like everybody else, I was taken by surprise, except perhaps a little bit earlier than most, by large language models. I got a call from Peter Lee, uh, in October, October '22. And actually, I didn't get a call. It was an email, like right out of a Michael Crichton novel. It said, "Zak, if you'll answer the phone, I can't tell you what it's about, but it'd be well worth your while." And so, um, I get a call from Peter Lee, and I knew him from before. He was a professor of, uh, computer science at CMU, and also department chair there. He went to ARPA, and then he went to Microsoft. And he tells me about GPT-4, and this was before any of us had heard about G-- uh, ChatGPT, which is initially, uh, GPT-3.5. He tells me about, uh, GPT-4, and he gets me early access to it when no one else knows it, uh, that it exists. Only a few people do. And I start trying it against hard cases, like child in the-- I get called down. I just remember from my training. I get called down to the nursery. It's a child with, um, a small phallus and a hole at the base of the phallus, and they can't palpate testicles, and they wanna know what to do, because I'm a pediatric endocrinologist. And so I asked GPT-4, "What would you do? What are you thinking about?" And it runs me through the whole, um- Workup of these very rare cases of, uh, ambiguous genitalia. In this case, it was congenital adrenal hyperplasia, where the making of excess androgens during, um, pregnancy and then subsequently during birth causes clitoris to swell from, uh, glands of the penis or of the phallus, and the liga mentora to fuse to form the shaft of the-- what looks like a penis. But there's no testicles, there's ovaries. And, um, so there's a whole endocrine workup with genetic tests, hormonal tests, ultrasound, and it does it all. And it blows my mind. It really blows my mind because very few of us in computer science really thought that these large language models would scale up the way they do. It was just not expected. And, you know, I was talking to, uh, Bill Gates about this after, uh, Peter Lee had introduced me to the problem, and he told me that his line engineers, that in Microsoft research, a lot of his fancy-- fanciest computer scientists did not expect this. But the line engineers at Microsoft were just watching the scale-up, you know, GPT zero, one, two, and they just saw it was gonna keep on scaling up with, uh, the, the size of the data and with the size of the model. And they said, "Yeah, of course, it's gonna achieve this kind of expertise." But the rest of us, I think because we value our own intellects so much, we couldn't imagine how we would get that kind of conversational expertise just by scaling up the model and the dataset.

3. 7:56 – 21:00

   The evolution of AI from the earliest versions to today’s neural networks, & the shifting definitions of intelligence over time

   1. PAPeter Attia7:56

      Well, Zak, that's actually kind of a perfect introduction to, to how I wanna think about this today, which is to say, look, there's nobody listening to us who hasn't heard the term AI, um, and yet virtually no one really understands what is going on. So if we want to talk about how AI can change medicine, I think we have to first invest some serious bandwidth in understanding AI. Now, you alluded to the fact that when you were doing your PhD in the early eighties, you were in the second era or the second generation of AI, which leads me to assume that the first generatia-- generation was shortly following World War, uh, Two, and that's probably why someone by the name of Alan Turing has his name on something called the Turing test. Um, so, so maybe you can talk us through what Alan Turing posited, what the Turing test was and proposed to be, and really what gen one AI was. We don't have to spend too much time on it, but clearly it didn't work. But let's maybe talk a little bit about the postulates around it and, and, and what it was.

   2. IKIsaac "Zak" Kohane8:58

      So after Wo- World War Two, we had computing machines, and anybody who was a serious computer scientist could see that you could have, uh, these processes that could generate other processes, and you could see how these processes could take inputs and become more sophisticated. And as a result, um, even shortly after, um, World War Two, we actually had artific- artificial neural networks, the per- the perceptron, uh, which was modeled, roughly speaking, on the ideas of a neuron that could take inputs from the, uh, environment and then have certain expectations. And if you updated the neuron as to what was going on, it would update the weights going into that artificial neuron. And so going back to, uh, Turing, he just came up with a, a test that said essentially, if a computational entity can maintain essentially its side of the conversation without revealing that it was a computer and that others would mistake it for a human, then for all intents and purposes, that would be intelligent behavior. And there's been all sorts of, all sorts of, uh, additional constraints put on it, and one of the hallmarks of AI, frankly, is that it keeps on moving the goal- goalposts-

   3. PAPeter Attia10:51

      Mm-hmm

   4. IKIsaac "Zak" Kohane10:52

      ... of what we consider to be intelligent behavior. If you woulda told someone in the sixties that a chess master-- the world chess masters were gonna be beat by a computer program, they'd say, "Well, that's AI. Clearly, that's AI." And if you would s-- and then when, uh, Kasparov was beaten by, uh, Deep Blue, by the IBM, uh, machine, people said, "Well, it's just doing search very well. It's searching through all the possible moves in the future. It also has-- knows all the grandmaster moves. It has a huge encyclopedic store of all the different grandmaster moves. This is no lo-- this is not really intelligent behavior." And then when-- if you told people it could recognize human faces and find your grandmother in a picture, on any picture in the internet, they'd, they'd say, "Well, that's, that's intelligence." And of course, when we did it, no, that was not intelligent. And then when, um, we said it could write a rap poem about Peter Attia based on your webpage, and it did that- Well, that would be intelligent, that would be creative. But then if you said it's doing it based on having created a computational model based on all the text ever generated by human beings, as much as we can gather, which is somewhere between one to six terabytes of data, and this computational model basically is predicting what is the next word that's gonna be said. Not just the next word, but of the millions of words that could be, what are the probabilities-

   5. PAPeter Attia12:41

      Mm-hmm

   6. IKIsaac "Zak" Kohane12:41

      ... of that next word, and that's, that is what's generating that rap. There's people who are arguing that's not intelligence. So the goal posts, posts around the Turing test keep getting moved. I just have to say that I, I no longer find that an interesting topic because it's what it's actually doing.

   7. PAPeter Attia13:02

      Yeah.

   8. IKIsaac "Zak" Kohane13:02

      And whether you wanna call it intelligent or not, that's, uh, that's up to you. It's like discussing whether is a, is a dog int- intelligent? Is a baby, uh, intelligent before it can recognize the, uh, constancy of objects? You know, initially, babies, if you hide something from it, it's gone, and when it comes back, it's a surprise. But at some point, they-- early on, they learn that there are, there's constancy of objects even when they don't see them. So there's, there's this spectrum of intelligent behavior, and I'd just like to remind you, myself, that there's a very simple computational model predicting the next word called a Markov model. And several years ago, people were studying songbirds, and they were able to predict the full song, the next note and the next note of the songbird just using a very simple Markov model. So from that perspective, I know we think that we're all very smart, but the fact that you and I, without thinking too hard about it, can come up with fluid speech, okay, so the model is now a trillion parameters. It's not a simple Markov model, but it's still a model. And perhaps later we'll talk about how this plays into, uh, the late, unfortunately the late, uh, Kahneman's, uh, notions of thinking fast and thinking slow, and this notion of system one, which is this sort of pattern recognition, which is very much similar to what I think we're seeing here, and s- and system two, which is the more deliberate and much more conscious kind of thinking that we pride ourselves on. But a lot of what we do is this sort of reflexive, very fast pattern recognition. I'll stop there.

   9. PAPeter Attia14:54

      Well, so, so if we go back to, you know, World War II, that's, to your point, where we saw basically kind of rule-based computing come of age. And, uh, you know, everybody-- anybody who's gone back and watched movies about the Manhattan Project or the decoding of, um, you know, all the sorts of things that took place, uh, like Enigma, for example. Um, you know, again, that's straight rules-based computational, uh, power, and, and we're obviously at the limits of, you know, we, we-- that can only go so far. Um, but it seems that there was a long hiatus before we went from there to kind of like maybe what some have called context-based, uh, computation, like what your, what your Siri does or Alexa, right? Which is a step quite beyond that. And then, of course, you, you would, you know, sort of go from there to what you've already talked about, Blue or Watson, where you have, you know, computers that are probably going even one step further. And then, of course, where we are now, which is, you know, GPT-4. So it's, it-- and I wanna talk a little bit about the, the computational side of that. But, but more what I wanna get at is this idea that, um, th- there's a, there seems to be a very nonlinear pace at which this is happening. And I, I hear your point. I'd never thought of it that way, um, but the, I, I hear your point about the goalpost moving. Um, but I think your instinct around majoring in the, the right thing is, is also relevant, which is let's focus less on the fact that we're never quite hitting the asymptote definitionally. Let's look at the actual output, and it is staggeringly different. So w- what was it that was taking place during the period of your PhD, what you're calling wave two of AI? What was the, what was the, uh, objective, and where was the failure?

   10. IKIsaac "Zak" Kohane16:55

       So the objective was, okay, we can now take-- So in the first era, you wrote, uh, computer programs in assembler language or in languages like, uh, Fortran. And there was a limit of what you could do. You had to be a comp-- real computational programmer to do something in that, uh, mode. In wave two, in the nine- in the 1970s, we came up with these rule-based systems where we said rules in what looked like English. If there is a patient who has a fever, and you get an isolate from the lab, and that, the bacteria in the isolate is gram-positive, then you might have a streptococcal infection with a probability of so and so. And these rule-based systems, which again, now, you're now programming in the level of human knowledge, not in computer code, the problem with that was severalfold. A, you're gonna generate tens of thousands of these rules, and these rules would interact in ways that you could not anticipate, and we did not know enough, and we could not pull out of human beings the right probabilities. And, um, what is the right probability of you have a, um- A fever, and you don't see anything on, uh, on the blood test. What, what else is going on? And there's a large set of possibilities, and getting all those rules out of human beings ended up being extremely expensive, and the results were not stable. And, and, and for that, and for that reason, because we didn't have much data online, we could not go to the next step, which is have data to actually drive these models.

   11. PAPeter Attia18:46

       What, what were the data sources then?

   12. IKIsaac "Zak" Kohane18:49

       Textbooks and, and journals as, uh, interpreted by human experts. That's why some of these were called expert systems, because they were derived from introspection by experts who would then come up with the rules, with the probabilities, and, and some of the early, um, work like, for example, there was a program called MYCIN, um, run by Ted Shortliffe out of Stanford, who developed a antibiotic advisor that was a set of rules based on what he and his colleagues sussed out from the different infectious disease textbooks and infectious disease experts. And it stayed only up to date as long as they kept on, uh, looking at the literature, adding rules, fine-tuning it. If there was an interaction between two rule, two rules that was not, uh, desirable, then you had to adjust that. Very labor-intensive, and then if there's a new thing, you'd have to add some new rules. If AIDS happened, you'd have to say, "Oh, there's this new pathogen. I have to make a bunch of rules." The probability is gonna be different if you're an IV drug abuser or if you're a male, a homosexual. And so it wa-- it was very, very hard to keep up, and in fact, people didn't.

   13. PAPeter Attia20:12

       What was the language that it was programmed in? Was this Fortran?

   14. IKIsaac "Zak" Kohane20:16

       No, no, this was, uh, it-- these were so-called rule-based s-systems. And so the languages, for example, for the system MYCIN was called EMYCIN, Essential-

   15. PAPeter Attia20:25

       I see

   16. IKIsaac "Zak" Kohane20:25

       ... uh, MYCIN. So these, these looked like English. And they-

   17. PAPeter Attia20:29

       So super, super labor-intensive.

   18. IKIsaac "Zak" Kohane20:31

       Super labor-intensive, and, and there's no way you could keep it up to date. And at that time, there was no electronic, electronic medical records. They were all paper records, so not informed by what was going on in the clinic. And so the three revolutions had to happen in order for, um, us to have what we have today, and that's why I think we, we had such a, a quantum jump,

4. 21:00 – 32:16

   How vast data sets, advanced neural networks, & powerful GPU technology have driven AI from its early limitations to achieving remarkable successes in medicine & other fields

   1. IKIsaac "Zak" Kohane21:00

      uh, recently.

   2. PAPeter Attia21:01

      And before-

   3. IKIsaac "Zak" Kohane21:02

      First we had to have-

   4. PAPeter Attia21:02

      Before we get to that, 'cause I, I mean, I, that-

   5. IKIsaac "Zak" Kohane21:04

      Yes

   6. PAPeter Attia21:04

      ... that's obviously, [chuckles] that's the exciting question. But I just wanna go back to-

   7. IKIsaac "Zak" Kohane21:07

      Yeah

   8. PAPeter Attia21:07

      ... the, the gen two. Were there other-

   9. IKIsaac "Zak" Kohane21:10

      Sure

   10. PAPeter Attia21:10

       ... industries that were having more success than medicine? W-were there applications in the military? Were there applications elsewhere in government where-

   11. IKIsaac "Zak" Kohane21:17

       Yes

   12. PAPeter Attia21:17

       ... they got a little closer to utility?

   13. IKIsaac "Zak" Kohane21:20

       Yes, so, uh, this-- there's a c-company which, uh, uh, was a remnant of a-- [chuckles] back in the, in the nineteen seventies, there was a whole bunch of computer companies around what we called Route One twenty-eight in Boston. And, uh, these were companies that, uh, were famous back then, um, like Wang Computer, like Digital Equipment Corporation, and it's a very sad story for Boston because all then-- that was before Silicon Valley got its pearl of computer companies a-around it. And, um, one of the companies, Digital Equipment Corporation, built a program called R1, and R1, uh, was an expert in config-configuring the mini computers that you ordered. So you wanted some capabilities, and it would actually configure all the industrial components, the, the processors, the disk, and it would know about all the exceptions and what you needed to know, what cabling, what memory, uh, configuration. All that was done, and it basically replaced several individuals who had that very, very, uh, rare knowledge to configure-

   14. PAPeter Attia22:31

       Mm.

   15. IKIsaac "Zak" Kohane22:32

       ... um, their systems. It was also used in several government logistics efforts. But even those efforts, although they were successful and used commercially, were limited, 'cause it turns out human beings, once you got to about three, four, five, six thousand rules, no single human being could keep track of all the p-- ways these rules could work. And so the-- we used to call this the complexity barrier, that these rules would interact in unexpected ways, and you'd get incorrect answers, things that were not commonsensical, because you'd actually not captured everything about the real world. And so it was very narrowly focused, and if the expertise was a little bit outside the area of focus, if, let's say it was a infectious disease program, and there was a little bit of influence from the cardiac status of the patient, and you had not accurately modeled that, its performance would degrade rapidly. Similarly, if there was, in Digital Equipment, a new model that had a complete different part that had been not included, and that there were some dependencies that were not modeled, it would, it would degrade in performance. So these systems were very brittle, did not show common sense. They had expert behavior, but it was very narrowly done. There were applications of medicine back then that survived till today. For example, already back then, we had these systems doing interpretation of ECGs-

   16. PAPeter Attia24:13

       Hmm

   17. IKIsaac "Zak" Kohane24:13

       ... um, uh, a-actually pretty competently, um, at least a first pass until they would be, uh, reviewed by an expert cardiologist. So you- That was, uh, there was also a program that interpreted, um, what's called serum protein electrophoresis, where you look at th-the, uh, proteins separated out by an ele-electric gradient to make a diagnosis, let's say, of myeloma or other protein disorders. And those also were deployed clinically, but they only worked very much in narrow areas. They were by no stretch of the imagination general-purpose, uh, reasoning machines.

   18. PAPeter Attia24:56

       So let's get back to the three things. There are three things that have taken the relative failures of first and second attempts at AI and got us to where we are today. So l-let's-- I can guess what they are, but let's, let's just have you walk us through them.

   19. IKIsaac "Zak" Kohane25:12

       Okay. So first is, um, the first one was just lots of data.

   20. PAPeter Attia25:16

       Mm-hmm.

   21. IKIsaac "Zak" Kohane25:17

       And we needed to have a lot of online data, uh, to be able to develop models of interesting performance and quality. So ImageNet was one of the, uh, first such, uh, data sets, uh, collections of millions of images with annotations, importantly. You know, this has a cat in it, this has a dog in it, this is a blueberry muffin, this has a human in it. And having that was absolutely essential to, uh, allow us to train the first very successful, uh, neural network models. And so having those large data sets was extremely important. The other, and there's equivalent in medicine, which is we did not have a lot of textual information about medicine until PubMed went online. So all the literature, medical literature, at least we have an abstract of it in PubMed, plus we have for a subset of it that's open access because the government has paid for it through grants. Um, there's something called PubMed Central, which has the full text. So all of a sudden, that has opened up over the last, uh, ten years. And then electronic health records, um, after Obama, uh, signed the HITECH Act, electronic health records, which also ruined the lives of many doctors, also happened to also generate a lot of text, um, for, uh, the use in these systems. So that's-- one is large amounts of data being generated online. The second was the neural network models themselves. So the perceptron that I mentioned that was developed not too long after World War II was shown by, uh, one of the pioneers of AI, uh, Marvin Minsky, to have fundamental limitations, uh, in that it could not do, uh, certain mathematical functions like what's called an exclusive or gate. Um, and because of that, people said these neural networks are not gonna scale. But there were a few true believers who kept on pushing and making more and more advanced architectures, and those, um, multi-level deep neural networks. So instead of having one neural network, we layer on top of one, uh, neural network, another one, another one, and another one, so that the output of the first layer gets propagated up to the second layer of net-- of neurons, to the third layer and fourth layer and so on.

   22. PAPeter Attia27:59

       And I'm sorry, was this a theoretical mathematical breakthrough or a technological breakthrough?

   23. IKIsaac "Zak" Kohane28:05

       Both. So it was both because having those, uh, an insight that these-- we could actually come up with all the mathematical functions we needed to, we could simulate them with these multi-level networks. Whereas it was a, uh, theoretical insight, but we'd have never made anything out of it if not for the fact of sweaty teenagers, mostly teenage boys, playing video games. In order to have sh- first-person shooters capable of running high resolution, uh, uh, pictures of aliens or monsters, um, in high resolution, twenty-four bit color, sixty frames per second, we needed to have processors, very parallel processors, that would allow you to render, to do the linear algebra that allow you to calculate what was g-- the intensity of color on every dot of the screen at sixty frames per second.

   24. PAPeter Attia29:11

       And, and, and that's literally just because of the matrix multiplication math that's required to do this. You have N by M matrices that are so big, and you're, you're, you're crossing and dotting huge matrices.

   25. IKIsaac "Zak" Kohane29:26

       Huge, huge, huge matrices. And it turns out that's something that can be run in parallel. So you want to have multiple parallel processors capable of rendering those images, again, at sixty frames per second. So basically, millions of bits on your screen being rendered at, uh, twenty-four or thirty-two bit color. And in order to do that, you need to have the, that linear algebra that you just referred to being run in parallel. And so these parallel processors called graphical processing units, GPUs, were developed, and the GPUs, uh, were developed by several companies, and some of them stayed in business, some of them didn't, but they were absolutely essential to the success of video games. Now, it then occurred to many smart mathematicians and computer scientists that the same linear algebra that was used to drive that computation Or images could also be used to calculate the weights of the edges between the neurons in a neural network. So the mathematics of updating the weights in response to, uh, stimuli, let's say, of a neural network, the updating of those weights can be done all in linear algebra. And if you have this processor that has literally-- So a typical computer has a central processing unit. So that's one processing unit. A GPU has tens of thousands of processors that do this one very simple thing, linear algebra. And so by having this parallelism that only supercomputers would have, typically, on your simple PC, because you needed to show the graphics at sixty frames per second, gave us all of a sudden these commodity chips that allowed us to calculate the performance of these multi-level neural networks. So that theoretical breakthrough was a second part, but it was-- would not have happened without the actual implementation capability that we had with the GPUs.

   26. PAPeter Attia31:50

       And so Nvidia would be the most successful example of this, presumably?

   27. IKIsaac "Zak" Kohane31:55

       Yeah, it was not, it was not the first, but it's definitely the most successful example. And there's a variety of reasons why it was successful and created an ecosystem of, um, of implementers who built their, uh, neural network, uh, deep learning systems on top of, uh, the Nvidia architecture.

5. 32:16 – 36:02

   An AI breakthrough in medicine: the ability to accurately recognize retinopathy

   1. IKIsaac "Zak" Kohane32:16

      So that-

   2. PAPeter Attia32:17

      Was there a moment when that... Would you, would you go back and look at the calendar and say this was the kind of year or quarter when there was escape velocity achieved there?

   3. IKIsaac "Zak" Kohane32:25

      Yeah. So it's probably around two thousand and twelve when there was, there was an ongoing contest every year saying, "Who has the best image recognition software?" And these deep neural networks running off GPUs were able to outperform significantly all the o- their other competitors in image recognition in two thousand and twelve. And that's, uh, that's very clearly when everybody just woke up and said, "Whoa, we knew about neural networks. We didn't realize that these convolu- convolution neural networks were going to be this effective. And seems that the only thing that's gonna be-- stop us is computational speed, the, and the size of our data sets." So that moved things very fast along in the imaging space with a, with very soon consequences in medicine. It was only, uh, six years later that we saw, uh, journal articles about recognition of retinopathy, uh, diseases affecting the, uh, retina, the back of your eye in diabetes, and a paper coming out of all places from Google saying we can recognize different stages of retinopathy based on the images of, uh, the back of the eye. And that also was a wake-up call 'cause, yes, it's, it's a part of the goalpost moving. It was great that we could recognize cats and dogs-

   4. PAPeter Attia34:01

      [laughs]

   5. IKIsaac "Zak" Kohane34:01

      ... in, uh, in, uh, webpages, but now all of a sudden, this thing that we thought was, um, specialized human expertise could be done by that same stack of software. Just if you gave it enough cases of these retinopathies, it would actually work well. And furthermore, what was wild was that there's something called transfer learning, where you look-- you tune up these networks, get them to recognize cats and dogs, and in the, the process of recognizing cats and dogs, it learns how to recognize circles and lines and fuzziness and so on. And you did a lot better in training up the neural network first on the entire ImageNet set of images and then on the retinas than if you just went straight to, "I'm just gonna train on the retinas." And so-

   6. PAPeter Attia34:58

      Hmm

   7. IKIsaac "Zak" Kohane34:58

      ... what was-- So that, that transfer learning was impressive, and then the other thing as a doctor was impressive to many of us. I was actually write-- asked to write an editorial for the Journal of Med-- of the American Medical Association in two thousand and eighteen when that Google article was written. What was impressive to us was that what was the main role of doctors in that publication? It was just twofold. One was to just label the images that were used for training. This is retinopathy. It's not retin- retinopathy. And then to serve as, um, judges of its performance. And that was it. All the, the rest of it was computer scientists working with GPUs and images, tuning it, and that was it. Didn't look anything like medical school, and you were having expert-level recognition of retinopathy. That was a wake-up call.

6. 36:02 – 42:59

   Third generation AI: how improvements in natural language processing significantly advanced AI capabilities

   1. PAPeter Attia36:02

      So you, you've alluded to the two thousand and seventeen paper by Google. A-Attention Is All That Is Needed, I think, is the title of the paper. Attention Is All You Need?

   2. IKIsaac "Zak" Kohane36:12

      Attention Is All You-- So that's not what I'm re-referring to. I was also referring to that, a two thousand and eighteen paper in JAMA where-

   3. PAPeter Attia36:18

      Ah, I'm sorry. Okay

   4. IKIsaac "Zak" Kohane36:19

      ... that was not a trick. You're, you're talking about the great paper, uh, Attention Is All You Need. That was about the invention of the transformer, which is a specific type of neural network architecture. I was talking about these are s- these were vanilla, barely vanilla, uh, convolutional neural networks, the same one that can detect, uh, dogs and cats.

   5. PAPeter Attia36:39

      Yeah, got it.

   6. IKIsaac "Zak" Kohane36:40

      But they... It was a big medical application, retina- retinopathy, two thousand and eighteen. Except for computer scientists, no one noticed the attention-is-all-you-need paper. And Google had this wonderful paper that said, you know, if we recognize not just, uh, text that collocates together, 'cause previously-- So we're gonna back-- Uh, away from images for a second. There was this notion that I can recognize a lot of similarities in text if I see which words occur together. I can implicate the meaning of a word by the company it keeps. And so if I see, uh, this word, and it has around it, uh, kingdom, uh, crown, uh, uh-

   7. PAPeter Attia37:38

      Throne

   8. IKIsaac "Zak" Kohane37:38

      ... castle, throne-

   9. PAPeter Attia37:40

      Yep

   10. IKIsaac "Zak" Kohane37:40

       ... it's about a king.

   11. PAPeter Attia37:41

       Mm-hmm.

   12. IKIsaac "Zak" Kohane37:42

       And similarly for queen and so on. And tho- that kind of, um, association, uh, w- uh, in which we created what was called, uh, w- uh, embedding vectors, which just in plain English, it's a string of numbers that says for any given word, what are-- what is the probability, how often do these other words occur-- co-occur with it? And just using that, those embeddings, those vectors, those lists of, of numbers that describe the co-occurrence of other words, we were able to do a lot of, uh, what's called natural language processing. We're sh- looking at text and saying, "This is what it means. This is what's going on." But then in the two thousand and seventeen paper, they actually took, uh, a next step, which was the insight that where exactly the thing that you're focusing on was in a sentence, what was before and after, the actual ordering of it mattered, not just the simple co-occurrence. That knowing what position that word was in the sentence actually made a-- the difference. That paper showed the performance went way up in terms of recognition. And that transformer architecture came from that paper, made it clear for a number of researchers, not me, that if you scaled that, uh, transformer architecture up to a larger model so that the position and the position dependence and this vector was learned across many, many more texts, like the whole internet, and you could ask-- you could train it to do various tasks. So this, this transformer model, which is called the pre-trained model, so s-- I apologize because it's very-- I find it very boring to talk about, because unless I'm working with, uh, fellow nerds, this transformer, this pre-trained model, this, um, can think of it as equivalent of an equation with multiple variables. In the case of GPT-4, we think it's about a trillion variables. So that means you have this like an equation where you have a number in front of each variable, a coefficient that's about a trillion long. And this, um, this, um, model can be used for various purposes. One is the chatbot purpose, which is predict, given these-- this sequence of words, what is the next word that's gonna be said? Now, that's not the only thing you could use this model for, but that's turns out to have been the breakthrough application of the transform mo- transformer, transformer model for text.

   13. PAPeter Attia40:54

       So is that, just to round out what you said earlier, Zak, would you say that is the three-- the third thing that enabled this third wave of, of AI? It, it is the, is the, the transformer?

   14. IKIsaac "Zak" Kohane41:04

       Actually, it was, it was not what I was thinking about. For me, I was thinking of the real breakthrough in data-driven AI, I, I put around the two thousand and twelve era. This is yet another-

   15. PAPeter Attia41:14

       Okay

   16. IKIsaac "Zak" Kohane41:14

       ... uh, we, we were all-- If you talked to me in two thousand and eighteen, I would, I would have already told you we're in a new heyday, and everybody was agreeing-- agree- would agree with you. There was a lot of excitement about AI just because of the image recognition capabilities.

   17. PAPeter Attia41:29

       Okay.

   18. IKIsaac "Zak" Kohane41:30

       This was an additional, um, capability that's beyond what many of us were expecting just from the scale-up of the neural network. So the, the three, just to be-- make sure I'm consistent, was large data sets, multi-level neural networks, and, uh, AKA deep neu- neural networks, and the GPU infrastructure.

   19. PAPeter Attia41:54

       Mm-hmm.

   20. IKIsaac "Zak" Kohane41:56

       And that bro-- you know, that brought us well through, um, the, the, uh, you know, two thousand and twelve to two thousand, uh, eighteen. The two thousand and seventeen, um, blip that became what we now know to be this whole, uh, large language model, um, transformer architecture, that's a complete unanticipated for many of us, uh, development. But that was already on the heels of a ascendant AI, uh, AI era. There was already billions of dollars of, uh, of frothy investment in frothy companies, some of which did well, and many of which did not do so well. So I, I think it j- the transformer architecture has revolutionized, um, well, revolutionized many parts of the human condition, I think. But it was already part of it, I think the third wave.

   21. PAPeter Attia42:58

       And

7. 42:59 – 58:16

   AI concerns & regulation: misuse by individuals, military applications, displacement of jobs, & potential existential concerns

   1. PAPeter Attia42:59

      so there's something about GPT where I feel like most people, by the time GPT-3 came out, or certainly by 3.5, this was now outside of the purview of computer scientists, people in the industry who were investing in it. This was now, I mean, becoming as much a verb as Google was in probably the early 2000s, right? Like everybody f- you know, there were clearly people who knew what Google was in '96 and '97, but, you know, by, by 2000, everybody knew what Google was, right? And I think, I, I feel like that, something about GPT 3.5 or 4 was kind of the tipping point where i- I don't think you can not know what it is at this point. And, um, I don't know if that's relevant to this story or, uh, meaning it-- does that sort of speak to what trajectory we're on now? Um, the other thing that I think, Zak, has become so, um, audible in the past year is the elevation in the discussion of how to regulate this thing, which seems like something you would only argue about if you felt that there were a chance for this thing to be harmful to us in some way that we do not yet perceive. So what can you say about that? Because that's obviously a nod to the technical evolution of AI, um, that, you know, very serious people are having discussions about pausing moratoriums, regulations, things like that. I, I can't-- you know, there was no public discussion of that in the '80s, right? Which may have spoke-

   2. IKIsaac "Zak" Kohane44:49

      Right

   3. PAPeter Attia44:49

      ... to the fact that in the '80s, it just wasn't powerful enough to pose a threat. So, um, can, can you maybe give us a sense of what people are debating now? What, uh, what, what, what is the smart, sensible, reasonable argument on both sides of this? And let's just have you decide what the two sides are. I'm, I'm assuming one side says, "Pedal to the metal. Let's go forth on development. Don't regulate this. Let's just go nuts." The other side is, "No, we need to have some brakes and barriers."

   4. IKIsaac "Zak" Kohane45:21

      Well, it's actually, uh, not quite that. So, um, you're absolutely right that, uh, the, you know, chatbots have now become a, a, a commonly used noun, and that probably happened with the emergence of, uh, GPT-3.5 and, uh, and it's, uh, in-- that appeared around, I think, December of 2022. But now, yes, because, um, out of the box, that pre-trained model I told you about could tell you th- things like, um, "How do I kill myself?" Um, "How do I manufacture a, uh, a toxin?" Um, it could allow you to do a lot of, uh, harmful things, so there was that level of concern, and we can tell about what-- we can talk about what's been done about those, uh, first order efforts. Then there's been a group of, uh, scientists who interestingly went from, uh, from saying, "We'll never actually get general intelligence from this particular architecture," to saying, "Oh my gosh, this, uh, technology is able to inference in a way that I had not anticipated, and now I'm so worried that either because it is malevolent or a- just because it's trying to do something that has bad side effects for humanity, it presents an existential threat." Now, on the other side, I don't believe are, is anybody who's saying, "Let's just do h- go heads down, and let's, um, let's just see how fast we can get to artificial general intelligence," or if they do f- think that, they're not saying it openly.

   5. PAPeter Attia47:12

      Can, can you just define AGI, Zak? I think we've all heard the term, but i- is there, is there a, is there a quasi-accepted definition?

   6. IKIsaac "Zak" Kohane47:21

      No. First of all, there's not, and I, I hate myself for even bringing it up-

   7. PAPeter Attia47:24

      [laughs]

   8. IKIsaac "Zak" Kohane47:24

      ... because it starts-

   9. PAPeter Attia47:25

      I was gonna bring it up before you anyway. You were, you-

   10. IKIsaac "Zak" Kohane47:27

       Yeah. I-

   11. PAPeter Attia47:27

       It was inevitable

   12. IKIsaac "Zak" Kohane47:28

       ... yeah, that's a, that was an unfortunate slip because, um, ar- you know, artificial general intelligence means a lot of things to a lot of people, and it-- and I slipped because I think it's, um, again, a moving target, and it's very much eye in the, of the eye in the beholder. So let's be-- you know, there's a guy called Eliezer Lut- Yudkowsky, one of the so-called doomers, and he comes up with great scenarios of how a sufficiently, uh, intelligent system could figure out how to persuade human beings to do bad things or pa- control of our, uh, infrastructure to, you know, bring down our, uh, our, uh, communications infrastructure or airplanes out of the sky. And we can talk about whether that's relevant or not. And on the other side, we have, uh, let's say, OpenAI and Google. But what's fascinating to me is that OpenAI, OpenAI, which working with Microsoft, generated, uh, GPT-4, weren't, were not saying publicly at all, uh, "Let's not regulate it." In fact, they were saying, "Please regulate me." Um- Uh, Sam Altman went on a world tour where he said, "We should be very concerned about this, and we should regulate AI." And he was before Congress saying, "We should regulate, uh, uh, AI." And so I feel f- a bit churlish about saying this because Sam was c- uh, kind enough to write a, um, forward to the book I wrote with Peter Lee and Harry Goldberg on, uh, GPT-4 and the revolution in medicine. But I was wondering, why were they insisting so much on r- regulation? And there's two interpretations. One is it's, uh, just a sincere, and it could very well, well be that, sincere wish that it be regulation- regulated so we check these machines, these programs, to make sure they don't actually do anything harmful. The other possibility, unfortunately, is something called regulatory lock-in, which means I'm going to-- I'm a very well-funded company, and we're gonna create regulations with Congress about what, uh, these... What is required, which boxes do you have to check in order to be allowed to run. And if you're a small company, you're not gonna have a bevy of lawyers, uh, with big checks to, uh, comply with all the regulatory requirements. And so I don't, you know, I s- I think Sam is pr- I don't know him personally. I imagine he's a very well-motivated individual. But whether it's for the reason of regulatory lock-in or for genuine concern, there has, uh, not been any statements of, "Let's go heads down." They, they do say, "Let's b- let's be regulated." Now, having said that, before you even go with the doomer scenario, I think there are, there are someone, uh, just as potentially evil that we have to worry about, another intelligence, that's human beings, and how do human beings use these great tools. So just as we know for a fact that one of the earliest users of GPT-4 were, uh, high schoolers trying to do their homework and solve, uh, hard, uh, puzzles given to them. We also know that, uh, various parties have been using the amazing, uh, text generation and interactive capabilities of these programs to spread misinformation, to run, uh, uh, chatbots, and there's a variety of malign things that could be, that could be done by th- uh, third parties using those, uh, th- these, um, engines. And I think that's, for me, the clear and present danger today, which is how do individuals decide to use these, these, uh, general purpose, uh, programs. If, you know, if you look at what's going on in the Ukr- Ukraine-Russian war, I see more and more autonomous vehicles flying and carrying weaponry and dropping bombs, and we see in our own military a lot more autonomous drones with greater and greater autonomous capabilities. Those are purpose-built to actually do, um, dangerous things. And, um, a lot of, uh, s- science fiction fans will, you know, refer to Skynet-

   13. PAPeter Attia52:23

       Mm-hmm

   14. IKIsaac "Zak" Kohane52:24

       ... from the, from the Terminator series, but we're literally building it right now.

   15. PAPeter Attia52:29

       A- and Skynet sort of i, in, in The Terminator, Zak, they kind of refer to a moment, I don't remember the year, like nineteen ninety-seven or something, and I think they talk about how Skynet became, quote, "self-aware" or, or something-

   16. IKIsaac "Zak" Kohane52:42

       Self-aware, that's right

   17. PAPeter Attia52:42

       ... self-aware. And somehow when it became self-aware, it just decided to destroy humans. And-

   18. IKIsaac "Zak" Kohane52:48

       They say w- we, we're a threat. That's right.

   19. PAPeter Attia52:50

       Yeah. And w- is, is self-aware, uh, movie speak for AGI? Like, what do you think self-aware means in more technical terms?

   20. IKIsaac "Zak" Kohane52:59

       Well-

   21. PAPeter Attia53:00

       Or is it super intelligence? Like, uh, there's so many terms here, and I don't know what they mean.

   22. IKIsaac "Zak" Kohane53:05

       Okay. So self-awareness means a, a, a process by which the intelligent entity can look back, look inwardly at its own processes and recognize its- itself. Now, that's very hand-wavy-

   23. PAPeter Attia53:23

       [chuckles]

   24. IKIsaac "Zak" Kohane53:23

       But, uh, I'm having a senior moment. The guy who wrote Gödel, Escher, Escher, Bach. Uh, well, let's see, what's his name? Gödel, Gödel, Escher, Bach. Oh yeah, Douglas Hofstadter. Douglas H- Hofstadter has probably done the most, uh, thoughtful and cl- and clear, um, writing about what, uh, self-awareness means. And it, I will not do it justice, but if you wanna r- uh, if you really wanna read a wonderful book that spends a whole book trying to explain it, it's called I Am a Strange Loop.

   25. PAPeter Attia54:14

       Hmm.

   26. IKIsaac "Zak" Kohane54:14

       And in I Am a Strange Loop, he explains how if you have enough processing power and you can represent the, uh, the processes that, if y- you have essentially models of the processes that you, that constitute you, in other words, you're able to look at what you're thinking, you may have some sense of self-awareness. Now, it's, there's a bit of an act of faith on that, and so, uh, many, many, uh, AI researchers don't buy that definition. So there's a diff- there's a difference between self-awareness and, uh, a Actual raw intelligence. You can imagine a super powerful computer that would predict everything that was gonna happen around you and was not aware of itself as an, an entity.

   27. PAPeter Attia55:02

       Right.

   28. IKIsaac "Zak" Kohane55:05

       And you-- it's-- the fact remains you do need to have a minimal level of intelligence to be able to be self-aware. So a fly may not be self-aware. It just goes and finds, uh, the good-smelling, uh, poop, and, you know, does whatever it's programmed to do on that. But dogs have some self-awareness and awareness of, uh, their surroundings. They don't have perfect self-awareness, like they don't recom-- uh, recognize themselves in the mirror, and they'll bark at, bark at that. Birds will recognize themselves in mirrors, and, uh, we are-- we recognize ourselves in, in many, many ways. So there is some correlation between intelligence and self-awareness, but these are not necessarily dependent functions. I feel like we got off the track. I apologize, Peter.

   29. PAPeter Attia55:57

       No, it's okay. I mean, I, I think what I'm-- so what I'm hearing you say is, look, there's, um, there are clear and present dangers associated with current best AI tools in that humans can use them for nefarious purposes. Um, uh, it seems to me that the most scalable example of that is still relatively small in that, you know, i-it's not, it's not, like, existential threat to our species large, correct?

   30. IKIsaac "Zak" Kohane56:27

       Well, yes and no. Um, if I was trying to do gain of function research with a virus-
8. 58:16 – 1:06:14

   How AI is enhancing image-based medical specialties like radiology

   1. IKIsaac "Zak" Kohane58:16

      the public seem to be concerned about it.

   2. PAPeter Attia58:18

      Yeah, no, tha-tha-those are, those are fair points. Let's now talk about that state of play within medicine. So, um, [clears throat] I liked your first example, almost one we take for granted. But, um, you go and get an EKG at the doctor's office, this was true thirty years ago just as it is today, you get a pretty darn good readout, right? It's gonna tell you if you have, uh, you know, uh, an AV block. It's gonna tell you if you have a bundle br-bundle branch block. It's gonna, it's gonna-- put it this way, they read EKGs better than I do [chuckles] .

   3. IKIsaac "Zak" Kohane58:47

      Yeah.

   4. PAPeter Attia58:47

      That's not saying much anymore, but they, but they do. Um, what, what, what was the next area where we could see this? It, it, I mean, it seems to me that radiology is a field of medicine, which is, of course, image pixel-based medicine, that would be the most logical next place to, to see AI do good work. What is the current state of AI in radiology?

   5. IKIsaac "Zak" Kohane59:16

      So in all the visual-based medical s- uh, specialties, um, it looks like AI can do as well as many experts. So it turns out that in recognizing-- So what are the image appreciation subspecialties? Pathology, when you're looking at slices of tissue under the microscope. Radiology, where you're looking at X-rays or MRIs. Dermatology, where you're looking at pictures of the skin. So in all those visual-based, uh, uh, specialties, the computer programs are doing by themselves as well as many, uh, experts. But they're not replacing the doctors because that image recognition process is only part of the, their job. Now, to be fair to your point, in radiology, we already today, before AI, in many hospitals would send X-rays by satellite to Australia or India, where they would be read overnight by s-a doctor or a specially trained person who had never seen the patient, and then the reports, uh, fil-filed, filed back to us. So because they're twelve hours away from us, overnight we'd have the results of these, those reads. And that same kind of function can be done automatically by AI. So that's replacing a certain kind of, uh, doctor.

   6. PAPeter Attia1:01:00

      But, but let me dig into that a little bit more. So let's start with a relatively simple type of image, such as a mammogram or a chest X-ray. So it's a single image. I mean, I guess with a chest X-ray, you'll look at a AP and a lateral, but let, let's-

   7. IKIsaac "Zak" Kohane1:01:13

      Yeah

   8. PAPeter Attia1:01:13

      ... just say you're looking at an AP or a, or a single mammogram. Um, a radiologist will look at that. A radiologist will have clinical information as well, so, uh, they will know why this patient presented, in the case of the chest X-ray, for example, in the ER in the middle of the night. Were they short of breath? Uh, do they have a fever? You know, do they have a previous X-ray I can compare it to? They'll have all sorts of information. Um, w- w- is it-- are we not at the point now where all of that information could be given to the AI to, um, enhance the pretest probability of whatever diagnosis it comes to?

   9. IKIsaac "Zak" Kohane1:01:53

      I am delighted when you say pretest pro- uh, probability. Those, that's, uh, don't talk dirty around me. Um, so that's-

   10. PAPeter Attia1:01:59

       [laughs] Love, love, love my Bayes' theorem over here.

   11. IKIsaac "Zak" Kohane1:02:02

       Yeah. Yep. So you just said a lot, 'cause what you just said actually went beyond what the, uh, straight convolutional neural netw- networks would do. 'Cause they actually could not replace, uh, radiologists, 'cause they could not do a good job of taking into account the previous, uh, history of the patient. Um, and it's required the emergence of transformers-

   12. PAPeter Attia1:02:31

       Mm-hmm

   13. IKIsaac "Zak" Kohane1:02:31

       ... where have both multi- multimodality. You have both the image and the text. Now, they're gonna do better than many, many radiologists today. And so there is, I don't think any [laughs] still any threat yet to radiologists as a job. You know, uh, one of the most irritating to, uh, computer science-- uh, to doctors' predictions was by Geoffrey Hinton, the, one of the, uh, intellectuals, leaders of neural network architecture. He said, I think it was in two thousand sixteen, that in, I have this approximately wrong, but in six years, we would have no need for radiologists, and that was just clearly wrong. And the reason it was wrong is, A, they did not have the, at least the capabilities that we just talked about, about understanding about the clinical context, but it's also the fact that we just don't have enough radiologists.

   14. PAPeter Attia1:03:25

       Meaning to do the training?

   15. IKIsaac "Zak" Kohane1:03:27

       To actually do the, to, to actually do the work. Um, so if you look at American medicine, and this is a big, uh... I'll let you shut me down, but there's, if you look at residency programs, we're not getting enough radiologists out. Now, we can, we get a, we have an overabundance of applicants for interventional radiology. They're making a lot of money, it's high prestige. But straight up, uh, radiology readers, not enough of them. Primary care doctors, uh, I go around medical schools and ask who's a, who's get-- who's becoming a primary care doctor. Almost nobody. And so primary care is disappearing in the United States. It's, uh, if you-- there, in fact, Mass General and Brigham announced officially they're not seeing primary care patients. And people are still going to dermatology, um, and they're still going to plastic surgery, but there's al- if, what I did, pediatric endocrinology, half of the slots nationally are not being filled. Pediatric, uh, developmental disorders like autism, those slots, half of them filled. PDID, there's a huge gap em- emerging in the available expertise. So it's not what we thought it was going to be, uh, that we had a, a surplus of doctors that had, that had to be replaced. It's just we have a surplus in a few focused areas which are very popular. And then for all the work of primary care, primary prevention, the stuff, kind of stuff that you're interested in, we have almost no doctors, uh, available.

   16. PAPeter Attia1:05:15

       Yeah, let's go back to the radiologist for a second because-

   17. IKIsaac "Zak" Kohane1:05:17

       Yeah

   18. PAPeter Attia1:05:18

       ... again, I'm fixated on this one because it seems like the most, um, w- well, the closest one to address. And again, if you're saying, look, we have a dearth of imaging radiologists who are able to work the emergency rooms, urgent care clinics, and hospitals-

   19. IKIsaac "Zak" Kohane1:05:35

       Ah

   20. PAPeter Attia1:05:35

       ... wouldn't that be the first place we would want to apply our best of imaging recognition with our super powerful GPUs, and now plug them into our transformers with our language models so that I can get, uh, clinical history, medical past history, previous images, current images, and I don't have to send it to a radiologist in Australia to read it, who then has to send it back to a radiologist here to che- Like, we-- I mean, if we're just trying to fill a gap, that gap should be fillable, shouldn't it?

   21. IKIsaac "Zak" Kohane1:06:10

       It is, and it, and, and that's exactly where it is being filled. And what

9. 1:06:14 – 1:15:05

   The use of AI by patients & doctors

   1. IKIsaac "Zak" Kohane1:06:14

      keeps distracting me in this conversation is that there's a whole other group of users of these AIs that we're not talking about, which is h- the patients.

   2. PAPeter Attia1:06:25

      Mm.

   3. IKIsaac "Zak" Kohane1:06:26

      And previously, none of these tools were available to patients. With the release of GPT three point five and four, and now Gemini and, um, Claude 3, they're being used by patients all the time in ways that we had not anticipated. And let me, let me give you an example. So there's a child who, um, w- was having trouble walking, having trouble chewing, and then started having intractable headaches. Mom so, uh, brought him to multiple doctors. They did multiple imaging studies, no diagnosis. Kept on being in, in intractable pain. She just typed into GPT-4 all the, the reports and asked GPT-4, "What's the diagnosis?" And GPT-4 said, "Tethered cord syndrome." She then went with all the imaging studies to a neurosurgeon and said, "What is this?" He looked at it, he said, "Tethered cord syndrome." And we have such a epidemic of misdiagnosis and undiagnosed patients. Part of my background that, uh, I'll just mention briefly, I'm the principal investigator of something, of the coordinating center of something called the Undiagnosed Network. It's a network with 12 academic hospitals going, uh, down the West Coast from University of Washington, Stanford, UCLA, to Baylor, up the East Coast, Harvard hospitals, NIH. And we see a few thousand patients every year, and these are patients who have been undiagnosed, and they're in pain. And that's a, just a small fraction of those who are undiagnosed. And yes, we bring to bear a whole bunch of, uh, computational techniques and genomic sequencing to actually be able to, um, help these individuals, but it's very clear that there's a much larger burden out there of misdiagnosed individuals. And because of that-

   4. PAPeter Attia1:08:27

      S- s- so quest-- but a question for you, Zak, which is, um, does it surprise you that in that example, the mother was the one that went to GPT-4 and inputted that? I mean, she had presumably been to many physicians along the way. Were you surprised that one of the physicians along the way hadn't been the one to say, "Gee, I don't know, but I got-- Let, let's see what this GPT-4 thing can do"?

   5. IKIsaac "Zak" Kohane1:08:52

      Most clinicians I know do not have what I used to call the Google reflex. I remember when I was on the wards with a bunch of-- and we had a child with dysmorphology. They look different. And I said to the, uh, fellows, this is after residency, "What is the diagnosis?" And they said, "I don't know. I don't know." I said, "He has this and this and this finding. What's the diagnosis?" And they said, and I said, "How would you find out?" They had no idea. And then I just said, "Let's take what I just said and type it into Google." And i- in the top three, uh, uh, re- uh, responses, there was the diagnosis. And that reflex, which they do use in their out, in their civilian life-

   6. PAPeter Attia1:09:38

      Hmm

   7. IKIsaac "Zak" Kohane1:09:38

      ... they did not have in the clinic. And doctors are in a very unhappy position these days. They're really being driven very, very hard, and they're being told to use certain technological tools. They're being turned into data entry clerks. They don't have the Google reflex. Uh, they don't have the reflex, let-- uh, who has the time to look up a journal article, let alone, and they don't do the Google, uh, uh, reflex. And if they, if they [laughs] and even less do they have the, um, let's look at the patient's, uh, uh, history and see what, what GPT-4 would come up, would come up with. I was gratified to see early on doctors saying, "Wow, look, I just took the patient history, plugged it into GPT-4, and said, 'Write me a letter of pre, of prior authorization.'" And they were actually, uh, tweeting about doing this, which on the one hand, I was very, very pleased for them because it was saving them five minutes to write that letter to the insurance company saying, "Please authorize my patient for this procedure." I was not pleased for them because if you use, uh, ChatGPT, you're using a program that is covered by OpenAI as a user, as opposed to a version of GPT-4 that is being run on protected Google, on protected Azure cloud by Microsoft, which is HIPAA covered. For those of you, your, uh, audience doesn't know, HIPAA is the, um, legal framework on which, under which we protect patient privacy, and if you violate it, you can be fined and even go to prison.

   8. PAPeter Attia1:11:23

      So in other words, if a physician wants to put any information into, uh, GPT-4, they better not identify it.

   9. IKIsaac "Zak" Kohane1:11:32

      That's right. So they just plunked in a patient note-

   10. PAPeter Attia1:11:34

       Yeah

   11. IKIsaac "Zak" Kohane1:11:34

       ... into ChatGPT. That's a HIPAA violation.

   12. PAPeter Attia1:11:37

       Right.

   13. IKIsaac "Zak" Kohane1:11:38

       If there's a Microsoft version of it, which is, which is HIPAA compliant, but it's not. Anyway, so these pa-

   14. PAPeter Attia1:11:44

       Yeah

   15. IKIsaac "Zak" Kohane1:11:44

       ... so these doctors are tweeting about putting, uh, so they were using it for, to improve their lives.

   16. PAPeter Attia1:11:49

       Yeah.

   17. IKIsaac "Zak" Kohane1:11:49

       The doctors were using it for improving the business, the administrative part of, of healthcare, which is incredibly m- incredibly important. But by and large, only a few doctors use it for, um, diagnostic, uh, acumen.

   18. PAPeter Attia1:12:04

       And, and then what about more involved, uh, radiology? So, you know, obviously, a plain film is, you know, one of the more straightforward things to do, although it's far from straightforward, as anybody knows who's, you know, stared at a chest X-ray. Um, but once we start to look at three-dimensional images, such as, you know, cross-sectional images of CT scans, MRIs, or even more complicated images like ultrasound and things of that nature, um, what, what is the current state of the art with respect to AI in the assistance of reading these types of images?

   19. IKIsaac "Zak" Kohane1:12:39

       So that's the very exciting news, which is, remember how I said it was important to have a lot of data. That was worth one of the three-

   20. PAPeter Attia1:12:49

       Yep

   21. IKIsaac "Zak" Kohane1:12:49

       ... ingredients. So all of a sudden, having a lot of data around, for example, echo card, echocardiograms, the ultrasounds of your heart. Normally takes a lot of training to interpret those images correctly. So there was a recent study from the, uh, Echo CLIP group led, I think, out of, uh, UCLA, and they took a million echocardiograms and a million textual reports of those and essentially trained the model both to create those, those embeddings I talked about of the images and of the text.

   22. PAPeter Attia1:13:31

       But this is, this is just to make sure people understand what we're talking about. This is not, here's a picture of a cat, here's a description, cat. The-- when you put an image in, you're putting a video in now. You're putting a multidimensional video because you have timescale, you have Doppler effects, you have, uh-- This is a very complicated video that is going in.

   23. IKIsaac "Zak" Kohane1:13:57

       It's a very complicated video, and it's three-dimensional, and it's weird view-- uh, three-dimensional, uh, views from different angles where you-

   24. PAPeter Attia1:14:05

       And it's dependent on the user.

   25. IKIsaac "Zak" Kohane1:14:08

       That as well.

   26. PAPeter Attia1:14:08

       In other words, the tech, the radiology tech can be good or bad. If I was the one doing it, it would be awful, right? If, if you have s- yeah.

   27. IKIsaac "Zak" Kohane1:14:18

       Here's the, here's the thing. So radiolog-- so the echocardiog-- the, uh, echo tech does not have medical school debt. They don't have to go to medical school. They don't have to learn calculus. They don't have to learn physical chemistry, all the hoops that you have to go through in medical school. They don't have the attitudinal debt of doctors. Um, and so in two years, they get all those skills, and they actually do a pretty good job.

   28. PAPeter Attia1:14:41

       No, no, they, they do a fantastic job. But my point is their skill is very much an important determinant of the quality of the image.

   29. IKIsaac "Zak" Kohane1:14:49

       So yes. But what we still require these days is a cardiologist to then read it and interpret it.

   30. PAPeter Attia1:14:56

       Right. That's my-- that, that's sort of where I'm going, by the way, is we're gonna get rid of the cardiologist before we get rid of the technician.
10. 1:15:05 – 1:21:11

    The potential for AI to augment clinicians & address physician shortages

    1. IKIsaac "Zak" Kohane1:15:06

       Let's-- My, my target in this conversation is nurse practitioners and physician assistants with these tools can replace a lot of expert clinicians. And there is a big open question, what is the real job for doctors in ten years from now? And I don't think we, we know the answer to that because you've, you've fast-forwarded the conversation just now.

    2. PAPeter Attia1:15:34

       Well, let's, let's think about it, right? So, so [clears throat] we still haven't come to proceduralists, so we still have to talk about the interventional radiologist, the interventional cardiologist, and the surgeon.

    3. IKIsaac "Zak" Kohane1:15:44

       Right.

    4. PAPeter Attia1:15:44

       And we, we can talk about the role of the surgeon and the da Vinci robot in a moment.

    5. IKIsaac "Zak" Kohane1:15:49

       Yeah.

    6. PAPeter Attia1:15:49

       But I think what we're doing is we're, we're kind of identifying the pecking order of, uh, of physicians and saying that-- a-and let's not even think about it through the lens of replacement. Let's start with the lens of augmentation, which is-

    7. IKIsaac "Zak" Kohane1:16:02

       Right

    8. PAPeter Attia1:16:02

       ... the radiologist can be the most easily augmented, uh, the pathologist, the dermatologist, the cardiologist who's looking at echoes and EKGs and stress tests. Uh, people who are interpreting visual data and using visual data will be the most easily augmented. The second tranche of that will be people who are interpreting language data plus visual data. So now we're talking about your internist, your pediatrician, where you have to interpret symptoms and combine them with laboratory values and combine it with a story and an image. Is that, is that a fair assessment in terms of tier?

    9. IKIsaac "Zak" Kohane1:16:42

       I think it's a, it's ab-ab-absolutely a fair assessment. My only quibble, it's not a quibble, is, and I'm gonna keep on going back to this, is in a place where we don't have primary care, which I'm claiming is increasingly-- The AAMC, the American Academy of, of Medical-- American Association of Medical Colleges estimates that in, in thir-- by thirty-five, that's only, uh, uh, eleven years from now, we'll be missing on the order of fifty thousand primary care doctors. As I told you, I can't get primary care at the Brigham or, uh, at MGH today. And in the absence of that, you have to ask yourself, how can we replace, uh, these absent primary care practi-practitioners with nurse practitioners, with, uh, physician assistants augmented by these AIs? 'Cause there's literally no doctor to replace.

    10. PAPeter Attia1:17:33

        So, so tell me, Zak, where are we technologically on that augmentation? Are we-- Which i-if, if we-- If NVIDIA never came out with another chip, if they literally said, "You know what? We are only interested in building golf simulators, and we're done with the progress of this, and, um, this is as good as it's gonna get," um, do we have good enough GPUs, good enough multi-layer neural networks that all you need is more data and training sets that we could now do the augmentation that has been described by us in the last five minutes?

    11. IKIsaac "Zak" Kohane1:18:10

        Yes. The short answer is yes. Let me get, make it be very concrete. Most concierge services cost, in Boston, somewhere between five and twenty thousand dollars a year. You can get this l- uh, very low cost, uh, concierge service that I'm just amazed that have, have, have not done the following, called One Medical. One Medical was acquired by Amazon, and they have a lot of nurse practitioners, um, uh, in there, and you can make an appointment, you can text with them. I believe that those individuals could order, uh, could be helped in ordering the right imaging studies, the right EKGs, the right, uh, medications, and, uh, assess your continuing, um, heart failure and only decide in a very few cases that you need to see a specialist cardiologist. Or a specialist endocrinologist today without-- It would just be a matter of just making the current models better, evaluating them, 'cause not all models are equal. And we, uh, a big question for us, this is a regulatory question, which is: Which ones do a better job? And they're not all equal. But I don't think we need technological break-breakthroughs to just make the current set of, uh, paraprofessionals work at the level of entry-level doctors. Let me quickly say, uh, the old very bad joke. What do you call the, uh, doct-- the medical student who graduates at the bottom of his class?

    12. PAPeter Attia1:19:51

        A doctor. [laughs]

    13. IKIsaac "Zak" Kohane1:19:52

        Doctor. And so if you could just merely get the t-- bottom fifty percent of doctors to be as good as the top fifty percent, that'd be transformative for healthcare. Now, there are other superhuman capabilities that we will-- that we can go towards, and we can talk about if we want, that do require the next generation of algorithms, NVIDIA, um, architectures, and data sets. But there's so much. If everything stopped now, we could already transform medicine. It's just a matter of the sweat equity to create the models, figure out how to include them in the workflow, how to pay for them, um, how to create a reimbursement system that, uh, and a business model that, uh, works for our society. But there's no technological, um, barrier.

    14. PAPeter Attia1:20:48

        So in my mind, there are basically-- everything we've talked about is take the best case example of medicine today and [clears throat] augment it with AI such as-- such that you can make-- you can raise everyone's level of care to that of the best, no gaps, and it's scaled out.

    15. IKIsaac "Zak" Kohane1:21:10

        Exactly.

11. 1:21:11 – 1:32:12

    The potential for AI to revolutionize early diagnosis & prediction of diseases: Alzheimer’s disease, CVD, autism, & more

    1. PAPeter Attia1:21:11

       Okay. Now let's talk about another problem, which is: Where do you see the potential for AI in solving problems that we can't even solve on the best day at the best hospitals with the best doctors? So let me give you an example. We can't really diagnose Alzheimer's disease until it appears to be at a point that, for all intents and purposes, is irreversible. Um, maybe on a good day, we can halt progression really, really early in a patient with just a whiff of MCI, mild cognitive impairment, um, maybe with an early amyloid detection and an anti-amyloid drug. But do you-- I mean, is it science fiction to imagine that there will be a day when an AI could listen to a person's voice, watch the movements of their eyes, study the movements or-- of their gait, and predict twenty years in advance when a person is staring down the barrel of a neurodegenerative disease and act at a time when maybe we could actually reverse it? I mean, how science fictiony is that?

    2. IKIsaac "Zak" Kohane1:22:24

       I don't believe it's science fiction at all. Do you know that looking at retinas today, uh, images of retina, s-straightforward convolutional neural net- neural networks, not even ones that involve transformers, can already tell you by just looking at your retina, not just whether you have retinal disease, but if you have hypertension, if you're a male, if you're female, how old you are, and some estimate of your longevity. And that's just looking at the back of your eye-

    3. PAPeter Attia1:22:54

       Mm-hmm

    4. IKIsaac "Zak" Kohane1:22:54

       ... and has seen enough, uh, data. When you-- When-- [sighs] I'm-- I did a study. I was, uh, I was a small player in a study that appeared in Nature in two thousand and five with, uh, Bruce Yankner. We were looking at, uh, frontal lobes of individuals, uh, who had died for a variety of reasons, often accidents, um, of various ages. And we saw, bad news for people like me, that after age forty, your transcriptome, the genes that are switched on, fell off a cliff. Like thirty percent of your transcriptome went down. And, um, and so there seemed to be a big difference in the expression of genes around age forty. And there was-- But there was one ninety-year-old who looked like the, the young guy, so maybe there's hope for some of us.

    5. PAPeter Attia1:23:49

       Mm-hmm.

    6. IKIsaac "Zak" Kohane1:23:51

       But, but then I thought about it afterwards, and there were o-other things that actually had much smoother functions, which are-- don't have s-- quite the falloff, like our skin. So our skin ages, and in fact, all our organs age, and they age at different rates.

    7. PAPeter Attia1:24:11

       You're, you're saying that the transcriptome of the skin, you did not see this cliff-like effect at a given age the way you saw it in the frontal cortex. Okay.

    8. IKIsaac "Zak" Kohane1:24:19

       That's right. And so different organs age at different, at different rates. Uh, but having the, the right data sets, um, and the ability to see nuances that we don't notice makes it very clear to me that, um, the early detection part, no problem. That's, uh, that's, is gonna be very straightforward. The treatment part, we can, we can talk about it as well. But, um, again, um, we had early on from the, uh, very famous Framingham Heart Study, a predictor of when you had-- going to have heart, uh, disease based on just a handful of variables. Now we have these artificial intelligence models that, based on hundreds of, of variables, can predict various other diseases, and it will do Alzheimer's, I believe, uh, very soon. I think you'll be able to see a combination of gait, speech patterns, um, picture of your body, picture of skin, and eye movements, like you said, will be a very accurate predictor. Um, I just-- we just published, by the way, recently, something about eyes. A very nice, uh, study, uh, where in a car, just by looking at the, at the driver, it can figure out what your blood sugar is. Because diabetics previously have not been able to get driver licenses sometimes because of the worry about them passing out because of hypoglycemia. So there was a very nice study that showed that you could just by looking, have ca-cameras pointed at the eyes-

    9. PAPeter Attia1:26:10

       Hmm

    10. IKIsaac "Zak" Kohane1:26:10

        ... to actually figure out exactly what the blood sugar is. So th-that's-- that kind of detection is, I think, fairly straightforward. It's a different question about what you can do about it.

    11. PAPeter Attia1:26:21

        Before we go to the what you can do about it, I just wanna go a little deeper on the, on the predictive side. Um-

    12. IKIsaac "Zak" Kohane1:26:26

        Yes.

    13. PAPeter Attia1:26:27

        You, you brought up the Framingham model or the-

    14. IKIsaac "Zak" Kohane1:26:29

        Yeah

    15. PAPeter Attia1:26:29

        ... Multi-Ethnic Study on Atherosclerosis, the MESA model. These are the two most popular models by far for, for looking at major adverse cardiac event, a major adverse cardiac event risk prediction. Um, but you needed something else to build those models, which was enough time to see the outcome, right? So you had to, in the Framingham cohort, which was the late '70s and early '80s, you then had the Framingham offspring, offspring cohort, and then you had to be able to follow these people with their LDL-C and HDL-C and triglycerides, and later, eventually, they incorporated calcium scores. So if today we said, "Look, we want to be able to predict thirty-year mortality," which is something no model can do today. This is a big pet peeve of mine, is we generally talk about cardiovascular disease through the lens of ten-year risk, which I think is ridiculous. We should talk about lifetime risk.

    16. IKIsaac "Zak" Kohane1:27:22

        Yep.

    17. PAPeter Attia1:27:22

        But I would settle for thirty-year risk, frankly. And if we had a thirty-year risk model where we could take many more inputs, and it-- I would absolutely love to be looking at the retina. I, I believe, by the way, Zak, that retinal examination should be a part of medicine today for everybody. I, I would take a retinal exam over a hemoglobin A1C all day, every day. I'd never look at another A1C again if I could see the retina of every one of my patients. Um, but my point is, even if effective today, we, we could define the dataset, and let's overdo it, and we can prune things later, but we wanna see these fifty things in everybody to predict every disease. H- Is there any way to get around the fact that we're gonna need thirty years to see this come to fruition in terms of watching how the story plays out? Or are we basically gonna say, "No, you're, you know, we're gonna do this over five years, but it will only-- you know, it won't be that useful 'cause a five-year predictor basically means you're already catching people in the throes of the disease"?

    18. IKIsaac "Zak" Kohane1:28:23

        I'll say three words: electronic health records. So that turns out not to be the answer in the United States. Why? Because in the United States, we move around. We don't stay in any given healthcare system that long. So very rarely will I have all the measurements made on you, Peter, uh, all your, all your g-glycohemoglobins, all your blood pressures, all your clinic visits, all the imaging studies that you've had. However, that's not the case in Israel, for example. In Israel, they have the, these HMOs, health maintenance organizations, and one of them, Clalit, I have a, a good relationship with because they published all the big COVID studies looking at the efficacy of, uh, the vaccine. And why could they do that? 'Cause they had the whole population, uh, available. And they have about twenty, twenty-five years worth of data on all their patients, and in detail, and famil-family relationships. So if you have that kind of data, and Kaiser Permanente also has that kind of data, I think you can actually come close. Now-

    19. PAPeter Attia1:29:39

        But you're not gonna be able to get retina, gait, voice. Uh, so, so it's-

    20. IKIsaac "Zak" Kohane1:29:44

        No.

    21. PAPeter Attia1:29:44

        Yeah. Because we still have to get those prospectively.

    22. IKIsaac "Zak" Kohane1:29:48

        Those you still have to get, but I'm gonna claim that there are proxies, rough proxies, but for gait, falls. Um-

    23. PAPeter Attia1:29:57

        Yeah

    24. IKIsaac "Zak" Kohane1:29:58

        ... and for, um, um, you know, hearing problems, uh, visits to the audi-audiologist. Now, these are noisier measurements.

    25. PAPeter Attia1:30:07

        Yeah.

    26. IKIsaac "Zak" Kohane1:30:08

        And so the-- so those of us who are, uh, da-data junkies like I am always keep m-um, mumbling to ourselves, "Perfect is the enemy of good." And so-

    27. PAPeter Attia1:30:20

        [laughs]

    28. IKIsaac "Zak" Kohane1:30:21

        Um, waiting thirty years to have the perfect dataset is not the right answer to help patients now. And there are things that we could know now that we-- that are knowable today that we just don't know because we haven't bothered to look. Give you a quick example. I, I did a study of, um, autism using electronic health records maybe fifteen years ago, and I saw there was a lot of GI problems. And I talked to a, a, a pediatric expert, and they said, it was a little bit dismissive, they said, "Brain bad, tummy hurt." I said, "But I'm seeing a lot of inflammatory bowel disease," like things that, you know, I, I-- just doesn't make sense to me that this is somehow effect of brain function. To make a long story short, we did a massive study. We're looking for tens of thousands of individuals, and sure enough, we found subgroups of patients who had immunological problems associated with their autism, and they had type one diabetes, inflammatory bowel disease, lots of infections. Those were knowable, but they were not known. And I had Frankly, parents come to me more thankful that for anything else I had ever done for them clinically, 'cause I was telling that these parents, telling these parents they weren't hallucinating, that these kids had these problems. They just weren't being recognized by medicine because no one had the, the big wide angle to see these trends. So, uh, without knowing the field of Alzheimer's the way I do other fields, I bet you there are trends in Alzheimer's that are, you can pick up today by looking at enough patients-

    29. PAPeter Attia1:31:56

        Yeah

    30. IKIsaac "Zak" Kohane1:31:56

        ... that you'll find some that have more frontotemporal components-
12. 1:32:12 – 1:41:55

    The future of AI in healthcare: integration of patient data, improved diagnostics, & the challenges of data accessibility & regulatory compliance

    1. PAPeter Attia1:32:12

       you've already alluded to the fact that, um, we're dealing with a customer, if the physician is the customer, who is not necessarily the most tech-forward customer. Um, and truthfully, like many customers of AI, runs the risk of being marginalized by the technology if the technology gets good enough, and yet you need the customer to access the patient to make the data system better, to make the training set better. So how do you see the interplay over the next decade of that, that dynamic?

    2. IKIsaac "Zak" Kohane1:32:52

       So that's really a, that's the right question 'cause you need, in order for these AI models to work, you need a lot of data on a lot of, on a lot of patients. Where is that data gonna come from? So there are some healthcare systems which, like Mayo, the Mayo, uh, Institute, uh, who think they can get enough data in that fashion. Um, there are some data companies that are trying to get relationships with healthcare systems where they can get de-identified data. I myself will, I w- I'm betting on something else. There is a trend where consumers are gonna have increased access to their own data. The 21st Century Cures Act was passed by Congress, and it said that patients should be given access to their own data programmatically. Now, they're not expecting your grandmother to, uh, write a program to access the data programmatically, but by having a right to it, it enables others to do so. So for example, Apple has something called Apple Health, and it has this big heart icon. And if you're one of the 800 hospitals that they've already hooked up with, like Mass General or Brigham, Women's, and you're a patient there, you c- if you authenticate yourself to it, if you give it your username and password, it will download into your iPhone, your labs, your meds, your diagnoses, your, um, uh, procedures, as well as all the, um, wearable stuff that, your blood pressure that you get as an outpatient, um, and, uh, various other, uh, forms of data. That's already happening now. There's not a lot of companies that are taking advantage of that, but that's a-- Right now, that data is available on tens of millions of Americans.

    3. PAPeter Attia1:34:44

       Now, isn't it interesting, Zak, how unfriendly that data is in its current form? Like, I'll give you just a silly example in our practice, right? So if we send a patient to LabCorp or Boston Heart, or pick your favorite lab, and we want to generate our own internal reports based on those, where we want to do some analysis on that and m- lay out trend sheets, we have to u- use our own internal software that it's almost impossible to scrape those data out of the labs, right? 'Cause they're sending you PDF reports, their APIs are garbage. Like, nothing about this is user-friendly. And so, so even if you have the, the My Heart thing or whatever, the My Health thing come on your phone, it's not navigable. It's not searchable. It doesn't show you trends over time. Like, uh, uh, is there, is there a more user hostile industry from a data perspective than the health industry right now?

    4. IKIsaac "Zak" Kohane1:35:45

       No. Um, no, and there's a good reason why, um, 'cause they're keeping you captive. But Peter, the good news is you're speaking to a real nerd.

    5. PAPeter Attia1:35:55

       [laughs]

    6. IKIsaac "Zak" Kohane1:35:56

       And let me tell you two ways where actually we could solve your problem. One, if it's in the Apple Health thing, A, someone can actually write a program, an app on the iPhone, which will take those data as numbers, and not have to scrape it, and it can run it through your own trend, uh, uh, trending programs. You could actually use it directly. Also, Gemini and GPT-4, you can actually give it those PDFs, and actually, with the right prompting, it'll actually take those data and turn them into tabular spreadsheets. It's not that-

    7. PAPeter Attia1:36:32

       But we can't, we can't do that because of HIPAA, correct?

    8. IKIsaac "Zak" Kohane1:36:37

       Yeah, absolutely. Uh, if the patient, if the patient gives them to you, if the patient gets it from the patient portal, absolutely, you can do that.

    9. PAPeter Attia1:36:43

       The patient can do that, but I can't use a patient's data that way.

    10. IKIsaac "Zak" Kohane1:36:47

        If the patient gives it to you, absolutely.

    11. PAPeter Attia1:36:50

        Really?

    12. IKIsaac "Zak" Kohane1:36:51

        Oh, yes.

    13. PAPeter Attia1:36:52

        But it's not de-identified.

    14. IKIsaac "Zak" Kohane1:36:55

        Doesn't matter. What, which p- which part are you getting hooked? I wanna-- There's many, uh, things that you're worried about.

    15. PAPeter Attia1:37:02

        I- if a patient says, "Peter, you can take my 50 LabCorp reports for the last-"

    16. IKIsaac "Zak" Kohane1:37:08

        Yes

    17. PAPeter Attia1:37:08

        "... ten years-"

    18. IKIsaac "Zak" Kohane1:37:09

        Yes

    19. PAPeter Attia1:37:09

        "... and you can run them through ChatGPT to scrape it out and give me an Excel spreadsheet that will perfectly tabularize everything that we can then run-"

    20. IKIsaac "Zak" Kohane1:37:18

        Yes

    21. PAPeter Attia1:37:18

        "... into our model to build trends and look for things," I didn't, I didn't think that was doable, actually.

    22. IKIsaac "Zak" Kohane1:37:25

        So it's not doable through ChatGPT 'cause your lawyers would say, "Peter, you're gonna get, um, A, a million dollars in fines from HIPAA. But if you use-- I don't want-- I'm not a shill for, uh, Microsoft. I don't o-own any stock. But if you do GPT on Azure Cloud, that's HIPAA protected, you absolutely can u-use it s- with, uh, patient consent. Hundred percent you can do it. And GPT is being used with patient data out of Stanford right now. Um, Epic's using GPT-4. Um-

    23. PAPeter Attia1:37:59

        That's great

    24. IKIsaac "Zak" Kohane1:37:59

        ... and it's absolutely, uh, legitimately, uh, usable by you. That's-- People don't understand that.

    25. PAPeter Attia1:38:06

        So we've, we've now just totally bypassed OCRs. Like, we do not need to waste our time with optical-- for people not in the acronyms, optical character recognition, which is fifteen years ago, what we were trying to do to scrape this data.

    26. IKIsaac "Zak" Kohane1:38:20

        Peter, let me tell you, there's-- New England Journal of Medicine, I'm on the editorial board there, and we just published, uh, three months ago, a picture of, picture of the week, a back of this seventy-two-year-old, and it looks like a bunch of red marks. Like, to me, it looks like somebody who scratched themselves. And it says blah, blah, blah, they had trouble sleeping.

    27. PAPeter Attia1:38:40

        This is the image of the week.

    28. IKIsaac "Zak" Kohane1:38:42

        Image of the week.

    29. PAPeter Attia1:38:42

        Yep.

    30. IKIsaac "Zak" Kohane1:38:43

        And so, um, and I took that whole thing, and I took out one important fact, and then gave it to GPT-4, the, the image and the text. And it came up with the f- the two things it c-thought it would be, either bleomycin toxicity, which I don't know what that looks like, and shiitake mushroom toxicity. What I'd removed is the fact that the guy had eaten mushrooms the day before.
13. 1:41:55 – 1:49:48

    The future of autonomous robotic surgery

    1. PAPeter Attia1:41:55

       move on to the next piece of medicine. So we've gone from purely the recognition, image-based, to how do I combine image with voice, story, text. Uh, you've made a very compelling case that we don't need any more technological breakthroughs to augment those. It's purely a data set problem at this point, and a willingness. Let's now move to the procedural. Is there, in our lifetimes, say, Zak, um, the probability that if you need to have a radical prostatectomy, which currently, by the way, is never done open. This is a procedure that the da Vinci, a robot, has revolutionized. There's no blood loss anymore. This-- When I was a resident, this was one of the bloodiest operations we did. It was the only operation, by the way, for which we had the patients donate their own blood two months ahead of time. That's how guaranteed it was that they were gonna need blood transfusions. So we just said, "To hell with it. Come in a couple months before, give your own blood, 'cause you're gonna need at least two units following this procedure." Today, it's, uh, it, it's insane how successful this operation is, uh, on, on a large part of, you know, the robot. But the surgeon needs to move the robot. Are we getting to the point where that could change?

    2. IKIsaac "Zak" Kohane1:43:17

       So let me tell you where we are today. Today, there's been studies where it's great. They just collected a bunch of YouTube videos of surgery and trained it up one of these, uh, mo- generative models. So it says, "Oh, they're putting on the sc- uh, scalpel to cut this ligament." And by the way, that's too close to the blood vessel. They should move it a little bit to the side. That's already happening. Um, based on what we're seeing with robotics in the general world, I think the da Vinci, controlled by a robot, 10 years is a very safe bet.

    3. PAPeter Attia1:44:03

       Really?

    4. IKIsaac "Zak" Kohane1:44:04

       It's a very safe bet. It's a very safe-

    5. PAPeter Attia1:44:06

       But in some way, in some ways, 10 years is nothing.

    6. IKIsaac "Zak" Kohane1:44:10

       It's nothing, but it's a very, it's a very safe bet. The fact is, right now, I can do a better job, by the way, just to go back to our previous discussion, giving you a genetic diagnosis based on your findings with a, than any primary care provider and interpre- interpreting the ge- uh, genomic test. Now, it can be outperformed by a, a, a-

    7. PAPeter Attia1:44:34

       Right. Uh, oh, and y- so are you using that example, Zak, because it's a huge data problem? Like in other words, that of c- that's obvious that you would be able to do that because the amount of data, I mean, there's three billion base pairs to be analyzed.

    8. IKIsaac "Zak" Kohane1:44:49

       Yeah.

    9. PAPeter Attia1:44:49

       So of course you're gonna do a better job.

    10. IKIsaac "Zak" Kohane1:44:50

        Yeah.

    11. PAPeter Attia1:44:50

        Are you saying-

    12. IKIsaac "Zak" Kohane1:44:51

        By linking, by, by, by linking it to symptoms.

    13. PAPeter Attia1:44:53

        Yeah, yeah. But you're saying-

    14. IKIsaac "Zak" Kohane1:44:55

        Yeah

    15. PAPeter Attia1:44:55

        ... surgery is a data problem because if you turn it into a pixel problem-

    16. IKIsaac "Zak" Kohane1:45:00

        Pixel and mo- you know, movement, um-

    17. PAPeter Attia1:45:02

        And degrees of freedom. Yeah. Infinite. Yeah.

    18. IKIsaac "Zak" Kohane1:45:03

        Degrees of freedom.

    19. PAPeter Attia1:45:04

        Yeah.

    20. IKIsaac "Zak" Kohane1:45:05

        That's it. That's it. And remember, um, there's a lot of degrees of freedom in moving a car around traffic, and by the way, lives are on the line there, too. Now, medicine is not the only job where lives are at stake. Driving a ton of metal at 60 miles per hour in traffic is also putting lives at stake, and last time I looked, there's several manufacturers who are saying that for some appreciable fraction of that effort, they're controlling multiple degrees of freedom with, with a robot.

    21. PAPeter Attia1:45:45

        Yeah, I very recently spoke with somebody, um, I won't name the company, I suppose, but it's one of the companies that's deep in the space of autonomous vehicles, and they very boldly stated, and I, I, they, they made a pretty compelling case for it, that if every vehicle on the road was at their level of technology in autonomous driving, y- y- you just, well, you wouldn't have fatalities anymore. Um, but the key was that every vehicle had to be at that level, right? It, because you, you s- if you still have some people driving and some autonomous, w- do you, do you b- I don't know if you know enough about that field, but do you, does that sense check to you?

    22. IKIsaac "Zak" Kohane1:46:21

        Yes, I do. I do. Well, first of all, I'm a terrible driver, and I am a better driver, it's not for an ad, but the fact is I'm a better driver 'cause I'm on a Tesla, uh, uh, 'cause I'm a terrible driver, and there's actually a, um, a very good s- um, message for medicine because I'll, I will, uh, paraphrase this. I know enough to know that I, I need to jiggle the steering wheel when I'm driving with a Tesla 'cause otherwise it will assume that I'm just zoning out. What I didn't realize is this. So I, I'm very bad. I'll pick up my, uh, phone, and I'll look at it, and it's looking at, I didn't realize it was looking at me, and it says, "Zak, put down the phone." So I, "Oh, okay." I put it, and I pick it up a- then I say, "Okay." Three min- minutes later, I pick it up again, and it says, "Okay, that's it. I'm switching off autopilot." Uh, so it switches off autopilot, and now I have to pay attent- full attention. Then I get home, and it says, "All right. That was bad. You do that four more times, I'm switching off autopilot until the next software update." And the reason I mention that is it takes a certain amount of confidence to do that to your customer base, saying, "I'm switching off the thing that they bought me for." In medicine, how likely is it that we're, we're gonna fall asleep at the wheel if we have an AI thinking for us? I think it's a real, it's a real issue.

    23. PAPeter Attia1:47:41

        Mm-hmm.

    24. IKIsaac "Zak" Kohane1:47:41

        We know for a fact, for example, back in the '90s, that, um, doses for s- a drug like ondansetron, where people would talk endlessly about how frequently it should be given and with what s- dose. The moment you put it in the order entry system, 90, 95% of doctors would just use the default there. And so how in medicine are we gonna keep doctors awake at the wheel, and will we dare to do the kind of challenges that I just described the car doing? So just to get back to it, I do believe, because of what I've seen with, uh, autonomy and robots, that as fancy as we think that is, controlling a da Vinci Robot will probably have less bad outcomes. You know, every once in a while, someone nicks something, and you have to go into full surgery, or they go home, and they, they, uh, [laughs] they, they die on the way home because they exsanguinate. I think it's just gonna be safer.

    25. PAPeter Attia1:48:41

        It, it's just unbelievable for me to wrap my head around that, but-

    26. IKIsaac "Zak" Kohane1:48:47

        It's gonna take more, it's gonna take more-

    27. PAPeter Attia1:48:47

        Tru- truthfully, it's impossible for me to wrap my head around what's already happened, so I, I-

    28. IKIsaac "Zak" Kohane1:48:51

        Yeah. Yeah

    29. PAPeter Attia1:48:51

        ... I, I, I guess I'll try to retain the humility that says, uh, I reserve the right to be startled. Um, again, there are certain things that seem much easier than others. Like I have an easier time believing we're gonna be able to replace interventional cardiologists, where the number of degrees of freedom, the complexity, and the relationship between what the image shows, what the cath shows, and what the input is, the stent, that gap is much narrower. Yeah, I can see a bridge to that. But when you talk about doing a Whipple procedure, when you talk about y- y- you know, what it means to car... you know, cell by cell, take a tumor off the superior mesenteric vessels, I'm thinking, "Oh my God."

    30. IKIsaac "Zak" Kohane1:49:36

        Since we're on record, I'm gonna say, I'm talking about your routine prostate-
14. 1:49:48 – 1:54:11

    AI & the future of mental health care

    1. PAPeter Attia1:49:48

       Let's go one layer further. [chuckles]

    2. IKIsaac "Zak" Kohane1:49:50

       Yeah.

    3. PAPeter Attia1:49:52

       Let's talk about mental health, right? So this is, this is a field of medicine today that I would also argue is grossly underserved, right?

    4. IKIsaac "Zak" Kohane1:50:00

       Yeah.

    5. PAPeter Attia1:50:00

       So everything you've said to date resonates. Um, I, I completely agree from my own experience that the resources in pediatrics and primary care, I mean, uh, these things are, are, are unfortunate at the moment.

    6. IKIsaac "Zak" Kohane1:50:13

       Harvard has, I think, 60% of the, of the med- of the, uh, undergraduates are getting some sort of mental health support, and it's completely outdoing all the resources available to the university health services. And so we have to outsource some of our mental health, and this is a very richly endowed university. In general, we don't have the resources.

    7. PAPeter Attia1:50:34

       Yeah. So, so, so here we live in a world where there's-- I think the evidence is very clear that when, when a person is depressed, when a person is anxious, when a person has any sort of mental or emotional illness, pharmacotherapy, ph- pharmacotherapy plays a role, but it can't displace psychotherapy.

    8. IKIsaac "Zak" Kohane1:50:54

       Of course.

    9. PAPeter Attia1:50:54

       So you, you have to be able to put these two things together. And the data would suggest that the knowledge of your psychotherapist is important, but it's less important than the rapport you can generate with that individual. Now, based on that, do you believe that the most sacred, protected, if you wanna use that term, profession within, within all of medicine will then be psychiatry?

    10. IKIsaac "Zak" Kohane1:51:20

        I'd like to think that, um, and I think-- But here's, here's some, uh-- I'd like to think that. And I'm not gonna ever-- Well, I shouldn't say that. If I had a psychat- psychiatric, uh, GPT speaking to me, I wouldn't think that it understood me. On the other hand, um, back in the 1960s or '70s, uh, there was a program called ELIZA, and it was a simple, uh, pattern matching program. It would just emu- emulate what's called a Rogerian therapist, um, where, "I really hate my mother." "Why do you say you hate your mother?" "Oh, it's because, uh, I don't like the way she fed me." "What is it about the way she fed you?" And just very, very simple pattern matching. And this ELIZA program, which was developed by Joe Weizenbaum at, at MIT, A, his own secretary would lock herself in, in her office to have sessions with this thing 'cause it's non-judgmental and a human being.

    11. PAPeter Attia1:52:31

        And I'm sorry, this was in the '80s?

    12. IKIsaac "Zak" Kohane1:52:33

        '70s or '60s.

    13. PAPeter Attia1:52:35

        Wow.

    14. IKIsaac "Zak" Kohane1:52:36

        Yeah. And, and it turns out there's a wh- that there's a large group of, of patients who actually would rather have a non-human, non-judgmental person who remembers what they've said last time, shows empathy verbally.

    15. PAPeter Attia1:52:53

        Hmm.

    16. IKIsaac "Zak" Kohane1:52:54

        And, you know, again, I wrote this book with Peter Lee, and Peter Lee made a big deal in the book about how GPT-4 was showing empathy. And in the book, I argued with him that this was not that big a deal. And I s- I said, "I remember from medical school being told that some of the most popular doctors are popular 'cause they're very deep empaths, not necessarily the best doctors."

    17. PAPeter Attia1:53:22

        Right. House is your great example in TV that-

    18. IKIsaac "Zak" Kohane1:53:24

        Exactly

    19. PAPeter Attia1:53:24

        ... you know. Yeah.

    20. IKIsaac "Zak" Kohane1:53:25

        And so I said, you know, for certain things, I might actually want-- But that's just me. And I think that for, um-- I could imagine a lot of, for example, with cognitive behavioral therapy being done and be found acceptable by a subset of human beings.

    21. PAPeter Attia1:53:46

        Hmm. Yeah, you might be right.

    22. IKIsaac "Zak" Kohane1:53:50

        It's just, it's not-- it wouldn't be for me, 'cause I'd say I'm just speaking to some stupid program. But if it's giving ins- you insight into yourself, and it, it's based on the wisdom culled from millions of, uh, patients, who's to say that it's worse? And it's certainly not judgmental, and maybe a little bit less.

15. 1:54:11 – 1:58:52

    How AI may transform & disrupt the medical industry: new business models & potential resistance from established medical institutions

    1. PAPeter Attia1:54:11

       So, so Zak, you're, you're born probably just after the first AI boom.

    2. IKIsaac "Zak" Kohane1:54:20

       Yeah.

    3. PAPeter Attia1:54:20

       You come of age in, you know, intellectually, academically in the second. And, um, and now in the mature part of your career, when you're at the height of your esteem, you're riding the wave of this third version, which I, I don't think anybody would argue is, is going anywhere.

    4. IKIsaac "Zak" Kohane1:54:43

       Yeah.

    5. PAPeter Attia1:54:43

       Um, as you look out over the next decade, and we'll start with medicine, what are you most excited about, and what are you most afraid of with respect to AI?

    6. IKIsaac "Zak" Kohane1:54:57

       So with re- with re- specifically with regard to medicine, what I'm most concerned about is how it could be used by the medical establishment to keep things the way they are, to pour con- pour concrete over practices.

    7. PAPeter Attia1:55:15

       Hmm.

    8. IKIsaac "Zak" Kohane1:55:15

       And what I'm most excited about is alternative business models, young doctors who create businesses outside the mold of hospitals. Hospitals are these very, very complex entities. They ha-- make billions of dollars, some of the bigger ones, b- but with very small margins, one to two percent. And when you make, have huge revenue but very small margins, you're gonna be very risk averse. And you're not gonna want to change. And so what I'm mo- excited about is, um, the opportunity for new businesses and n- and new ways of delivering to patients insights that are data-driven. What I'm worried about is hospitals u- doing a bunch of information blo- blocking, um, and regulations that will make it harder for these new businesses to get creative because they don't wanna be... understandably, they don't, they don't wanna be disrupted, and that's the danger.

    9. PAPeter Attia1:56:24

       I- in that latter case or that, that case that you're afraid of, Zak, can patients themselves work around the hospitals with these new companies, these disruptive companies, and say, "Look, we have the legal framework that says I own my data as a patient. I own my data," so I know the h- believe me, we know this in our practice, just because our patients own the data doesn't make it easy to get. I- there is no aspect of my practice that is more miserable and more inefficient than data acquisition from hospitals. Uh, it's, it's, it's actually comical. It's actually comical.

    10. IKIsaac "Zak" Kohane1:57:00

        It's absolutely comical. I'm willing to p- and I do pay hundreds of dollars to get my data for my patients with rare d- and unknown diseases in this network extracted from the hospitals because it's worth it to pay someone to, to do that extraction. Yeah, but now I'm telling you it is doable.

    11. PAPeter Attia1:57:19

        Yeah. So you're saying because of that, um, are you confident that, that the, the, the legal framework for patients to have their data coupled with AI and companies, do, do you think that that will be a sufficient hedge against your, your biggest fear?

    12. IKIsaac "Zak" Kohane1:57:35

        I think that, um, there will, that this is not unlike my ten-year prostatectomy by robot prediction. I'm not as certain, but I would give better than fifty percent odds that in the next ten years there'll be a company that, at least one company, that figures out how to use that patient's right to access through dirty APIs-

    13. PAPeter Attia1:58:01

        Yep

    14. IKIsaac "Zak" Kohane1:58:02

        ... AI to clean it up, provide decision support with hu- human doctors or paraprofessi- uh, health paraprofessionals to create, uh, alternative business, I'm convinced because just the, the, um, demand is there. And, um, and I think that you'll see companies that are even willing to put themselves at risk. What I mean by that, are willing to take the medical risk on that if they do better than a certain level of performance, they get paid more, or if they do worse-

    15. PAPeter Attia1:58:35

        They don't get paid

    16. IKIsaac "Zak" Kohane1:58:36

        ... they get [inaudible] .

    17. PAPeter Attia1:58:36

        Yeah.

    18. IKIsaac "Zak" Kohane1:58:38

        I believe there are companies that are gonna be in that space, but that is because I don't want to, uh, underestimate the medical establishment's ability to squish-

    19. PAPeter Attia1:58:47

        [laughs]

    20. IKIsaac "Zak" Kohane1:58:48

        ... um, threats. So we'll see.

16. 1:58:52 – 2:03:22

    Potential positive & negative impacts of AI outside of medicine over the next decade

    1. PAPeter Attia1:58:52

       Okay. Now let's just pivot to AI outside of medicine. Same question. What are you most afraid of, um, over the next decade? Not, so maybe we're not talking about, uh, s- you know, self-awareness and Skynet, but next decade, what are you most afraid of, and what are you most excited about?

    2. IKIsaac "Zak" Kohane1:59:12

       What I'm most afraid of is, um, a lot of the ills of social networks being magnified by use of, um, these AIs to further accelerate the, um, cognitive chaos and, uh, vitriol that, uh, fills our, um, our social experiences on the net, and I think it's co- it could be used to accelerate them. So that's my biggest fear.

    3. PAPeter Attia1:59:48

       I saw an article two weeks ago that, um, was, was an individual, I do- I can't remember if they were currently in or formerly part of the FBI, and they s- stated that they believed, I think it was somewhere between seventy-five and ninety percent of, quote-unquote, "individuals" on social media were not in fact individuals. Um, d- would, would you d- I, I don't know if you spend enough time on social media to, to, to have an, have a point of view on that.

    4. IKIsaac "Zak" Kohane2:00:16

       Unfortunately, I have to, uh, admit to the fact that my daughter, who's now twenty years old, but four years ago she bought me a mug that says on it, "Twitter addict." And, um, and so I spend enough time. Um, I would not be surprised if some large fraction, um, are bots, and it's gonna get, could get worse, and it's gonna be harder to actually distinguish reality from human beings, harder and harder and harder. And, uh, so that's the, that's the, that's the real problem. Um, and that's gonna g- as w- we are fundamentally social animals, and if we cannot understand our social context and we s- uh, we cannot trust it at, in any, in most of our interactions, it's gonna make us crazy, um, or I should say crazier.

    5. PAPeter Attia2:01:09

       Yeah.

    6. IKIsaac "Zak" Kohane2:01:10

       Um, and my most positive, uh, aspect is I think that these tools can be used to, um, expand the creative expression of all people. If you're a poor driver like me, you're gonna be a better driver. If you're a lousy musici- musician but have a great ear, you're gonna be able to express yourself musically in ways that you could not do before. I think you're gonna see filmmakers who were never meant to be filmmakers before express themselves.

    7. PAPeter Attia2:01:42

       Hmm.

    8. IKIsaac "Zak" Kohane2:01:42

       I think human expression, it is going to be expanded because just like, um, printing press allowed, uh, all sorts of, uh- In fact, it's, it's a good analogy because the printing press also created a bunch of wars because it made a lot of people to make clear their opposition to the church and so on, uh, made, uh, enabled a, a number of bad things to happen. But it allowed also expression of all literature in the ways that would have not been possible without the printing press. So I, I... That's-- I'm looking forward to, um, yes, human, human expression and creativity. And, you know, if you've-- I'm sure, I can't imagine you haven't played with some of the picture generation or music generation capabilities of AI. Or if you have, I strongly recommend. You're gonna be amazed.

    9. PAPeter Attia2:02:32

       I have not. I, I, I am ashamed maybe to admit that I, my, my interactions with AI are limited to really, you know, ChatGPT-4 and, and basically problem-solving. Like, you know, uh, solve this problem for me [chuckles] right? And, and by the way, I think I'm doing it at a very, you know, JV level. I, I could really up my game there. There, there are-- You know, just before we started this podcast, I thought of a problem I've been asking my assistant to solve, and because, A, I don't have the time to solve it, and I'm not even sure how I would solve it, it would take me a long time. I've been asking her to solve it, and it's actually pretty hard. And then I realized, oh my God, why am I not asking ChatGPT-4 to do it? So I just started typing in the question. It's a bit of an elaborate question, and as soon as we're done with this podcast, I'll probably go right back to it. But I haven't done anything creatively with it. Um,

17. 2:03:22 – 2:08:02

    The implications of AI achieving a level of creativity & expertise comparable to exceptional human talents

    1. PAPeter Attia2:03:22

       what I will say is, what does this mean for human greatness, right? So, you know, right now, if you look at a book that's been written, you know, and someone who's won a Pulitzer Prize, you, you sort of recognize, like if you... I, I don't know if you read Sid, if you read Sid Mukherjee, right? He's one of my favorite writers, uh, when it comes to writing about science and medicine. And when I read something that Sid has written, I think to myself, there's a reason that he is so special. He, and he almost alone, can do something we can't do, right? I've written a book, doesn't matter. I could write a hundred books. I'll never write like Sid, and that's okay. I'm no worse a person.

    2. IKIsaac "Zak" Kohane2:04:07

       Well, probably not.

    3. PAPeter Attia2:04:08

       I'm no worse a person than Sid, but he has a special gift that I can appreciate, just as we could all appreciate watching an exceptional athlete or an exceptional artist or musician. Does it mean anything if that line becomes blurred?

    4. IKIsaac "Zak" Kohane2:04:24

       That's the right question. That's the right question. And yes, Sid writes like poetry. Um, but here's an answer which I don't like. I've heard many times people said, "Oh, you know that, uh, Deep Blue beat Kasparov in chess, but chess is more popular than it ever was, even though we know that the best chess players in the world are computers." So that's one answer. I don't like that answer at all.

    5. PAPeter Attia2:04:54

       Hmm.

    6. IKIsaac "Zak" Kohane2:04:54

       'Cause it does-- Uh, because the fact that I don't think if-- I mean, let's put, if we create Sid GPT-

    7. PAPeter Attia2:05:04

       [laughs]

    8. IKIsaac "Zak" Kohane2:05:06

       And, and Sid wrote, you know, Alzheimer's, the second greatest malady, and it wrote it in full Sid style, but it was not Sid.

    9. PAPeter Attia2:05:17

       Yeah.

    10. IKIsaac "Zak" Kohane2:05:17

        But it was just as empathic. Family references, you know, walking along, uh-

    11. PAPeter Attia2:05:23

        Right. The history, the n- the, the weaving of history with story, with science. Yeah.

    12. IKIsaac "Zak" Kohane2:05:27

        But if it did that-

    13. PAPeter Attia2:05:28

        Yeah

    14. IKIsaac "Zak" Kohane2:05:28

        ... and it was just a computer, how would you feel about it, Peter?

    15. PAPeter Attia2:05:31

        I, I mean, Zak, you are asking the jugular question. I would enjoy it, I think, just as much, but I, I don't know who I would praise. Like, I think m- maybe I have in me a, a weakness/tendency to want to idolize. You know, I'm not a religious person, so my idols aren't religious, but I do tend to love to see greatness.

    16. IKIsaac "Zak" Kohane2:05:56

        Yes.

    17. PAPeter Attia2:05:56

        I love to look at someone who wrote something who's amazing and say, "That, that, that amazes me." I love to be able to look at, you know, the best driver in the history of Formula One and study everything about what they did to make them so great. Um, so I'm not sure what it means in terms of that. What-- And I don't, I don't know how it would change that.

    18. IKIsaac "Zak" Kohane2:06:15

        I grew, I grew up in Switzerland, in Geneva, uh, and even though I have this American accent, both my parents were from Poland. And so the, the reason I have an American accent is I went to international school with a lot of Americans. But I-- all I read was whatever my dad would get me from, uh, England in science fiction, so I'm a big science fiction fan. So let me go science fiction on you to answer this question. Not gonna be in ten years, but it could be in fifty years, is you, you'll have idols, and the idols will be, yes, um, uh, Gregor- Gregorovich wrote a great novel, but you know, AI 521, what their, their understanding of the human condition is wonderful. I cry when I read their novels. And they'll just be part, uh, they'll be a, a part of the ecosystem, and this is where-- You know, they'll, they'll be entities within us. Whether they are self-aware or not will become a philosophical question. You know, let's, let's not go that narrow path, that disgusting rabbit hole where I went. I wonder, is, does Peter actually have consciousness or not? You know, does he have the same processes as I do? And we won't know that about these, or maybe we will, but will it matter if they're, if they're just among us? And they'll have brands, um, they will have... Yeah, they'll have brands, they'll have companies around them, they'll be superstars. And, uh, they'll be, you know, Dr. Fubar from, uh, Kansas, trained on, um, ay- Ayurvedic medicine, the key p- person for alt- uh, alternate medicine, not a, not a human. But we love what they do.

18. 2:08:02 – 2:14:10

    Digital immortality & legacy: the potential to emulate an individual's personality & responses & the ethical questions surrounding it

    1. PAPeter Attia2:08:02

       Okay, last question. How long until, from at least an intue- intellectual perspective, we are immortal? So if I died today, my children will not have access to my thoughts and musings any longer. Will there be a point at which, during my lifetime, an AI can be trained to be identical to me, at least from a goalpost perspective, to the point where after my death, my children could say, "Dad, what should I do about this situation?" And it can answer them in a way that I would have.

    2. IKIsaac "Zak" Kohane2:08:51

       So Peter, um, that's a great question. And people normally say, "Great question," because they're trying to buy themselves time to answer it. It's a great question because that was an early business plan that was generated shortly after P- GPT-4 came out. In fact, um, I was talking to, um, very briefly to Mark Cuban, and he-- because he saw GPT-4, I think he, uh, got, uh, tra- uh, trademarks or copyrights on his voice, all his work and likeness so that someone could not create a Mark who responded in all the ways he does. And I'll tell you that it sounds crazy, but there's a company called Rewind.ai, and I have it running right now. And everything that appears on my screen, it's recording. Every sound that it hears, it's recording. And if, um, it's doing-- If characters appear on the screen, it'll do-- it'll OCR them. If a voice appears, it'll voice to s-- And then if I have a q-question, I said, "When did I speak with Peter Attia?" It'll find it for me. It will-- I'll say, um, "When-- Who was I talking about, who was I talking about, um, AI and, uh, Alzheimer's?" And they'll find this video on a timeline. Now, all that data-

    3. PAPeter Attia2:10:41

       H-how many terabytes of data is this, Zak?

    4. IKIsaac "Zak" Kohane2:10:45

       Amazingly small. It's just gigabytes.

    5. PAPeter Attia2:10:48

       No, how, how is that possible?

    6. IKIsaac "Zak" Kohane2:10:50

       Because, A, it compresses it, it compresses it down in real time with using Apple silicon. And second of all, you and I, you're old, and you don't realize that gigabytes are not big on a standard Mac that has a terabyte. That's a thousand gigabytes.

    7. PAPeter Attia2:11:06

       Yeah.

    8. IKIsaac "Zak" Kohane2:11:07

       And so you can compress, uh, audio immensely. And what it's doing is it's actually not taking, um, um, video. It's just taking multiple snapshots-

    9. PAPeter Attia2:11:18

       I see

    10. IKIsaac "Zak" Kohane2:11:18

        ... every time the-

    11. PAPeter Attia2:11:19

        Okay

    12. IKIsaac "Zak" Kohane2:11:19

        ... screen changes by a certain amount.

    13. PAPeter Attia2:11:20

        Yeah, it's not trying to get video resolution per se.

    14. IKIsaac "Zak" Kohane2:11:23

        No.

    15. PAPeter Attia2:11:23

        Yeah, okay.

    16. IKIsaac "Zak" Kohane2:11:24

        No. And it's, and it's doing it. And I can see a timeline, and I-- it's quite remarkable. And so that is enough, in my opinion, data, so that with enough, uh, conversations like this, someone could create a pretty good approximation of at least public Zak.

    17. PAPeter Attia2:11:46

        So then the next question is, is-- are-- is Zak willing to have Rewind AI on a recording device, his phone with him twenty four/seven in his private moments, in his intimate moments, when he is arguing with his wife, when he's upset at his kids, when he's having the most amazing experience with his postdoc. Like if you think about the entire range of experiences we have from the good, the bad, the ugly, those are probably necessary if we want to formulate the essence of ourselves. Um, you envision a day in which people can say, "Look, I'm, I'm willing to take the risks associated with that," and there are clear risks associated with doing that, "but I'm willing to take those risks in order to have this legacy, this data set to be turned into a legacy."

    18. IKIsaac "Zak" Kohane2:12:35

        So I, I think it's actually pretty creepy to, um, to come back from the dead to talk to your children.

    19. PAPeter Attia2:12:39

        Hmm.

    20. IKIsaac "Zak" Kohane2:12:40

        But, so I actually have other, other goals, like being able to-- Here's where I take it. We are being monitored all the time. We have iPhones. We have Alexa devices. I don't know what is actually being stored by whom and what. And people are gonna use this data in ways that we do or don't know. I feel it's us, the little guy, if we have our own copy, and we can say, "Well, actually, look, this is what I said then."

    21. PAPeter Attia2:13:09

        Yeah, that was taken out of context. Yeah. [chuckles]

    22. IKIsaac "Zak" Kohane2:13:11

        Yeah. That was taken out of context. And I can do it. I have an assistant that can just find it, and find exactly, and find all the time I said it. I think that's, that's good. I, I still, I, I don't think it's-- I think it's messing with your kid's head to have, uh, you come back from the dead, uh, and give advice, even though m-they might be tempted. Technically, I think it's gonna be not that difficult.

    23. PAPeter Attia2:13:34

        Wow.

    24. IKIsaac "Zak" Kohane2:13:35

        Um, and just again, speaking about Rewind AI, again, I have no c-- I have no stake in them. Um, and I don't think I-- I think I might have paid them for a, a license to run on my computer. But it's also, the microphone is always on. So when I'm talking to students in, in my office, uh, it's, it's, uh, taking that down. So yeah, I'm sure some peop-- There are some moments in my life where I w-- I don't wanna be on record, but it's, uh, there are big chunks of my life that are actually being stored this way

19. 2:14:10 – 2:16:15

    Parting thoughts

    1. PAPeter Attia2:14:10

       Well, Zak, this has been, um, this has been a very interesting discussion. I, I've learned a lot because I, I think, you know, I, I probably came into this discussion with about the same level of knowledge, m- maybe slightly more than the average person, but clearly not much more, um, on just the, the general principles of AI, the evolution of AI. Um, and I think what I come away from this, I come away from this much-- I, I guess if anything surprises me, and, and a lot does, but nothing surprises me more than the timescale that you've painted for the evolution within my particular field and your particular field, which is medicine. I had no clue that we were getting this close to that level of intelligence.

    2. IKIsaac "Zak" Kohane2:14:58

       So Peter, if I were you, and, and this is not an offer 'cause I'm too busy, but you're a capable guy and you have a great network. If I was running the clinic that you're running, I would take advantage of now. I would get those, those videos and th- those, those, uh, sounds and get all my patients with, with, of course, their consent to be part of this and to actually follow their progress. Uh, not just the way they report it, but by their gait, by the way they look. Um, you can do great things in what you're doing and advance the state of art. You know, you're asking who's gonna do it. You're doing, uh, some interesting things. You can be pushing the envelope using these technologies as just another very smart, comprehensive assistant.

    3. PAPeter Attia2:15:49

       Zak, you've given, you've given me a lot to think about. I, uh, I'm, I'm grateful for your time and, and obviously for your, your insight and years of dedication that have, uh, allowed us to be sitting here having this discussion. Thank you very much.

    4. IKIsaac "Zak" Kohane2:16:02

       It was a great pleasure. Thank you for your time. [upbeat music]

Episode duration: 2:16:15