Short Summary: Deep dive into systems biology and personalized medicine, exploring how technology and data can revolutionize health care, artificial intelligence, biotechnology, and the future of medicine.
About the Guest: Dr. Lee Hood is a pioneering scientist with a 60-year career in biology, notably at Caltech and the University of Washington. He co-founded the Institute for Systems Biology and has significantly contributed to molecular immunology and the Human Genome Project, holding a PhD in biology.
Note: Podcast episodes are fully available to paid subscribers on the M&M Substack and to everyone on YouTube. Partial versions are available elsewhere.
Episode Summary: Delve into the evolution of biological research from traditional methods to systems biology. Dr. Hood explains the shift from studying individual parts of biological systems to understanding their interconnectedness. He discusses the implications of big data in biology, particularly in medicine, emphasizing predictive, preventive, personalized, and participatory healthcare approaches. The conversation also touches on the integration of AI in medicine, limitations of current drug development strategies, and potential of new therapeutic avenues like peptides.
Key Takeaways:
Systems Biology: Understanding complex biological systems by analyzing how individual components interact.
Data-Driven Health: Use of genomic and phenomic data can lead to personalized health strategies, enhancing wellness and preventing chronic diseases before they manifest.
AI in Medicine: AI can augment human capabilities in medicine, acting as a vast knowledge base to assist physicians in diagnosis and treatment, potentially leading to a partnership model between AI and human doctors.
Chronic Disease: Much of the chronic disease burden could potentially be mitigated through lifestyle changes rather than solely through pharmaceutical interventions.
Future Drug Development: The traditional focus on single-target drugs might shift towards multi-modal strategies, recognizing diseases like Alzheimer's might be a metabolic disorder.
Peptides and New Therapies: Small peptide drugs are emerging as potential new treatments due to their ability to interact with a range of biological molecules, offering new possibilities beyond traditional small molecule drugs.
Environmental Impact on Health: While personal behavior significantly influences health outcomes, environmental factors like exposure to toxins can also play critical roles in disease development.
Related episodes:
M&M #204: Preventive Medicine, Personalized Nutrition & Changing Your Microbiome | Momo Vuyisich
M&M #47: Metabolism, Blood Sugar, Microbiome, Diet, Aging, Digital Phenotyping & Personalized Medicine | Nathan Price
*Not medical advice.
Support M&M if you find value in this content.
Episode transcript below.
Full AI-generated transcript below. Beware of typos & mistranslations!
Lee Hood 1:43
I'm a scientist that has had now a 60 year career in science, and really got started at Caltech in 1970 when I was a young assistant professor, and deciding that my big problem that I wanted to attack in the future was human complexity, and in thinking about it in the 1970s there were two issues to begin with that were really challenging. One was the issue of we didn't have the technologies to generate a lot of data, and a lot of data are really key to be able to dealing with complexity. And number two, if we generated all that data, we didn't have very good ways of thinking about how it could be used to decipher complexity. So my lab initially focused on molecular immunology, which was studying in the ultimate complexity in human biology, and it started developing technologies. And over the next 20 years, we developed six instruments that let us manipulate and and synthesize and sequence proteins and genes, and one of these was the automated DNA sequencer. So from bringing engineering to biology, which we did in that first 20 year period, to actually getting involved in the human genome project because of the first human automated DNA sequencer, was a second paradigm change that was really I embraced fully and and I was a part of a committee in 85 that looked into the whole question of whether we should do the genome or not, and we ended up being split six to six on that question, where the six opposed to it were opposed because it was big science that would take the money away from small science and so forth. But in the end, it started in 1990 and, of course, was finished in 2003 and it gave us a framework, a really important framework, for beginning to think about the source code of complexity and human organisms. And the next thing that happened is I really started thinking about how to deal effectively with the complexity of the data we were starting to generate. And that led to thinking about systems biology and about generating new technologies even more effectively, not by having a big lab that had all of the diversity of science you needed to generate technologies, but by creating a department to do so. And Cal Tech biology was absolutely against creating a new biology department. So Bill Gates made it possible in 1992 to. Me to move to the University of Washington and set up such a department, and it was really spectacularly successful. But in the end, I decided I wasn't moving systems biology and systems thinking fast enough. So in 2000 I created the Institute for Systems Biology with two other really terrific co founders, and that really pioneered global holistic thinking about biology and and about disease, and in the early 2000s applying systems thinking to healthcare made me realize, one, that there was going to be a discipline of systems medicine, and two, to think about what would be an ideal type of health care. And I decided at that time it should be p4 predictive, preventive, personalized and participatory. We can talk later about what each of those mean. So
Nick Jikomes 6:06
so if I step back for a sec, you've had a long career, and basically so you've been you've been a part of the evolution of biology in terms of how it's conducted. So if we go back to the Human Genome Project, so I'm just a little boy at this time. So this is, this is for my, my time in science. You know, that was a big deal. It took many years to get that first human genome done, just one, whereas today, you know, you can give me my genome sequence, and I don't know, probably by the end of the day, at some tiny fraction of the cost. And so you were part, you were a big part, of building some of the technologies that enabled all of this stuff to happen more rapidly and more cheaply. And a natural consequence of that technological progress is larger and more complex data sets. And we're at the Institute of systems biology. What is systems biology in terms of its relationship to these bigger, larger data sets that depend on these technological innovations. And I guess another way of saying that in simpler terms is, what is systems biology and what, as opposed to, like, simpler or older forms of biology? How do those things differ? Exactly
Lee Hood 7:16
So, I think the analogy is the analogy of a car. If you wanted to take a systems approach to understanding how a car worked. One, you would take the car apart, and you'd get each of the individual parts. Two, you'd try to understand what that part did and what it was and how it was described. And then three, you'd have to understand how the parts were all connected together to build a coherent object. And then four, you'd need to understand the dynamics of how those parts could actually move the car and make predictions about what you could do with it, and the same is true in biology. If you want to understand Alzheimer's, we have to be able to understand a great deal about the complexity of the brain. But more than that, we have to understand the fundamental parts that deal with cognition, how they're interrelated, and what happens to them during the disease process, and raise questions about how we can prevent what actually happens, which is the destruction of neurons and the loss of the interactions that lead, successively to the inability to speak and communicate and eventually even to think
Nick Jikomes 8:43
and so what are, what are some of the things you guys are working on today here, in terms of systems biology projects, just broadly speaking,
Lee Hood 8:54
well, the biggest projects we're working on now really started in 2013 and 14, where we took 108 individuals and we decided to analyze their genomes in their phenomes longitudinally, taking multiple samples of blood and so forth.
Nick Jikomes 9:16
So genome is all the genes they have. Phenome is all the proteins. So the
Lee Hood 9:21
genome is basically the whole source code of DNA. It's the sequence of the entire 23 pairs of human chromosomes and everything that encode the 20,000 or so genes, which are the molecular machines that actually execute many of the biological functions and so forth. So what systems biology attempts to understand is how those informational components work in conjunction with one another. So in a sense, the genome is the. Code of life, and for a human, it specifies, in general ways, what exactly the human can potentially become, okay, but that potential is modulated by two things that arise from individual, one, their behavior, diet, exercise, sleep, all of those kind of things and to their environment. And these two things together probably constitute 90% of what determine how you change as you develop from birth to death, and that's called your phenotype. And the way we measure this changing phenotype is to look at the phenome, which are we see in systems biology, the blood being a window into health and disease, because it bathes all the organs in your internal body, and we can capture molecules secreted by each of these organs and read out their behavior in various ways. Or we can look at your gut microbiome, microbes that are present in your gut through which your diet and your pills and everything else must filter to get into your blood. So a
Nick Jikomes 11:23
phenome, a phenome would be the set of all the phenotypes or traits that someone has, and for any given trait that's going to be a function of a combination of your genotype, what's in your genome, and how that is being allowed or enabled to be expressed through environmental and behavioral interactions. So So for example, if we measured, say, blood cholesterol in you or me, we would get some number for LDL. Our genetics have something to do with that, but also our diet and our sleep patterns and our life history are interacting with what's in our genome, and that's what gives us the phenotype that we would measure at any given time point in our life.
Lee Hood 12:00
That's exactly right. So the phenome is essentially any feature in a human organism that changes over time. So the phenome can be measured by imaging. It can be measured by blood analytes. It can be measured by sequencing your gut microbiome to determine the ratio of number of species, and it can be determined by wearing wearables like Fitbits and aura rings that essentially measure your normal physiology, your exercise and Sleep and so forth.
Nick Jikomes 12:40
And so how does thinking about the phenome, thinking about a systems biology approach and all the data we have? So there's not only data that we can collect at a mass scale in laboratory environments like this, but as you just mentioned, wearables are very common. It's now quite easy for even an average person to collect quite a bit of data on themselves, from getting blood work done, from wearing an aura ring, or whatever it may be. No connection to our ring, by the way, we're just we're just chatting off the cuff. All of these things are giving us data. We're getting measurements of our phenome. How does sort of the systems biology approach and how we think about using this data. How does that start to get us thinking about medicine from the standpoint of prevention, rather than diagnosing and fixing things after they go wrong? Yeah.
Lee Hood 13:34
So the basic idea with this data driven approach, where you have the genome and phenome measurements of individuals is basically one. We can look at the genome by itself and find that there are of the order of eight different classifications of actionable possibilities where variants can lead to if you have a particular kind of variant, you know exactly what to do to cure the ill that it will cause. Okay, and these fall into a whole series of categories. There are 81 variants that have been designated by the American College of Medical Genetics that are actionable. And an example of this is there's a variant called malignant hypothermia. And this is a variant where, if you take the wrong kind of anesthetic during a surgery, it can absolutely destroy your thermal control mechanism, and you heat irreversibly, and it kills you. And so that's a very dramatic if you have that gene, we know now which anesthetics you need to stay around, but knowing that you have it is key. Or it'll kill you. So that's that's one category. There's a second category of variants that make it impossible for you to use certain drugs. And if you, your physician knows you have those variants, he can call for other drugs that will avoid that. And there's a third category of genes that are called polygenic features, that are a collection of genes that give you a genetic risk for a variety of diseases and normal types of physiology. So for example, there is a polygenic score, say, for Alzheimer's, that says you have a very elevated risk if you have all of these variants, each of which contribute a very little piece to that whole score. And those variants actually tell us some of the machinery that causes Alzheimer's once they're analyzed. So anyway, the genome leads to a lot of actionable possibilities, but the phenome does too, and and what we did, for example, in the first study we looked at 108 people, is we did. We took the six different types of data, the genome variants and and we actually had five phenome measurements. And so we could show there were statistical correlations at each time point in individuals that were really striking in their their disposition. And these correlations of different kinds of features in the genome and phenome led us to the medical literature and to actionable possibilities that could actually either improve your wellness or let you avoid disease. And an example of one was with myself, I was incredibly low in vitamin D, and so I was prescribed to take 1000 international units. Yeah, start bringing it up.
Nick Jikomes 17:12
I mean, your fair skinned were in Seattle, yeah.
Lee Hood 17:14
But it didn't touch it. And it turned out I had two gene variants that blocked the uptake of vitamin D.
Nick Jikomes 17:21
Oh, you're saying you took supplemental vitamin D, but it didn't move the needle for you. It at 1000
Lee Hood 17:25
international units. I needed to do 15,000 internet so
Nick Jikomes 17:30
if all you knew is that your vitamin D was low from a blood test and your doctor said, Well, let's take the standard recommended dose for something, nothing would happen, and you wouldn't know that
Lee Hood 17:40
you wouldn't, you wouldn't know that unless you did the test again. So it's putting these two things together led to an actionable possibility that could and vitamin D level is important for a lot of different reasons, and in Seattle, there are a lot of people that have low vitamin D, yeah, and most of them with 1000 international units, can bring it right up to normal, but some might be like you, and that's some like me, and by a factor of 10, therefore you should know about your genome as well as your genome well.
Nick Jikomes 18:12
So I want to this brings me to some interesting questions. So as fascinating as it would be to talk about the biology here, I want to talk about some of the practicalities. So, you know, you just mentioned a few so, so, so the the hope here is I can get my genome sequenced. I can have phenome type data, I can have wearables. I can have all of this stuff. I can collect all of this data. But how do I get to actionable insights from that? Well, in theory, right? All of this stuff could be plugged into some master database, and then I could bring it to my doctor, and they could simply tell me what to do. So let's you know if I'm assuming I'm the average person, I don't have a PhD in biology, I don't know how to interpret all of this data myself. I need someone to do that for me. Ideally, that's going to be a physician that I see face to face. But there's a lot of data, a lot of complexity that gets really big, really fast here, and frankly, I'm not sure if I brought all of the biomarkers I could bring to a primary care doctor here in Seattle, would they even themselves be able to interpret it all for me and tell me what to do? How do we think about the implementation side of all of using all of this data to guide health when sort of the expertise required to dissect all of this phenome style systems biology data is so vast that even the average primary care physician might not themselves have the expertise to interpret it.
Lee Hood 19:36
I would say 99% of the physicians out there wouldn't have the faintest idea how to interpret 95% of the data that you can bring them Okay, so how do we go about it? Well, I think this wraps up three interesting problems in this whole approach to data driven health. So one problem. Is, how do we persuade people that this is a good way to move toward health? So it means you have to educate the people about what the genome is and about what the phenome is and why these things could be really powerful in assisting you. And you know, the ultimate objective we have of this is health of the individual, and the health can best be reflected in a significant expansion of your health span. So typically, the health span of normal individuals goes into the 60s or so, and then after that, an awful lot of people decay very, very rapidly. We think with data driven health that we're generating, we'll be able to take people into the 90s and have them absolutely mentally alert, physically capable and so forth. So a really interesting question that comes up sociologically is, what are you going to do with the extra 20 years or so? Yeah, and how does that affect retirement? And how are you going to pay for I mean, they're just a marvel of but the really important point is data driven, health can absolutely optimize your wellness, and that's important for development, for education, for jobs, for creativity, for happiness, all of those. I mean, health underlies all of those in a fundamental way, and it transforms individuals, and it can transform citizens in a country in the same way, because you have a lot of creative people that can go on and find so health is really good. A, how do we educate them? And B, once we've educated them, what we did with a company called aravail, which was a scientific wellness company that used these techniques, was we had coaches, and a good coach could easily handle 500 patients or so. We ended up, over a four year period, recruiting about 5000 people or so, and so we were able to improve their wellness and help them avoid disease and really striking ways.
Nick Jikomes 22:32
So let's get a little more concrete with that. When you say a coach, are these physicians, or are they just people that are trained to interpret the data sets that you were using. They are
Lee Hood 22:41
generally not physicians. They're people that usually came out of a background of nutrition or of some some nursing and things like that, who are especially trained to be able to to be able to take the actionable possibilities, and arivale generated more than 200 of them, and explain the relevant ones to individuals needed to practice to those actionable possibilities. And I think the coaches were remarkably successful. But we think in the future, this is going to get into the millions and 10s and hundreds of millions of people, and if you need a coach for every 1000 individuals, coaches don't scale. So what we're beginning to study now is we've created a chat bot that has the ability. For example, we took a chat bot and had them one read the book I wrote on the age of scientific wellness in all its detail. That's your book. That's my book, and with Nathan price, and then we had the chat bot look at all the lectures I had on, on out in the literature and everything like that. And so now this chat bot can answer any questions on data driven health at any level you want to ask the question, appropriate for a patient, approach, appropriate for a physician or appropriate for a data scientist?
Nick Jikomes 24:29
So given, so I want to tie a few things together there. So earlier, you said, you know, just given, sort of the vastness of the data sets we're talking about that that can be generated. Now you said something like the vast majority of physicians won't know how to interpret the vast majority of this data. That's not necessarily a knock on physicians themselves. It's just the complexity here. The biology is so complex and the data sets are so vast, nobody can have total expertise in every domain. Nobody is going to be able to interpret every single piece of data. You know, it's just not possible for a single human being to do that. But, you know, one thing that was implicit in what you just said is, you know, we're developing AI systems now that can read the entire internet basically in, you know, in a day, and they can be trained on different text sets of text to different levels of expertise, a lot of this stuff can be baked into chat bots. We know that. You know already today, some of these large language models are rivaling or even surpassing the diagnostic capabilities of human physicians that have been working for decades. What does this say about sort of the future of doctor patient interactions and the role for human beings in the role of the physician. Are humans just going to get surpassed by AI, given the vastness and the complexity of the data we're talking about? No,
Lee Hood 25:51
not at all. I think the way to look at it is, in the future, physicians will be partners with AI, and that AI will be the enormous catalog that will have give them at a fingers tip, all of the expertise you need in all domains of medicine, so you can begin to think about Treating your patients in ways no physician has ever been able to treat it before. So the physician is the grand integrator and and the chat bot or or large language model transformer essentially supplies information a in the context of the actionable possibilities for a given patient, in the context of their genome and their phenome at that time, and B then provides the evidence that those actionable possibilities are medically valid, because physicians quite rightly, want to be assured that they're there. They're reaching valid conclusions for their patients and everything. So I think the really important thing is for us to educate physicians now as to the enormous power of education in transforming how they can practice health care in the future.
Nick Jikomes 27:29
Another thing I want to ask you about that I think is really interesting, and it's an interesting problem here. So I want to ask about the quality of the scientific literature, the protocols and the selection of literature we use in the training process of these AI models in order to ensure that they are you know, that they're giving us good, robust, valid information that is actionable and is correct. So the challenge here is kind of interesting, because you might think, okay, if I train a large language model on all of the peer reviewed work, all the peer reviewed scientific publications in a given domain, it's now going to be like the super expert scientist in this field, and it's going to give me the best, most robust results. But of course, as you know, a lot of just because something is published in a peer reviewed journal doesn't mean it's good science. There's a lot of junk out there due to either, you know, sloppiness of work, sometimes even fraud, that's, that's, there's been a lot of big controversies lately out there. And just to give one example of how this like could quickly get really messy. So a field that I follow closely includes the work of John Speakman, who's a scientist who studies metabolism and diet. He just came out with a paper that basically quantified something that many of us have known already for years, which is that there's an entire field of study that's often called nutrition epidemiology, and it's based on these food frequency questionnaires where they just ask people to remember what they ate in the past few days or weeks or months or even years. And all of the results are based on that data. And the data is basically garbage. You can argue about exactly how garbage it is, but these are not precise measurements, and basically it was just quantified how off a lot of these data sets are. But you know, if I was, you know, if I was a large language model guy, I could say, well, I want to create something. I want to create an AI doctor or AI nutritionist. I'm going to give it all of the nutrition, epidemiology literature, peer review papers published by scientists at Harvard and Stanford and all these big name places with brand name journals behind them, and yet that training data would be largely bogus and give a lot of bad results. I'm just using that as an example. How do you think about how we train the AI systems on the literature, given that we can't just give it all of literature, because there's so much that there's a huge mixture of really good, really bad and really mean. Yoker science out there?
Lee Hood 30:00
No, no, I think you have to be able to give them all literature and then give them the power to discriminate between good and bad science. And the power to do that is going to come from what are called Knowledge graphs. And these are graphs of the entire span of biomedical information that have been annotated, and there are a couple of them that are, one very large Google and one that's been annotated in enormous detail. And those are highly accurate standards that these large language models will ingest and compare everything it gets against those high quality data, annotated networks, annotated communities of interaction and so forth. So I think, and it's absolutely critical in the beginning that we make sure we deal with the possible not only errors in in data, which I think we can really deal with effectively, but the hallucination of the Transformers as well. And I think with the combination of this highly annotated PubMed, we're easily going to be able to do that, and it'll be a context. It'll be a baseline of reality and truth against which everything can be checked. Yeah,
Nick Jikomes 31:31
it sounds like basically what you're saying is there are ways to set up training protocols such that even though you are giving the AI quote all of the data, you can also give it human labeled subsets of the data that give it a sense of a standard by which it could compare all the rest of the things to something that a human is pre specified as being rock solid. And this can be used in a way to ensure that these AI systems are generating robust results and they're not just sort of blindly giving equal weight to everything that we feed them.
Lee Hood 32:06
Yeah, and in fact, what I could suggest to your listeners is that you read a book written recently by Henry Kissinger and Craig Monday called Genesis. And this is a description of the history of AI and how it's come about. It's an Henry Kissinger, like the Henry Kissinger, the Henry Kissinger, it was a it was unfinished when he died, and and and and Craig finished up the book, but it's a spectacular book because it tells you about the history. It gives you a clear idea of how powerful these things are going to be, and then it really talks about the sociology and ethics of how we deal with hallucinations and how we give the large language models a human soul so they can make appropriate kinds of judgments between good and bad, whatever it is. Okay, so? And one kind of thing that's been suggested is suppose, along with everything else in science, you gave them the sociology and history of science and philosophy and ethics and a whole variety of other kinds of things like that, you could begin To give them the integrated human wisdom that could influence their view of the entire internet of data out there and so and so forth. Yeah,
Nick Jikomes 33:50
so this would be large language models that aren't just trained on everything in PubMed, but they've also read Aristotle.
Lee Hood 33:58
It's really a very interesting idea. Okay,
Nick Jikomes 34:03
interesting. I didn't I had no idea Henry Kissinger was had been writing about that subject.
Lee Hood 34:07
Oh, his last two books were really on on AI, yeah,
Nick Jikomes 34:12
interesting. So where do you think? I mean, there's a lot of places we can go with this. Weird. I mean, where do you think AI is going to be going, broadly speaking, as it relates to science and medicine? And I don't, I don't just mean, you know, I think, basically, I would love if you could kind of just riff on what you just said, you know, I think it's easy to imagine AI systems that are just, just narrowly trained on, you know, diagnostic tasks, or the scientific literature to do these sort of robotic or Spock like, you know, you know, impersonal, dispassionate analyzes and and diagnostic tasks for humans. But what you're getting at is, you know, we're probably going to see ais that are not just trained to have scientific. Or medical expertise, but they are also trained in ways that give them what we might call a kind of human wisdom. And where do you see that going? Well,
Lee Hood 35:10
I think they'll be trained in ways that have very deep knowledge. And I think one the prediction is within a few years, we'll have large language models that far exceed the capacity of expert scientists to be able to deal with the scientific facts of their information, and probably both Microsoft open AI and Google now are generating the next generation of transformers, and I think even those people have no idea of just where they're going and what they're going to be able to do, except that they're going to far exceed the previous generation, whose capacities we know at least a little bit about. So I think the potential for the future is is very great in amassing knowledge and and being able to gain new insights, but it also requires a careful formulation to keep it in a sense, under control. And under control is defined very differently by different kinds of people and so forth, but, and that's what I think the book Genesis tried to get at how we could give, how we could give large language models, a sense of responsibility with regard to our ethics and sociology.
Nick Jikomes 36:53
And there's sort of lots of different views on how the whole AI thing could play out. Generally speaking, you might even say there's new ideologies, or potentially even new religions forming in in this realm. So one of the more optimistic views is basically that, well, we're about to hit what some people call the singularity AI is going to sort of integrate, there's going to be human machine interfaces. We're basically going to integrate with artificial general intelligence, and effectively we will all, or at least some of us, become totally immortalized and escape the limitations of human biology that humanity has historically been limited by. The more pessimistic view is that, you know, we're basically going to just create an AI that gets rid of humanity because it decides that's what's best to do. Where in that spectrum of optimism to pessimism, do you sit? And what do you think sort of the future of human mortality is going to look like?
Lee Hood 37:56
Well, I think the responsibility we have as humans is to be able to educate these intelligences so they're in sync with our values and our philosophy and our way of looking at things. I mean, the one I mean, I think, I think there are still interesting issues. It's one thing to have infinite knowledge and to be able to connect it all. Does that give you the ability to creatively jump a gap that has never been jumped before and make new fundamental realizations about things humans clearly have done that repeatedly out in the context of their scientific development. And I think this is whether these intelligences will be able to do that, I think is an open question. But the most important part is we have to educate them in a proper way that is consistent with a partnership that is acceptable to us, that I think is very important, and we don't know exactly what that implies at this point in time,
Nick Jikomes 39:15
yeah, and I guess that's, that's a big, open question. That's also the scary thing, yes, we're sort of absolutely we sort of, I mean, if I sort of reflect what you say back to you and interpret it a little bit, on the one hand, you know, two things are true simultaneously, we would probably all agree in some way, shape or form, that these things need to be trained in a very mindful fashion, so that there's, you know, what people call AI alignment, that the AI is going To stay aligned with humanity. Two problems there that immediately emerge. One, we don't actually know how to achieve that alignment. We're just we know we're flying by the seat of our pants. We're just sort of hoping we achieve it. But no one, I don't think, really knows how to ensure that happens. And two is that when we say that we want the AI aligned with us, an implicit assumption that. That I think is worth exploring is, you know, who are we referring to when we say us? Because, you know, we're all humans. Everyone shares one common humanity in the sense that we all belong to the same species. But you know, the entire history of humanity is a story of factions of humans competing against one another and and so you know, the question of how the AI is trained and who it is trained by, is very important, because I don't think any one group knows exactly how this is going to play out. But also, if history is a guide at all, there's certainly no guarantee that whoever is at the helm of training these things is going to have all of humanity in mind, rather than just a subset.
Lee Hood 40:43
So I agree with everything you said there. Basically, I think the really, the most fundamental belief I have is once you let the genie out of the box, you're not going to put the genie back in the box, and so we have to deal with the genie that's out there, and all we can do is hope we deal with it in ways that are responsible by according to our own ethics. And I admit there'll be very different groups that might have very different kinds of ethics, and that's a given that we have to deal with?
Nick Jikomes 41:21
Yeah. I mean, I guess yeah, in some sense that that is just the naked truth. The genie is out of the bottle, and we're just gonna have to deal with it, and different groups are gonna have different thoughts. And
Lee Hood 41:36
I think it's like, it's like politics today. I mean, many people have differing, divergent opinions on what's happened, but I think the really important thing for all of the the sides in this political domain are, let's take the most optimistic opportunity. Let's take advantage of what circumstances have given us again. The genie is out of the box. We're not going to put it back and change anything, but let's do the best we can, each of us, in our own ways, with what we're facing.
Nick Jikomes 42:19
I want to ask you some stuff about Aging and Longevity, but also just how we think about medicine and healing, broadly speaking. So, you know, sort of underlying all of this is sort of your background in systems biology and all of the stuff that you're doing. And there's many ways we could take this, but you know, historically, if we just step back and take sort of a bird's eye view, there's been different, sort of general approaches to human health and medicine. You might even call it different ideologies of medicine, one of them, and I'm sort of half making this up on the fly, so I'm using my own terminology here. It might conflict with what other people say, but this is, this is me speaking, you know, one, one view is what we might just call the traditional, the traditional biomedical drug development view is you think of diseases as things that have gone wrong in the body, sort of like the way that you would think about something going wrong in your car. The car is this complex machine. The Nuts and Bolts get rusty. Over time, things get loose. The radiator goes funky, you know, something overheats, and after it goes wrong, you have to go in and take something out or put something in. You have to fix individual pieces. So there's this sort of reductive approach to drug development that that's been a big part of medicine the last several decades, right? You develop a single molecule that binds ideally to one receptor in one way, that has one clear effect and a minimum of side effects, and you know that has its merits and its strengths and weaknesses, just like any other approach. Sort of another kind of approach that some people have is what we might call the holistic or healing approach. And I guess, you know, one of the basic ideas here would be, you know, people with this mindset tend to think, well, you know, our bodies are these evolved, organic things. They are complex. They have a lot of built in feedback mechanisms. The healing is sort of already known. The body knows how to heal itself, and it's all if it's not fundamentally broken in its healing capacity, it has everything it needs the body can be enabled to heal itself. For example, there are viewpoints by physicians and scientists going back quite a long ways that you know, perhaps we can think of cancer as a failure of the healing process itself. How do youth, when you think about Aging and Longevity and how our health deteriorates over time, sort of, how do you start to think about it at those high levels, in terms of, you know, is, is fixing something like aging or just disease generally? Is it more a matter of waiting for things to go wrong and fixing specific problems, or is there. There a way in which we should think about these things as being a degradation of the body's own organic healing processes. So
Lee Hood 45:07
I think that's really a terrific question, and the heart of it underlies the fatal flaw in contemporary medicine, which is almost entirely focused on disease and dealing with disease. And I would argue real health should do two things. One, it should be able to optimize wellness, to take individuals well beyond what they might be normally, and make them far Weller and age far more effectively. That's one aspect and two, we should be able to diagnose most diseases, chronic diseases, especially early, so we can reverse them when they're simple and easy to change. That means we should never have chronic diseases, and this the statistics are striking. In our $4.5 trillion health care bill today, 86% of our dollars are spent on chronic diseases. So if we really could deal with those. We could put enormous resources into the optimization of wellness and healthy aging, and I can talk to the whole idea of aging now. And for example, my my nonprofit called phenome health is very interested in data driven health of the type we've been talking about. Recently, we allied with the Buck Institute for aging research in Marin County, and the reason we did that is because scientific wellness, which is what the data driven health is pushing. Is one side of the coin of healthy aging, and the other side is what is called Gero science, the ability to look at each of the 12 hallmarks of aging and come to understand how we can use those insights to optimize the aging process, and the bringing and merging of these two things together, scientific wellness with Gero science, I think, is really going to push us to where we can get the kind of health spans we've been talking about, and the focus will entirely be on optimizing wellness, on improving healthy aging and avoiding chronic disease, a fundamentally different orientation that what we have in our $4.5 billion health care budget today, and of course, a really interesting sociologic question is, how are we going to make the transition? Yeah,
Nick Jikomes 48:09
so to what extent do you think combating chronic disease is a matter of developing new drugs, treatments and interventions, versus avoiding the things that cause it and the things that avoiding the things in the environment that are creating it in the first place.
Lee Hood 48:33
So I would say there are three different ways you can get to chronic diseases. One is, you can have a fundamental flaw in your genetic material, and they're probably the only way we're going to easily cure it is, is through drugs, okay?
Nick Jikomes 48:55
And that's so that definitely happens to a minority of people, but it does happen, and because it's an intrinsic thing, that's why you need an intervention of some kind, right? Yeah, that's right. Okay, what's number two?
Lee Hood 49:05
So number two is you can have defects and lead to chronic disease through your behavior. And diabetes is absolutely the classic account. Obesity is a classic example. And to a certain extent, neurodegenerative diseases can be an example. And there, what we know is you can have profound impacts on these with changing life habits. And if you had to pick one life habit that I think is really fundamental to healthy aging is exercise. Exercise, I think it dominates in a really striking way. Sleep is important, diet is important, stress, those are all things that we have to be able to deal with. But I think exercise really can. Trump, in major ways, a lot of these, these chronic kinds of issues. And then the third thing is the environment that can give the toxins. And yeah, it's known, for example, that black mold in houses where you live in certain people can induce absolutely classic Alzheimer's disease. If you move the people out of those toxic environments, it reverses. If you do it early, if you do it early enough, yeah, if you, if you've lost neurons, then yeah, you're, you're beyond, yeah. But So it's things like that that we're going to learn from data driven health, and we'll be able to prevent all of those different kinds of things. And so going back to the genome again, they're really several different approaches to to repairing a genetic defect, and, for example, the gene that causes hyper hypercholesterolemia. So one is drugs, and there is a drug that can deal with that very effectively, not classic statins. You need a much more powerful drug. But a second is if we could actually in the early embryos, genetically modify those genes to make them normal, and we have the gene engineering techniques to do it, although there's a long ways to go with all the technical details that are complexities arising from but in principle, in The future, we're going to be able to re engineer things so you won't have to worry about that. And, and, and I think a third thing is we may be able to to generate new and much more effective drugs.
Nick Jikomes 52:00
And I do want to talk about things like new and more effective drugs or therapies. So as we briefly mentioned, you know, when we say the word drug today, it's basically taken for granted that we're talking about a small molecule drug which has been specifically developed through the biopharmaceutical protocols that that are in our society, and those have been engineered to try and have a lot of specificity. We tend to like to develop drugs that have one or very small number of targets, right? They bind to one protein only, ideally. And the idea there is, we want it to have a very specific, narrow effect, and we want to minimize the side effects or any collateral damage that might happen. And that, you know, there's a whole host of challenges and criticisms that can go with that approach. What do you think about sort of, the traditional drug development approach as it has existed up until now in the US? And what do you what are some of the new avenues being explored sort of as alternatives to small molecule high specificity drugs,
Lee Hood 53:05
okay, well, that, that is a very interesting question. So in my view, I can tell you, one of the most spectacular failures in the pharmaceutical industry is developing drugs against amyloid for Alzheimer's. And over the last 12 years, there have been more than 500 clinical trials, a billion dollars plus, and almost all of them have failed by most reasonable kinds of criteria. And the reason they failed is because amyloid is almost certainly a result of and not a cause of
Nick Jikomes 53:52
to what extent. So what was underlying that is that was that just a hopeful mistake, where the drug companies were mistaking a consequence of disease, for a cause of disease, or was there literature that was incorrectly, for some reason or another, pointing to that as the causal agent? Well,
Lee Hood 54:09
I think clearly there was literature that tried to point to it as a causal agent, but I think it was a Lemmer like mentality. It was a simple, easy explanation. And you and something you could easily make drugs against, and they made drugs that can remove amyloid and and, in fact, one of the really interesting systems is, is, if you strip amyloid out of many patients that have Alzheimer's, they you make them worse. Wow, because amyloid has a really important role in facilitating inter neuron communication, and you take it away, so in your worse off than were before. So, so that was there are other people that have Alzheimer's that have no amyloid, right? So I've heard
Nick Jikomes 54:57
it goes in both directions, right? You have people who die with lots of the. Plaques that never had it, and then people who have it, they don't have a lot of the plaques. And this would sort of fit, at least with the idea that the plaques might represent, in some way, shape or form the body's attempt to fix the problem, rather than being the problem. But I
Lee Hood 55:14
think more and more, what's coming to be the view now is the idea that amyloid that Alzheimer's is a metabolic disease, yes, and it's a metabolic disease of the brain and of the brain's ability to handle and deal with cholesterol and things like that effectively. And if you look at it in that way, there become a whole series of reasonable things you can think about doing. And I will say again, at early stages, Alzheimer's is a disease that can be dealt with with diet and exercise really effectively, as as the first major successful trial, and Alzheimer showed us the finger trial out of Finland, and it showed both of these exercise and diet had a remarkable impact on the transition of high risk people into Alzheimer's.
Nick Jikomes 56:18
It seems almost like so in your description of sort of what's played out in the attempt to create drugs for Alzheimer's, you know, people were basically mistaking or misapprehending The relationship between correlation and causation. And they were trying to make a drug against something which a drug could be made against. And they were perhaps mistaking the body's attempt to fix a problem for the cause of the problem itself. It almost seems like there might be an analogy here between that and sort of what has happened historically with something like statins, cholesterol and cardiovascular disease, where people said, Ah, cholesterol levels are really high in certain people. This must be the cause. We can easily make a drug that lowers these things, and the drugs do lower that marker, but they don't always lower disease or death risk. Is this sort of just a general feature or shortcoming of the sort of drug developed mindset as it has traditionally existed where there is this sort of, perhaps myopic focus on looking at single, druggable targets, and not sort of thinking in a systems biology way about what the body is doing and why. I
Lee Hood 57:28
think that's exactly the explanation. I mean, for all of these complex chronic diseases by the time they're diagnosable, if you look at their related networks, you see that they become terribly disease perturbed, and that's why a single drug in cancer will almost never cure the cancer, because that's only one of seven things that are wrong. Yeah. So a really interesting question is, okay? That means we should have combinatorial drugs in the future, and the FDA actually makes it virtually impossible to think about doing that, because to get a combinatorial grub group, you have to do them individually and in pairs and so forth. So I think we're going to have to develop entirely new, multi dimensional drug strategies, multimodal drug strategies in the future, and how we're going to do it, I think is, is a really fascinating question.
Nick Jikomes 58:31
What are some of the emerging therapeutic biological agents that people are working on that are not of the small molecule drug class. I know, you know, people talk a lot about peptides, for example, what are, what are some of these alternative approaches to small molecule drugs? And how are they fundamentally similar or different?
Lee Hood 58:52
Well, I think an area that's going to be very exciting is going to be peptides as drugs in the future. And
Nick Jikomes 59:00
peptides are just basically small strings of amino acids. There are small strings of
Lee Hood 59:03
amino acids, and for example, in the field of proteomics, the study of all the proteins and the human body, one of the exciting new areas is, we've just in the last few years, discovered that there are more than 7000 small peptide proteins that average perhaps 30 residues in length or so, that play very important roles in communication and the translation of certain kinds of information. And it's these molecules that, in a sense, could be models as drugs that could have the ability to manipulate a wide variety of range of molecules and to be able to manipulate not only the through the interaction with other proteins effects, but through the interactions with nucleic acids and. Are all the different types of informational RNAs and so forth. So I think peptides are going to be a really important drug in the future.
Nick Jikomes 1:00:10
What do you think about the new GLP one drugs being used for weight loss today? I'll just keep it vague. What's your general take on them? Are? Are, I mean, obviously they do work for quite a large fraction of people in terms of the measurable output of weight loss. But do you think they're going to have downstream consequences that have been overlooked or that we don't know about yet? And you know, do you think you know natural, you know, sort of modified peptide drugs like that is, is that the future, or are there potentially going to be some downside risks that we haven't appreciated yet? Well,
Lee Hood 1:00:51
I think the GLP one inhibitors are absolutely fascinating drug and of course, they were initially created for weight loss, and they certainly did that. And I think what's striking is recently, a million person study was released to show that they've had remarkable effects on diabetes, and in fact, they're becoming a drug of choice for chronic type two diabetes and so forth. I think they they've had effects on drug dependencies, alcoholism and things like that, that have been very remarkable. And of course, the weight loss has been, has been really remarkable. I think they do have side effects that have to be monitored really carefully. And the most serious side effect is you lose probably as much muscle as you lose fat when you take these things. And for young people, you can get by. You should exercise and do those things. For old people, it can be a disaster. It can make you incredibly susceptible to falls and the things that can but the other thing is that it can cause an acute necrotizing pancreatitis, which is devastating for the people that get it, and that should be recognized immediately by physicians, and the drug should be stopped. I had, I had a friend that had to lose more than half his pancreas because of this necrotizing effect from taking from a doctor who didn't realize that his stomach ache was something more than a trivial kind of effect. So obviously, these drugs are, you know, they so they do have side effects. They have to be monitored. But the one thing for the weight loss people is an awful lot of people, when they go off, them bounce right back up again, and they're right back where they were before. So do you want to take them for the rest of your life? Or maybe you can take them at really minimal doses for longer? Those are all questions that remain to be answered. But it's a fascinating drug, and it does a lot of good things. Yeah,
Nick Jikomes 1:03:16
one of the things worth mentioning here sort of this gets to this sort of brings us back to the notion of systems biology. You know, historically, when you studied biology in school in a formal setting, you know, by our nature as humans who have to communicate using language, we like to parse the world into categories that we can label with our words. And so, you know, if you're a medical school or graduate school, you learn from a metabolism textbook, and then you have a neuroscience textbook, and then you have, you know, you have all of these different categories, immunology and so on and so forth. But obviously, all these things are fundamentally connected in ways. And you know, one thing that's sort of a hot area right now is the increasing recognition. And, you know, I think it's fair to call this more of a systems biology way of thinking, the increasing recognition that a lot of the things that we thought of as simply neurodegenerative or simply metabolic, they're actually fundamentally the same in their causes, at least in some cases. So for example, you mentioned earlier that Alzheimer's might be best thought of as a metabolic problem. And a lot of diseases that we think of as simply brain diseases. They are diseases of the brain, but they are sort of metabolic in their origin. They rely on fundamental metabolic systems that are common to all cells. And so why I'm bringing this up is the ozempic thing is really interesting to me. These GLP one drugs, because I don't know if you know about this example, but there was actually a weight loss drug used in Europe in the 2000s that was successful. It caused people to lose weight. It was called Ramona bond, and it was a cannabinoid based drug that antagonized the cannabinoid system. And sort of the punch line here is that it was pulled from the market because people were having suicidal ideation and severe depression and these neuropsychiatric issues. And. Basic idea is, well, if you antagonize the cannabinoid system, you will lower hunger and feeding, but you lower sort of the will to live, because the reward system can't be cleanly parsed into just food and these other little buckets. It's sort of all integrated. And the maybe pessimistic view of what's going on with these weight loss drugs that one might take is, yes, they are causing people to have less interest in eating food, which is having the desired side effect of weight loss, but they might also be having these neuropsychiatric side effects, and there's at least one study out that's starting to suggest that might be the case. Thinking from a systems biology perspective, maybe this isn't surprising. You can't cleanly parse the body into like a weight loss bucket and a neuropsychiatric bucket and so forth, because all these things are sort of integrated and connected. At the end of the day, what do you think about I mean, what do you think about that, generally speaking, Should we be worried that something that we think of as a weight loss or metabolic drug is likely going to have these sort of knock on psychiatric type effects, absolutely.
Lee Hood 1:06:04
And I think, you know, the million person study that we, we talked about now, it's, it's really relatively short term. We need a million person study that is five years out, and we'll see all of the other kinds of complications that can come up, and I'm sure they're going to be more I agree entirely with you. And you know where this multi disciplinarity being an essential feature dealing with the disease was expressed really clearly recently, is in long COVID, you could have long COVID that attacked virtually any of your major systems, and that meant it wasn't just neurologic, it wasn't just respiratory, it wasn't just GI or cardiovascular, it was all of those. So how do we get medicines? How do we get medical systems that have these integrated kinds of care? I think that's absolutely a critical kind of question. And again, going back to our earlier conversation on AI giving physicians the broad spectrum to be able to deal with the entire domain of medicine is really an exciting idea with regard to the requirement for multi disciplinarity,
Nick Jikomes 1:07:32
yeah, what? What are you guys doing at phenome health today? What are you working on? Is this purely a research organization. Are you building any like consumer facing technologies or products that people can use?
Lee Hood 1:07:45
So our purpose is, is really three or four fold, one, to generate the data we need that's going to give us compelling evidence for the power of this approach. Number two, it's to be able to take those data and analyze them in deep and effective ways, and that includes using the tools of AI as well as statistics and classic data analysis. And then three is to generate partners that we can transfer this knowledge to who have the patients that we'll be able to to be able to transfer this data driven approach to in a really effective manner. In fact, I've just been talking with a series of longevity clinical centers in places throughout the world, Singapore, Switzerland and so forth. And they're very interested in taking on these kinds of opportunities. And then three, we're developing the new technologies, both AI and measurement technologies that can let us look in whole new areas of data space or new kinds of information on the phenome to give yet more rounded views of a lot of the topics we'd like to analyze.
Nick Jikomes 1:09:20
What do you think about some of the new I guess you could call these sorts of systems biology inspired products out there that people can use to get get information about their phenome. So for example, I just talked to a scientist from a company called VIOME. And basically you can send in saliva, blood and or stool samples, and they measure, they do transcriptomics of the microbiomes. They don't just measure which microbes are present or absent, but they're measuring gene expression levels of all of those microbes. And this sort of gives you information that the basic product is they do that, and they tell you sort of nutrients or foods that you want to avoid completely. To eat some of or eat more of, and the idea is that you can use this type of data to achieve a site of precision nutrition, to achieve precision in which nutrients andor foods you as an individual should be eating. Are these things that you're excited about? Do you think these products are at a place where they really will work? Really will work and tell people something useful today?
Lee Hood 1:10:25
So I would say, one, the microbiome in that study is going to be an incredibly important part of health, and with our own studies, we can say unequivocally how your microbiome evolves in your 60s and 70s and 80s correlates beautifully with whether you're healthy and and so forth. So optimizing the evolution of your microbiome to be healthy is one interesting possibility. Number two, we've shown that a statin has two different effects. We have one, lowering LDL, which is a good effect, and two, it increases the incidence of type two diabetes, which is a bad effect. And we've been able to show beautifully in looking at the gut microbiome that there are different types of bacteria there that can facilitate or inhibit either one of those two kinds of effects. So the whole point is, drugs are filtered through your microbiome, and we have to be able to understand for all drugs in the future, what, what effect that's going to have? So what I would say is, we're at the very earliest stages. And I think people that make dramatic claims about what you can do from gut microbiome data, I think a lot of it, you ask them if it's clinically verified, there are very few things that are really clinically verified in a way a physician would feel comfortable, and I'm very skeptical of some of the claims many of the companies are making out there. And that isn't even to say they're wrong, it's just to say we're not there yet. We you, you can get people to spend a lot of money and not get much of an effect, if you really don't know about it. So it's an exciting area. A lot is going to happen in the future, but it hasn't happened yet. So what's
Nick Jikomes 1:12:34
an area that's sort of active for you, in the lab today, in your research lab, that you're particularly excited about and that you think we're going to make some substantial progress on in the near future.
Lee Hood 1:12:48
Well, one kind of diagnostic tool that we're very excited about is in the blood, there are exosomes, little vesicles that are snipped off from cells, and often snipped off more from cells that are diseased. And we're developing the techniques now to be able to quantify in individual episomes, both nucleic acids and proteins, and we've looked in very preliminary way at type one diabetes and shown a striking predisposition of the episomes and several individuals so infected with molecules that are critical in the pancreas. So we think they may turn out to be very powerful blood diagnostic tools for diseases and and the nature of what's packaged in the episomes may really give you insights into which biological networks In those organs have become perturbed.
Nick Jikomes 1:14:00
How much of, how much of the you know, this is sort of a vague question, but I would just love to get your general sort of take on this. How much of the chronic disease burden today do you think, roughly speaking, could be addressed by simply removing toxic stuff from the environment. So for example, if we cleaned up the food system, cleaned up the water, cleaned up the air, whatever exactly that means to you, how much of the chronic disease burden is coming from that rather than non environmental sources.
Lee Hood 1:14:36
I you know, that's an really excellent question that we don't know the answer to at this point in time. So I would argue that my own gut feeling is your own personal behavior is a far. Important determinant of your going into the major lease of familiar chronic diseases than the environment is. And we've we've seen some interesting examples in the measurements we made. For example, in the airfield program, we had four or five individuals with blazingly high levels of mercury. And of course, that is really lethal for brain cells and things like that. And it turned out that in three cases, it was cured by them stopping to eat tuna sushi in excessive amounts of the mercury in the in in fish, at the top of the, the the life scale and everything, yeah. And in the other one, she removed the filling. She had mercury fill, old Mercury album fillings, yeah. And that flipped her right around.
Nick Jikomes 1:15:59
So what are the symptoms of high mercury? Well, if
Lee Hood 1:16:03
you let it go, it they they mostly start to be neurologic things where you you're losing memory and things like that. So, so, but high mercury is not something you want to have in your blood. So there are things like that we can cure. What we're looking at now, which I think is really fascinating, is we're looking at the effects of forest fire smoke, on pregnancy, and there are really striking effects, both on the mother and on the fetus with regard to high exposure to smoke. So it's so the idea that when you're in those exposed areas, you've it's really essential to put on these n 95 masks that you can filter out these small particles and everything. Something
Nick Jikomes 1:17:03
I want to ask you about before we run out of time. You know you met, you mentioned that Henry Kissinger had written these books about AI before. I noticed while I was waiting, you've got, you've got quite a library here in your office. And it's not just science textbooks. What are some of the specific books or just general topic areas that that have been really influential for you as a scientist, that aren't science books.
Lee Hood 1:17:32
Well, I would say a really fascinating book was the 100 year life, and this was written by an economist and a sociologist about what would be the impact on society if we all lived effectively out to 100 years. And it discussed education. It discussed the fact that you'd probably have to learn to have repeated new jobs, because your jobs would, many people's jobs would? It talked about what retirement met. It talked about as, I mean, a really important aspect was, how are you going to live in your 90s? I mean, and all these and and the implication was that you would almost certainly want to be engaged in creative things through much of this, this life, the idea of you could quit would would certainly lead to complications, and of course, the implications of health care and so forth were just barely mentioned in there, which I thought was very interesting. But the book is absolutely a fascinating read. The Invention of Nature by Andrea Wolf is one of the best books I've ever read. And it's about, it's, it's a story of Von Humboldt and his interactions with Goethe, where Goethe said to him, the key to science is thinking big. And he was really the first systems biologist ever, because he did absolutely remarkable kinds of things in every way, and just seeing how a single intellect could transform the world of science. For example, He had 200 people from around the world measuring magnetic fields by a simple way. And he first showed that that there was an inversion of this thing over time. He was the first that recognized, are you
Nick Jikomes 1:19:52
talking about the inversion of the magnetic poles of the earth? That happens? He
Lee Hood 1:19:56
was the first one that measured how the. Uh, alpine vegetation changed as the elevation went and how stratified it was. He, he was the first that actually saw the the match of South America and Africa, and said the things that probably I mean, he just made room and and he was absolutely responsible for convincing Jefferson to buy the Louisiana Purchase, which transformed the US. I mean, here's a guy that, I mean, is just a remarkable individual. I mean, he was a scientist, but he was an adventurer and a polymath philosopher and polymath in every way. Yeah, yeah. So,
Nick Jikomes 1:20:48
so what? So do you think? You know, I mean, we've talked about some big topics here. You've had a very long career, but you're still clearly coming into work every day. Do you think that you're just going to continue working for as long as you possibly can. And and what does that what does that word work even mean to you? So think of it as work.
Lee Hood 1:21:09
I don't really think of it as work. I mean, it's a passion I've had to kind of decipher complexity and with phenome health, I want to be around when we see this revolution in healthcare that moves from the focus on disease to wellness and prevention, and that's going to be within the next 10 to 15 years, and I'd like to be here and see it all happen and everything.
Nick Jikomes 1:21:38
Do you have any thoughts on at all on So, so in my view, there's this sort of phenomenon playing out right now that has to do with technology and access to data, where, for all of my life, really up until very recently, and you know, for the way that things have been done for many decades, in order to become a doctor, in order to become a scientist, in order to become basically, a member of productive society, you had to obtain credentials through these institutions, right? You have to go to college and get a degree, then you have to get your PhD, or you have to go to medical school, or what have you. And the reason you need to do that is because the knowledge required for those types of professions. It requires you to be tutored and to learn directly from experts that have, you know, acquired the right type of knowledge. But of course, we're in this new sort of situation now with AI in the Internet where you know, you don't need to say, go to Harvard to get a Harvard education, you can really learn all of the same things that you would get in the classroom at any of these prestigious institutions, right online, for free. To a large extent. How do you see sort of the future of education and the credentialing system interfacing with you know, who can, who can and who can, cannot, cannot, or who will become, say, engineers, physicians, scientists. Do we need to be? You know, is it going to be necessary to actually have formal credentials? Or can people just acquire this knowledge for free? And sort of the proof will be in the pudding, in terms of what they can display, in terms of what they can build and what they know. So
Lee Hood 1:23:22
I think the answer is probably both. But what I will say personally, what was most valuable in my education at all levels were individuals that influenced me in personal conversation and and, for example, changed the nature of my future horizons and what I want to do and what I thought I could do. I mean, in high school, I went to a small, very small high school in in Shelby, Montana, 446 in my graduating class, but I had four of the best teachers I ever had. And you know, one said the most important thing you can learn in high school is to speak and think. And he was my debate teacher, and he was absolutely right in so many different ways. I think debate was maybe the most valuable single thing I did. But there was another who'd gone to Caltech as a meteorologist in the Navy who decided he'd send to Caltech smart people. And he had to talk for a year to persuade me to go, because I looked at the requirements and we had no calculus yet I had to take advanced math, which he said, you'll do, don't worry about it, and, and, and I ended up going to Cal Tech and that. So I think, I think hands on with people, is a critical, important part of the education, and you're never going to get that. I. Yeah, through zoom, and you can learn facts, but you can't synthesize and integrate the same way you can with deep conversations with really profound people. And at Cal Tech and at Johns Hopkins, where I went to medical school, I had really terrific mentors that just changed how I thought about so many things in important ways. And there are courses I could have learned by zoom. I do agree with that, but the insights and the passion and the commitment and the stretching the horizons, those are different things. Yeah, it's
Nick Jikomes 1:25:42
sort of interesting to think about this. So you're talking about the influence that individuals had on you and you interface with them, you know, in person, you just can't get that type of experience through your laptop at home. What's interesting to think about there is, you know, I would say I have a similar view. You know, individuals influenced me, individual mentors and professors and teachers that I had. And you know, even though the credentials I have, the degrees from the brand name universities that I have, or what are pointed to and what people recognize, a much more, you would know a lot more about me and how I think from knowing the lineage of thinkers that I directly interfaced with, but there's sort of no way to display that for people and to communicate that except through these credentials, these degrees that we have. Is, you know, as time moves on, do you think that will sort of evolve new credentialing mechanisms that take sort of a finer grained approach in terms in terms of telling you, like, who someone learned from. So instead of just I went to Harvard, you could sort of communicate to someone. I actually, I actually interface directly with this lineage of thinkers, which is very different from the other people that also went to Harvard. Yeah,
Lee Hood 1:26:57
absolutely. I mean, I think you can go to Harvard and get a mediocre education, because there are an awful lot of people there that are only interested in research, and they're not good my son went to Harvard and he had really mixed results there, I would say. But I think the idea of interacting with people is absolutely essential. And, I mean, it's the debate on, you know, everybody was home during COVID, and getting them to come back is is not simple, and it's even with the crew with phenome health, we have people in different cities and everything. And that has advantages, because you can get enormous talent that you might not be able to get here, but they're they're not. We don't sit and brainstorm the way those of us in Seattle are here, yeah, so it's, I think hands on touching is really critical to the evolution of people's thinking, yeah. And the evolution of thinking is different than knowledge. Yes, it's how you use that knowledge that's so important. Yes, and
Nick Jikomes 1:28:11
your computer of capacity for change and the way in which you change Exactly, yeah, yeah. Interesting. We've covered a lot of ground. Lee, is there anything that you want to reiterate that we've spoken about, or any final thoughts you want to leave people with to do with systems biology or sort of the future of medicine and biology, broadly speaking,
Lee Hood 1:28:33
I would just say that biology and medicine are in the most exciting state ever in my entire career, and I really look forward to the next 10 or 20 years at seeing revolutions that we can't even begin to imagine, part due to AI, but part due to Fundamental changes in society and their attitudes toward health,
Nick Jikomes 1:29:02
all right? Well, Professor Lee hood, thank you very much for your time. Thank you for inviting
Share this post