SIDDHARTHA MUKHERJEE:
We use the word generative AI, but, but it’s important to realize that humans are capable of generative thinking. And generative thinking is very important because it’s not just logical deduction. It’s not just putting two and two together. We uniquely have the capacity to be creative, to create, to generate ideas. And those ideas are unique in the sense that unlike any other species, those ideas affect our capacity of defining who we are. That’s what human beings are.
REID:
Hi, I’m Reid Hoffman.
ARIA:
And I’m Aria Finger.
REID:
We want to know what happens if, in the future, everything breaks humanity’s way.
ARIA:
We’re speaking with visionaries in many fields, from art to geopolitics, and from healthcare to education.
REID:
These conversations showcase another kind of guest. Whether it’s Inflection’s Pi, or Open AI’s ChatGPT, each episode we use AI to enhance and advance our discussion.
ARIA:
In each episode, we seek out the brightest version of the future and learn what it’ll take to get there.
REID:
This is Possible.
REID:
In our first season of Possible, we didn’t address human health. It’s an immense, important topic with seemingly intractable and far-reaching challenges. It’s unclear where it’s best to begin. You could start at the cellular level or launch into massive healthcare systems. There are regional, technological, policy, and financial layers and infrastructure to health. It’s a Gordian knot with everything entangled. How are we supposed to cut through?
ARIA:
We could spend a full season on everything related to human health, but we opted to divide it into two episodes for season two. One is focused on health and medicine, and another on disease and diagnostics. So now, these two aren’t mutually exclusive or collectively exhaustive, but they do portray two sides of the same coin: health and illnesses, and the products and systems we have built around them. And to your Gordian knot metaphor, Reid, can we untie it? Or can AI serve as Alexander’s sword and cut straight through it? So, let’s find out.
REID:
There are so many challenges in the medical field that humans have been trying to crack for centuries. Pick cancer, for example. It’s been around for 4,000 years, but doctors and scientists are still figuring out how to best treat it. Medicine is making great strides, but we have so many unanswered questions.
ARIA:
But I am confident that AI is going to help us break through the hardest and most urgent problems around disease. I mean, just think about it. Take Paxlovid. Using modeling and simulation, scientists used AI to not only decide which components to make, but also help find the molecules that could deliver Paxlovid in pill form versus intravenously. AI helped reduce computational time in this drug’s development by 80 to 90%. That’s amazing.
REID:
Or how about CRISPR? It’s a gene-editing technology that can do everything from treat sickle cell anemia to engineer more delicious mustard greens. AI can help predict the activity of RNA-targeting CRISPRs, which holds promise for developing new methods to prevent or treat viral infections.
ARIA:
Our guest today is one of the experts at the forefront of medical discoveries and developments like these, especially around cancer research. And he’s here to help us rethink how we understand human health and the ways that AI can help. Siddhartha Mukherjee is a cancer physician, researcher, and author. His 2010 book, The Emperor of All Maladies: A Biography of Cancer, won the 2011 Pulitzer Prize for General Nonfiction. His most recent book, The Song of the Cell: An Exploration of Medicine and the New Human, is also a must-read. Reid, what I loved is it’s so obvious that you and Sid have been tossing around these questions and having these conversations for years and years. What stood out to you most about the discussion with Sid that everyone’s about to hear?
REID:
I think it’s the combination of deep scientist and kind of data with humanist, with a question of—you know, part of the reason why, for example, I wanted him to make sure he talked about art some, and he was like, “well, but you know, I don’t want to be stealing,” you know, kind of like, “my amazing Sarah Sze partner who is so much better at this.” But it’s like, yeah, but you, you care about art as much as you care about science. And as part of how you are a humanist and you care about humanity. And that reflects everything through when he goes super deep on technology, to thinking about AI, to thinking about like, okay, well, what are the ways we should think about who we’re becoming as humanity? And it was just, I hope that we can even, we’ve even managed to capture even a small portion of that, that joyous humanity person in him. So, Aria, for you, what was something that really stood out from what Sid said?
ARIA:
So at the top we talked about how we didn’t sort of tackle healthcare head-on in season one, but I love today how Sid’s conversation hit back to almost every episode we had. And in particular, I mean, he talked about how AlphaFold—which was created by Mustafa Suleyman, who we interviewed earlier last season—how he uses it every single day as part of his cancer research and breakthroughs. And so it is really lovely to hear him as a humanist and a scientist, but truly, it’s all being woven together about how we can reduce disease, improve longevity, and, you know, ultimately improve humanity.
REID:
We sat down and talked to Sid about everything from CAR T-cell therapy, to the future of humanity and moral philosophy, to the future of medicine. Here’s our conversation with Siddhartha Mukherjee.
ARIA:
Sid, thank you so much for being here today. It was just awesome revisiting, you know, all of your books and articles, et cetera, that I had read throughout the years. And you had once said that you did your medical training in reverse. You started training as an immunologist, then a biologist, and then you got your medical training. So how did you get here today?
SIDDHARTHA MUKHERJEE:
So I am, obviously I’m an oncologist by training. I treat cancer patients. My specific area of interest is blood cancers—leukemias, and lymphomas. But really my interest is in novel therapeutics. I develop new therapies. I make new drugs, as I like to tell people, and that is actually the main area of my interest. And that spans not just leukemias and lymphomas, but also breast cancer, immunology, a huge effort in osteoarthritis, in cartilage regeneration, and blood stem cell regeneration—all of which seem totally disconnected, but actually are deeply connected. So your question is, how did I get here? I just, you know, I’m not going to repeat a big résumé, but I started off as a student at Stanford, where I had the opportunity to interact with Paul Berg. Paul was the person who created recombinant DNA.
SIDDHARTHA MUKHERJEE:
He was the person who basically invented the technology that allows us to take two pieces of DNA from two completely different creatures—or even an artificial piece of DNA—and stitch them together into one single piece. And that allows us to do all sorts of things, including cloning. I then went to Oxford, where I trained in immunology—another big milepost—where I trained with Alain Townsend. From Alain, I picked up the idea of, you know, that medicine and science, biology, are sort of intricately linked, like two strands of DNA. A yin and a yang. Then I came to medical school, where I actually began to sort of really explore these ideas in more detail, worked as a postdoc in David Scadden’s laboratory at Harvard Medical School. That’s the non-résumé part of my history, because it tracks the intellectual journey that, that, that I, that I took.
ARIA:
Is there something that excites you most about cancer research right now?
SIDDHARTHA MUKHERJEE:
There are many things that are exciting about cancer research. I’ll just lay out a few of them, and I’ll try to organize them a little bit so that we can understand them. I mean, so the fundamental basis of cancer treatment research—or whatever you want to call it—is prevention, cancer prevention. And it’s been traditionally underfunded, but of course it gives you the most bang for the buck. And for a long time, we didn’t understand some fundamental features about cancer prevention. Namely there was possibly a universe of cancer-causing agents that we weren’t trapping. These are agents that people have been thinking about for a long time—inflammatory agents, things that cause inflammation. But inflammation is a kind of a clown-car kind of word. You know, if you open it, all sorts of various things come out, and you don’t know what’s, what’s real and what’s not real.
SIDDHARTHA MUKHERJEE:
So we have begun, and by we, I mean the scientific community. There are two or three recent important papers—one from Charles Swanton’s lab in London, a couple of others from other laboratories around the world—where we have begun to unpack a new understanding of how the link between inflammation, possibly between inflammation and obesity, and cancer. Because there’s been always this suspicion that somehow we were missing a series of very important human carcinogens, and we didn’t know how they worked. I liken this to a little bit, you know, in this typology, I would say that these are unknown unknowns. So cancer, you know to find carcinogens, it was very easy to find known unknowns. So you know, the mechanism. You know that they cause DNA damage, they cause mutations. But you don’t know the chemical.
SIDDHARTHA MUKHERJEE:
So that’s a known unknown. You know the mechanism by which it works, so you can find the chemical. But then there’s a universe of, as Donald Rumsfeld famously put it, of unknown unknowns. And these are things where you don’t even know how the, how it causes cancer, but there’s an epidemiological link. So for instance, air pollution was an unknown unknown. We knew that there was an epidemiological link. People who happened to live in highly air-polluted areas, carbon or coal-car-exposed areas, got cancer. But no one knew exactly how and why these things cause cancer. We are beginning to understand that. How does obesity cause cancer? Again, we’re beginning to understand that. How does diet relate to cancer risk? We’re beginning to understand that. So, the whole of cancer prevention is now moving towards these really mysterious unknown unknowns.
SIDDHARTHA MUKHERJEE:
And I think that’s super exciting in cancer prevention. Now, in early detection—which is the middle of the pyramid—there’s a whole new series of things that are going on, some of which involve AI. So, can we use new technologies, potentially generative algorithms, to figure out people who are at higher risk? Do we screen them differently? Do we find early cancers? Does that help? Big, big, big questions that will be answered in the next decade or so. In the field of cancer treatment—that’s the area that I work most closely in—I mean, the exciting thing, of course, is the use of, I would say not just the immune system, but the body’s entire natural physiology to act against cancer. That ranges from the use of the immune system—obviously new medicines that unleash or reactivate the immune system to recognize cancer CAR T cells, so T cells that have been genetically engineered to kill cancer. It’s a field that I work in very closely. The use of diets in combination with drug therapies to potentially attack cancer cells. The use of new technologies—including technologies that have been fast forwarded by AI, to find new medicines, new drugs against cancer. And so all of that, I think, in the arena of cancer treatment, is where there’s a lot of exciting stuff going on.
ARIA:
And one of the things that you have been very concerned about is making sure everyone benefits from the great, you know, health, medicine breakthroughs that are happening. And so you mentioned CAR T-cell therapy. Would love to hear just a little bit more about what are you doing there? What is it about sort of democratizing access to this, especially in your work in India?
SIDDHARTHA MUKHERJEE:
Yeah, so I founded a company called Immuneel, which is going to bring the first CAR T therapies to India. CAR T therapies are basically genetically modified T cells, which are then unleashed against cancer. It’s incredibly complex technology. You basically have to extract T cells from the body, genetically modify them using a virus, and then return them to the body and weaponize them, and return them to the body now to attack cancer cells. You can imagine that this technology, you know, obviously evolved out of very many incredible discoveries here. But we, I and two of my colleagues, decided that we would try to democratize it. And, you know, these therapies cost, cost around—or I should say, are priced—around about $500,000 in the United States. And, as a friend of mine once said, “half of infinity is still the infinity.” So it’s infinity for someone in India, a child in India.
SIDDHARTHA MUKHERJEE:
And so we decided that we would try to figure out and cut every one of those costs. And now I’m talking about costs, not price. And we’ve been able to bring the technology to India. That was the first part of the journey. We’ve treated about 25 patients so far with astonishing data. Cured children who would otherwise die of leukemias and lymphomas. In fact, our cure rate is exactly the same as the cure rate that CAR T’s have in the United States. This is for very particular leukemias and lymphomas. This is the first CAR T—we have several in the pipeline—but really the trick has been to create a sense of self-confidence. If you can send a rocket to the moon from India, 1.3 billion people, we can make a car T. And we have. So, and you know, there are CAR T’s now available in China. They’re available in several other countries. And so one of the efforts has been to use all the tools available—especially, you know, the kind of hacking engineering that we can really have talent in in India—to reduce the cost and increase the accessibility. So it’s a journey. We’ve just started it. But it’s probably one of the most exciting things. I met one of the children that we, we cured, and it was probably the most important medical moment in my life.
ARIA:
That’s amazing. Thank you for that work.
REID:
Wow. Awesome. In addition to being in a journey about understanding cells—what they mean in the human body, how you can trigger them, how you can evolve them. We also, of course, have an artificial intelligence part of, you know, your gesture. One, general reflections on AI, but also with reflections to medicine and cancer?
SIDDHARTHA MUKHERJEE:
Sure. It’s a little intimidating to talk about AI to you and Aria, because, you know, it’s like bringing coals to a Newcastle. But anyway, I could give some general reflections, but let me focus later, more importantly, on medicine. So my general reflections on AI, of course, is that it’s one of the transformative technologies. It’s one of those accelerators. I have absolutely no doubt that it’s going to be, and has already started becoming, showing itself, as a— it has all the signs, it may not have achieved all of its potential, but it has all the signs, of it being one of those accelerators. I often think about, remember the game Snakes and Ladders?
SIDDHARTHA MUKHERJEE:
So you know, I think of accelerators as ladders. So all of a sudden you are sort of trudging along. You know. One-two-three-four-five-six, and you land on seven, and there’s a ladder, it takes you to 18. So it may not have achieved all of its potential, and it certainly has not, but it has all the signs, the smell, the feel, of one of those accelerators. Recombinant DNA, an accelerator. Language, of course, perhaps the most important human accelerator that ever existed. But it has, you know, it has, it has the capacity, I use those two examples very particularly. Language has a generative component. It, you know, language, as I often say, we think in language, but then language dictates how we think. And it, in fact, many people have pointed out that this recursion is extraordinarily important.
SIDDHARTHA MUKHERJEE:
AI has that feel about it. Recombinant DNA I also used, again, as a very particular example, because recombinant DNA allowed us to change who we were. Again, a very recursive quality about it. AI has this recursive quality. We’re making things that change—again, to go back to my fundamental definition of humans—we’re making something that will change who we are. AI allows us to make those things faster. And, in fact, itself is something that will change who we are. And so it has this recursive quality. That’s why I chose those two incredible accelerators, but as those two particular examples. AI in medicine, okay, well, thus far, very early days. There’s a lot of hype around it, but very early days. Recently a flotilla of papers—one in Nature, one in various other major publications—that talk about sort of, you know, basically absorbing and beginning to curate medical knowledge, which would be just the first step.
SIDDHARTHA MUKHERJEE:
So, in other words, you know, things that we couldn’t do before, which is using AI as a diagnostic companion, because it has absorbed virtually all of medical knowledge. Very important step in terms of diagnostics. So, you walk into an emergency room with a rash, and, you know, you are, the physician that is there has a certain set of medical knowledge. And all of a sudden that has now become not just quadrupled or quintupled, but now the entire archive of medical knowledge is available to him or her. And, and so obviously, one step, and we’re moving in that direction. But in terms of what’s exciting to me—and we can talk more in depth about that—is not just the capacity to ingest the entire corpus of knowledge and then incorporate that corpus into diagnostic decisions, into medical decisions, but the generative aspect of it.
SIDDHARTHA MUKHERJEE:
Which is to say, “can we make something completely new in medicine based on, on AI alone?” And there, I think we—and I’m actually involved with some of this myself—we’re doing some things that we haven’t ever done before. So, for instance, I’m involved in an effort, we’re trying to make drugs completely out of computation alone. So the old way of making drugs was to use basically empiric ways, or use nature as your, as your guide. And, you know, highly successful. Aspirin comes from nature. Incredible drug. But then came a time when we were using rational drug design and some synthetic chemistry. So, a great example of rational drug design and synthetic chemistry is the drug Lipitor. So, you know, the pathway by which heart attacks—myocardial infarction—occurs was tracked down to elevations in cholesterol, in particular, you know, by Brown and Bishop. That then led to the identification of the receptor.
SIDDHARTHA MUKHERJEE:
And the, and Lipitor is a mechanism by which, you know, we can change the dynamics of how much cholesterol is in your—the harmful kind of cholesterol—is in your blood. So that is an example of rational drug design. We are in, we are entering, potentially, a third phase in terms of drug design in which we don’t even need some of these empiric tools. In other words, we can design a drug potentially completely using computation alone. And generative AI is an incredibly important component to this. And the first such drug, when it appears, will change the entire paradigm of drug discovery. Because we’ll be able to use, presumably, generative AI to, to create not just one drug, but a universe of drugs. And just to help you understand the problem, AI has already solved—as you know, with the help of computational biology—it’s already solved our capacity to fold proteins in space. This was, this is a huge problem, and this was taken on by DeepMind. The program is called AlphaFold. It’s freeware. I basically use it every day.
REID:
Pi, could you explain?
Pi:
Well, to put it simply, AlphaFold is a revolutionary AI program that can predict the 3D structures of proteins. This is a really big deal because knowing the structure of proteins is crucial for understanding how they function. And that, in turn, has all sorts of applications in fields like drug discovery and medical research. Before AlphaFold, predicting protein structures was a really time-consuming and difficult process. But now, it’s much more efficient and accessible, thanks to this technology. It’s one of the most important advances in AI in recent years, and it’s already having a big impact on science and medicine.
SIDDHARTHA MUKHERJEE:
But that solves only half the problem, right? So you’ve got the protein that you want to target, but now you want to take a chemical—a small molecule or another protein—and you want to dock it into the first one. So it’s a, you’ve got the lock, but now you have to find the key. And again, we are beginning to use AI. AI helped us define the shape of the lock, but now AI is helping us define the shape of the key. And once we find out how to make the key, there will be a whole universe of materials—I’m not even just talking about drugs—a whole universe of materials that can interface with biology. Whether they be nanomaterials, whether they be proteins, new chemicals, small molecules, whatever it might be—that is a ladder in Snakes and Ladders, and certainly a major ladder for medicine. So, you know, I think that’s what’s exciting to me. Both the, both the knowledge base and the corpus, and then, of course, the generative aspects—making new things that weren’t there before.
REID:
When people think about which risks should we be taking to make acceleration in science, to be making acceleration in medicine—because the normal kind of unfortunate dialogue is to say, “well, we should just get all that without any risks,” and you can’t get it without risks, you can’t get it at all, you can’t get acceleration—which risks do you think would be wise for people to consider taking, like that we as, as humanity, as scientists, should lean into taking more?
SIDDHARTHA MUKHERJEE:
I think that the lack of public sharing of medical data. And I think we’ve leaned a little bit strongly towards privacy. And in fact, there is a free lunch here. I think modern cryptography has achieved enough that it wouldn’t, it could make it really, really tough—not impossible, but really, really, really tough—to de-encrypt private information. But the sharing of that information has become incredibly difficult. You know, we can’t do that across genetics. We can’t do that across risk. You know, AI will soon allow us to mine medical information extremely well. Even unstructured medical information, medical notes. We may be able to figure out who’s at a great risk for a chronic kidney disease in the future, or, or, you know, losing their eyesight in the future. In order to do that, we have to be able to access these unstructured medical notes, which is an incredible corpus of knowledge. And, you know, someone like you, I mean obviously would realize that the only way to do this is information and the sharing of information. So I think that’s absolutely crucial.
ARIA:
One of the places this is going to take us is in Song of the Cell, you spoke of cell manipulation, and you talked about creating, sort of, the new human. And I was excited about one of the examples you gave because you talked about bone marrow transplants. In my previous work at DoSomething we did bone marrow drives where we got college students to swab their cheeks so they could enter the bone marrow registry. And, you know, we saw hundreds and thousands of lives being saved. You know, one of my very best friends donated bone marrow to a seven-year-old. And you were talking about how, “yeah, this seven-year-old is now a new human because they have DNA from the bone marrow from someone else.” But that’s like child’s play when we’re talking about bone marrow donations. Like, what do you define as the new human? And, and where could it go as we’re seeing AI enter the equation?
SIDDHARTHA MUKHERJEE:
Well, I think part of the reason I used, I used the word “new human” in this book as a provocation because I wanted to provoke the idea that, in fact, we’ve been making what I call new humans all the while. And, in fact, every time we’ve made them, there’s been a specter of dismay and extraordinary amounts of caution and extraordinary amounts of alarm. And perhaps that’s a good thing. These are, these may be helpful things because they allow us to calibrate—they are real stats. They allow us to calibrate, you know, the reach of technology with the potential important pieces that stop technologies. And so I’ll give you one example: When the first blood transfusions were attempted, some people thought that, you know, the person who would be, who would come out of that blood transfusion would be a different person—because, you know, this was in another time, when blood transfusions were less successful.
SIDDHARTHA MUKHERJEE:
A more important example: When the first baby was born through IVF, Louise Brown, you know, Louise Brown’s parents got hate mail saying, you know, “you have, you have basically changed the mechanism of human reproduction and thereby changed what it means to be human. Louise Brown is not human.” And yet she is, you know, there she is being as human as she’s now, she has children herself. And there are a hundred thousand, more than a hundred thousand, several hundred thousand, babies born with IVF. And you don’t even know. Similarly, when the first bone marrow transplants were, were done, people said, “well, you know, you’re making some kind of bizarre chimera between one human and another.” And now, you know, we walk through a hospital ward, and there’s some people with bone marrow transplants and some people without bone marrow transplants.
SIDDHARTHA MUKHERJEE:
There are people walking around with a kidney, with the heart, from another person. So, you know, in the biological sense, we’ve been creating these new humans for a while. And so I wanted the word “new human” to be a provocation for people to begin to understand that this is not, sort of, suddenly today’s news. This news has been around for a while. We’ve dealt with these ideas for a while. And in fact, I would say we haven’t dealt badly with them. We’ve dealt with them with caution. We’ve created boundaries. We’ve created barriers. We’ve been generally okay, I would say. Our record isn’t, may not be an A+, but it’s certainly not a, not a C-, in terms of our capacity to regulate, to understand what the limits are, what the boundaries are, in this. Now I’m not an expert on how far AI can push these boundaries and make them basically—push them beyond our control.
SIDDHARTHA MUKHERJEE:
And I’ll flip this question back to both you, Aria, and Reid, in terms of answering this question. But for me, what the history of biology, and thereby the history of human intervention technologies, teaches us, is that we haven’t—our, our track record isn’t that bad. And so I’m, I’m hopeful. The alarmist concerns that we have around how AI will come and take over everything can be mitigated by our, our understanding of what, what we have done in history. I don’t know, I’ll flip this question back to you, Reid, and back to you, Aria.
REID:
I’d say that the, that the question around thinking about, you know, what is—humanity will become, like, this is one of the parallels within AI. It’s like when the printing press came out, people were like, “oh my God, this is going to destroy, you know, certain key elements of humanity. We’re going to be eroding our cognition because it makes us less dependent upon our memory. You know, it’ll create misinformation that’s, that’s floating around.” You’re like, well, that sounds like a lot like what some of the people are saying about AI right now. And yet, I think it’s hard for us to imagine the humanity that we, we have become and aspire to be collectively, you know, all eight billion of us without books, without, without written things, and how books kind of evolve us. And then the question, of course, is—is as it gets closer to, to not just the tools we use, but who we are—like, you know, genetics and so forth—it becomes a lot more, you know, kind of challenging to think through. And I think we will want to evolve deliberately our genetics.
REID:
It starts easily, of course, as everyone’s noted with like, well, let’s fix key highly dysfunctional genetic diseases. Like, why have those as, as suffering? But then it gets to, kind of, questions around like, well, what if you had genetics that allowed, you know, kinda higher-on-average IQs, or higher-on-average longevities, or other kinds of things? And, and then people argue against it by saying, “well, but are you going to start choosing blue eyes?” And you’re like, well, we do have that bad human instinct too. We’re going to have to try to keep that and keep all kinds of diversity in what we’re doing. We’re going to have to navigate all that. But that, that evolution is part of what, who we are as human beings and what we need to do. And so I think it’s, it’s important to address, and not just kind of back into. But I don’t think we’re going to be able to fully address it until we start kind of dipping our toes into the water. I don’t know, Aria, if you think something differently.
ARIA:
Well, I mean, I remember, Reid, we were, we were at like a ten-person dinner, and we were going around the table talking about, you know, if you could right now get ten extra points of IQ, you know, would you do that? Would that be ethical? And I think all these questions are so tough. And, and Sid, you know, you yourself talk about the difference between desire and disease. And like, certainly curing disease is okay, but even now it’s so interesting, you know, when you talk about kids getting cochlear implants, and then, you know, the deaf community is upset because they feel like a piece of their culture is going. So what is interesting to me about it is sort of the intersection of culture and what is human. Like, will different cultures and different societies come out differently on this because something is more, more important to them? But I definitely think that the next decade, at least, we’re hopefully going to be dealing with diseases and curing them. And then, and then the desires will come. And I have, I have no doubt that we’re going to be doing things in 15 years that today just seem beyond the pale and in 15 years seem pretty normal.
SIDDHARTHA MUKHERJEE:
Yeah, I’ve talked a lot about this in the two recent books, The Gene and The Cell—this distinction between disease and desire. And another way to put it is the distinction between enhancement and, and emancipation. So emancipation from terrible things that happen to us, enhancement being going somehow beyond emancipation, doing something that is, that is, that is desirous to us. And the, I suppose, the good news is that genetic technologies, at least, and cellular technologies, are pretty limited because of biological constraints based on everything that we know. You cannot simply, aside from raising the hideous prospect of eugenics, you cannot simply genetically manipulate yourself into having even blue eyes because that trait is controlled by a hundred odd genes. Not that particular trait, but most complex human traits, including most complex human traits that range from, I’ll give you one example—human height is highly genetically determined. Tall people, tall tend to produce tall children. Short people tend to produce shorter children. So we know that there’s a genetic component, but if you dissect the genetics, it turns out to be hundreds of genes. And this has become a pretty understandable pattern. Most human traits are controlled by a hundred, if not several hundreds of genes. And it’s not easy to manipulate several hundreds of genes. It’s just not technically easy. It may become at some point in time. The difference, I think, between all of this and computational technologies is that you can enhance without manipulating genetics. So you can enhance, and I would say you can—and if you want to put it in, sort of, using Richard Dawkins terminology—you can change the meme without changing the gene. In other words, and you can potentially change cultural memes without changing human genetics.
SIDDHARTHA MUKHERJEE:
In other words, your children will inherit it. Well, they will inherit the idea, but they won’t inherit it through biological mechanism. They’ll inherit through a cultural, social mechanism, et cetera. That is what is, to me, both fundamentally, extraordinarily interesting about some of this, some of the, some of the universe that’s going around in AI and its intersections with genetics. But it’s also simultaneously quite frightening, because you know, I hear the voices. I think the voices are real. I don’t think, you know, I may be an optimist in general, but I, you know, I don’t think that I’m completely sanguine about this interaction, this social genetic interaction, the gene environment interaction that will almost certainly be disrupted by, by AI. You know, gene environment interactions are crucial to the, to us. To humans.
SIDDHARTHA MUKHERJEE:
Now, AI is, you know, often people use “disruptive”. To me, what’s interesting is that AI is going to disrupt human-gene environment interactions. And by that I mean so many different things. I’ll give you a very, a very colloquial example as it were. You know, AI has the capacity to disrupt, you know, one of our many gene-enviroment interactions is our conception of, of what is beautiful. Some of that is, is determined by, you know, social memes. Some of it is just determined by history. Some of it is determined by genetics. Some of it is determined by, you know, a whole variety of things. AI has the capacity to disrupt that idea of what is beautiful. And so in, in a way, that’s what bothers me.
SIDDHARTHA MUKHERJEE:
What bothers me is AI’s capacity to disrupt, not just generally speaking technology, but the way humans are. You know, going back to this idea that creativity—if AI starts competing or disrupting or squashing the human creative impulse, I think we’re in deep, deep trouble. On the other hand, if AI causes an efflorescence in human creativity, then we’re not in deep trouble. I think it’s going to be very beautiful to go where we’re going.
REID:
Well, one of the things that I think is going to be the structural deep rub here is the need we have as a tribe and as a species to have broad-ranging experimentation and taking some risk, together with individual desires to having controlled and contained downside.
SIDDHARTHA MUKHERJEE:
The big concerns here are maintaining diversity, empathy, equality—many things that define us as humans. Not just the, you know, sapiens, as I say, remember, is not just a technological term.
SIDDHARTHA MUKHERJEE:
It also involves, and I said this right at the outset, it also involves this whole range of creative processes that we, that we can do and, and are able to do. And this range of creative processes is possible, I suspect, because of these other qualities that we have. You cannot disrupt it, you cannot take the, you cannot take empathy out of the creative process. You can’t write a novel, I think, without having, you know, without having empathy or, or generating empathy in characters. I know Reid, we once spoke about, you know, one of your criteria for reading a novel is, within the first few pages, can you, do you have empathy for one of the characters? And if you do, you continue to read. Maybe I’m very broadly speaking. You can’t write a novel without the feeling of empathy.
SIDDHARTHA MUKHERJEE:
You cannot write a book, you cannot create art without appreciating that human beings are different. By the way, this is one of the reasons, I think that, one of the reasons that fascist art turns out to be bad art. But you can’t you can’t create the most beautiful things we create without understanding that, you know, you and I are different. And you may, you might hate something, and I might like something. But anyway, but I think that’s the rub. The rub is, you know, can we continue to be homo creativists, and will that be enhanced, or will that be somehow disrupted by AI and, potentially by genetics, by biology, biological intervention? And if either of those two happens, and especially if those two happen in collusion, I think there’s, there’s a prospect of very bad things happening.
SIDDHARTHA MUKHERJEE:
And on the other hand, you know, of course, if there’s an efflorescence, that would be a wonderful thing.
REID:
I 100% agree. And that reminds me of another thing that we cannot let this podcast go without asking you, what is humanity? What are we now? What are we turning to be? What are the levers? You know, one of the things, as you know, Sid, from our conversations, you know, as, I and we tend to say, “Hey, we’re, we evolve through our technology.” You know, whether it’s, you know, clothing, glasses, fire, or medicines and so forth. Because I think one of the things that I love about you is you are, it’s many things—is a scientist and a humanist. How would you put the, kind of, from your world perspective, science perspective, that question of what is humanity? What journey are we on? What are the ways that we are, you know, kind of, evolving? Not in the purely just genetic, you know, adaptive sense, but in the, this question of bringing up with recombinant DNA, bringing up with science perspective, bringing up with technology, what’s that, what’s that lens forward on humanity?
SIDDHARTHA MUKHERJEE:
Yeah, I mean, it’s a deep question, Reid, and you and I have talked a lot about it. You know, I think that the, it’s not a coincidence that our species is homo sapiens, we concentrate a lot on the homo but really should be concentrating a lot on the sapiens. The sapiens being we are a species that is capable of a particular kind of thinking. And I think that particular kind of thinking is, now has become, a word in, in common use.
SIDDHARTHA MUKHERJEE:
You know, we use the word generative AI, and we’ll talk about that “I”, but it’s important to realize that humans are capable of generative thinking. And generative thinking is very important because it’s a particular kind of thinking. It’s not just logical deduction. It’s not just putting two and two together, but it’s the capacity to say, “here’s fire. I’m going to use fire, which is hot, to make food. And by changing the quality of food, I’m going to change the quality of what I produce, and I will cultivate a new kind of crop, which I couldn’t previously eat, but now I can eat because I can transform it using fire.” Now, that seems like a trivial example, but it totally changed the course of human history. There were foods that we could not eat, which, with the use of fire, we were able to eat all of a sudden.
SIDDHARTHA MUKHERJEE:
And those were the very foods that we could cultivate. And once we could cultivate them, we could create homes around them, and those homes became villages. And, you know, I’m sort of recapitulating a particular kind of history, but it’s a generative history. It’s, again, it’s not, it’s not just logical deduction. It’s the fact that we are, you know, we have a creative impulse inside us. The question you asked is very important. Now that you, now that we understand that, you know, that the sapiens part of it is, is a, there is a creative or generative impulse to make new ideas come around, the question is, what are humans and what do we look forward, what we look forward to? Well, one thing that humans are becoming, and perhaps are definitionally who we are, is that we are a species that uniquely has the capacity to control our destiny, potentially by interfering with our own evolution.
SIDDHARTHA MUKHERJEE:
We are a species that, unlike any other species that we know—and maybe we’ll encounter one in some other planet—but, unlike any species that we know, has the capacity to drive our own evolution. In other words, we have a recursive process, uniquely recursive process, which allows us to do things to ourselves that change ourselves. That spans the gamut of technologies, from recombinant DNA—changing our genomes, changing the genomes of ourselves, changing the genomes of crops around us—all the way to generative AI, in which we create new ways or new machines that can come close to, or are coming close to, the way we generate our, our thoughts. So in some ways, that is becoming the definition of who we are. Of course, there’s a biological entity of who we are. And it can’t be ignored. You know, there will be people on your podcast who will absolutely talk about how important it is to not neglect those biological impulses.
SIDDHARTHA MUKHERJEE:
You know, we are also hormonally driven, four-limbed creatures that you know, know, are capable of love, and envy and passion, and terrible crimes and incredible inventions. There’ll be people on, and that is absolutely true. There’s no disagreement there. But I think, what my perspective on this—which perhaps is, you know, which I share with you—is that we uniquely have the capacity to be creative, to create, to generate ideas. And those ideas are unique in the sense that unlike any other species, those ideas affect our capacity of defining who we are. So it’s almost a recursive definition that no other species have. We make ideas that change who we are. That’s what human beings are. And that’s simultaneously the most exciting and also the most dangerous thing that we can do to ourselves—to make ideas that change who we are.
REID:
And, and part of—I, obviously we’ve talked about this a lot, I completely agree—there’s accelerants, but it’s what we’ve been doing already as the, you know, history recapitulation. And so it’s not, it’s not like, oh, suddenly, it just, the accelerant. Now sometimes, you know, differences of degree are differences of kind, but it’s kind of the question about how to navigate it. And it’s one of the things I think is, if I would guess, what would be the next, the most important thing vis-à-vis technology over the next 20 to 100 yearsm it’s how are we doing that generative self-evolution in this? And it applies at the individual level, applies in society level, applies at the humanity level, applies at the planet level. The kind of scientist-humanist is a very important voice in this kind of shaping, because we’re not going to get, we’re going to get, you know, all kinds of, you know, humanity is a cacophony of perspectives.
REID:
And it’s like, well, which of these voices do we orchestrate into the, into the chorus, into the song of, of who we become?
SIDDHARTHA MUKHERJEE:
I think Reid, you have a particularly important perspective on this as well, because, you know, I think that the word is exactly right. So there’s a cacophony of voices here, and we need to separate out from that cacophony the parts that fit into the bigger song and the parts that should actually probably be left behind from the bigger song. But the important pieces that there are, I think there are, there are, I would say, four or five important parts of that song. In other words, I’m, I’m playing a game. I’m playing a game in which, you know, a spaceship is cast out into space, and you have, you can put five members on it, and they’re going to discuss, they’re going to be concerned with discussing, the future of human beings.
SIDDHARTHA MUKHERJEE:
Who would you choose and who would you chuck from that spaceship, right? So here are my choices. I’d love to hear yours, but here are my choices. My choices, we need a pure humanist—probably a philosopher, a moral philosopher. We need a historian—because someone needs to remind us what, where we have to go, well, where we’ve come from, sorry. We need, and the third person is, I would say we need a pure scientist. Presumably a biologist, a computer scientist, or some of some sort or at least a pure scientist. I’m just sort of using the term abstractly. But then there are two other people that I would put in that ship, and they are, I call them translators. The first translator is a scientist-humanist, someone who can speak the two languages.
SIDDHARTHA MUKHERJEE:
And the second is, is another translator, a technologist-humanist. And they may be the same person. But we need two translators. Because the pace at which we are moving, it seems to me that we need to at least understand this pentagon of dimensions. Where did we come from? Which is the historian’s dimension. What should we be doing? Which is the moral philosopher, the humanist idea hopefully tinged by an understanding of human history. Number three is what capabilities, what does science look like? And remember, I, you know, the—I’ve, I’ve embodied in the, in the pure humanist and the historian, the wealth of literature and music and, you know, the wealth of art. I’ve, I’ve tried to embody them in, in those. So please don’t, you know, as you very well know, my wife’s an artist. I’m not excluding someone who will, who will make cave paintings.
SIDDHARTHA MUKHERJEE:
Then we need someone to tell us what science looks like. And, you know, in that person is embodied, perhaps the physicist, the tech, you know, computer scientist. And then we need two translators who will, who will basically make peace between the three kingdoms. And so that, so that would be my idea of who to put on that spaceship.
REID:
Well, it’s an excellent question. You know, part of the thing that you and I have also talked about a bunch is how questions of, like which questions you ask really shape the journey and is actually kind of the important progress in science and everything else. So I will also, you know, kind of essay a response. And, you know, the translator thing, I’ll have to think about a little bit. As the kind of key role, I was kind of, I think myself, I would hope the hack would be, is that everyone would bring their translational capability with them.
SIDDHARTHA MUKHERJEE:
Yes, maybe. Yes.
REID:
It’s an absolutely essential characteristic. I agree with the humanist, and specifically philosopher as a version of humanists. Obviously there’s a large scope of humanists—everything from artists, anthropologists, everything else, literature, you know, even, you know, kind of, you know, scholar of religion and so forth—there’s the whole range. But I think the philosophy is the orientation. And you know, obviously, a sweet spot in my heart, because of my own Oxford studies. I also agree with the pure scientist. Although I think in this case, partially because of what I’m letting the other three, I wouldn’t say computer scientist. I think biological would be like anyone in the kind of biomedicine area, I think would be the area where I would orient. And, you know, no disrespect to the physicists and many other folks, here partially because of that routing. Now, as the kind of the entrepreneur reinvention I will, you know, get historians everywhere scoffing, and I will omit the historian from the five person spacecraft.
REID:
And I’ll add a technologist-inventor because part of that is, specifically that creation of new technology, which is a very, you know, kind of rare skill set. Just like, kind of the elite, you know, kind of scientist, you know, like yourself or Bob Langer or other, you know, Eric Lander or other people I would consider [laugh], you know, as, as part of this you know, on the, on the pure scientist and kind of biological side. Even with like, for example, Lander’s background in mathematics. So one would be the technologist-inventor. I think another one would be some kind of, and it, the interesting question is, kind of social leader. Because bringing in the dynamic of us as a tribe, because we have this as homo sapiens, homo techne—we are both individuals, and we’re all tribal creatures.
REID:
We’re kind of these, we’re these bizarre dualities, and you have to bring the tribal thing in too. And so that was, you know, somewhere in that, I don’t think that’s a politician—especially modern politicians don’t think big enough. But like, you know, someone who is a kind of a humanist group leader, and I think there are some very enlightened religion people that could fit in that category. I would want to be very selective of which ones, because there’s some also very unenlightened ones. But the, but the right enlightened ones from multiple different religious traditions could be, could be quite interesting.
SIDDHARTHA MUKHERJEE:
There’s one seat left, I think.
REID:
Yes. One seat left. Again, partially because it’s so much of the forward motion—the way to the future is the journey forward—I might take kind of a science fiction writer.
SIDDHARTHA MUKHERJEE:
I’m, I’m dying to hear who you, who you would choose as your science fiction writer.
REID:
Well you know, there’s, there’s a set—Kim Stanley Robinson is interesting.
SIDDHARTHA MUKHERJEE:
Yep.
REID:
You know I’ve actually been recently rereading David Brin’s Uplift series.
SIDDHARTHA MUKHERJEE:
Yes, yes. I know them very well. I’m, I’m a big fan. Yeah.
REID:
Exactly. So I’m just kind of thinking through, you know and, you know, partially I’m dating myself because these are people I was reading when I was reading science fiction like every day. You know, these are the folks, there’s, I’m sure there’s more modern folks that, that would also be you know, well, ought to be included. But, you know Neal Stephenson, I mean, there’s a stack of folks that I could, I could think of as interesting people.
SIDDHARTHA MUKHERJEE:
Well, this is the fun. This is, this is, I think, a fun game for all your listeners. Whether you, you can pick five seats or six seats, and then you could have a vigorous debate about who to put in, put in the five seats. But, but I think, I think I’m moved to thinking— and sort of humbly thinking, I should say—that it’s probably important to have a tribal leader. Otherwise this will become another cacophony on a spaceship instead of a cacophony on Earth. Yes. I think I agree with that, with that.
REID:
The translation function is totally key. I’m just hoping that maybe in the individuals you can get that as a feature.
SIDDHARTHA MUKHERJEE:
Yes. They can carry their own, they can carry their own little translator baggage.
REID:
On a slightly more lighthearted note, we have a set of questions we ask each of our guests. They’re fun. You can answer at brevity or length, as entertains you, or as moves you. And so I will kick it off. Is there a movie, song, or book that fills you with optimism for the future?
SIDDHARTHA MUKHERJEE:
There’s a book which I love, which is Richard Rhodes’ The Making of the Atomic Bomb. It, it seems to me that it would, it would be a kind of a depressing book. But what it does is it reminds you that after the making of the atomic bomb, there was an incredible international response. And so that book is a, is a great reminder that we can do terrible harm, but then we can backtrack our way and learn from that harm that we’ve done. And that we learn empathy from that harm. There’s, the last section of the book, which has a testimony of many of the people who were at Hiroshima and Nagasaki, you know, you, you are moved to tears because we learned from that experience what not to do. So anyway, that would be my, my book of choice. A fat book, though. So fair warning. [laugh]
ARIA:
I was going to say adding to the list. But now concerned. Go, Reid.
REID:
For sure. Well, and actually one of the things I love about it is it’s the—look, look, part of walking and running to the future is you will stumble on your way, but correction of that is important. And I think that’s a great kind of note of optimism. Out of curiosity, knowing you, because I would just want to add specifically also a piece of artwork.
SIDDHARTHA MUKHERJEE:
Um yes. So, you know, I mean, I’ll change tax here. I, so I assume I’m not allowed to use my wife’s work, which contains an incredible sense of explosion and optimism? I’m, I happen to marry, be married to Sarah Sze, one of the great artists of our generation. But I’m the, I’m the follow-on spouse. But I love the, I love Jackson Pollock’s work, for instance. And you know, I would nominate any one of his many, many, many paintings for this. So that would be, that would be my choice. For what it’s worth, I’ll throw in a great, a piece of music, “All of Me” by Billie Holiday.
ARIA:
Love it. So this next question, it can be personal, professional, and you can take it wherever you like, but is there a question that you wish people would ask you more often?
SIDDHARTHA MUKHERJEE:
Oh. I, there’s a, so how about, how about this? How about I tell you a question that, which people won’t ask me? I have the worst sense of direction that I know in a human being. And I think that I actually—and I’m quite serious about it—I think I actually have a development, I think there’s something in my brain that didn’t develop correctly. We have a, we happen to be, have a house which is on the ocean, and there’s an exit that, that we’re supposed, that I’m supposed to take. And I’ve gone through this, through this exit, I think more than 500 times off the highway. And every time I’m taking the exit, I think to myself, ‘am I a hundred percent sure that I should be taking this exit?’ So don’t ask me where we should be going. Reid experienced this personally, because, because we were walking around in Japan, and, and it was, it became very clear to Reid that I was the, I was the, I was the person who was almost certain to lead us in the wrong direction.
REID:
Very funny. So where do you see progress or momentum outside of your industry that inspires you?
SIDDHARTHA MUKHERJEE:
So, I’m going to give you a non-technological, or answer outside technology. I think our debate around technology has actually become quite sophisticated. So I think our capacity to have a real back-and-forth between the sort of the techno-optimists and the techno-pessimists has become more grayscale. And I think the discipline of arguing. or figuring out—and, you know, people can call it bioethics or, you know, AI ethics or whatever you want to call it—I think that debate has become more sophisticated. And I’ve started actually enjoying it. I’ve started listening to it, enjoying it. I’ve started thinking about it. I think that that’s, it’s kind of, it’s a sophisticated debate. And I think, you know, as I said, it’s some art, it’s got history to it. But I’m enjoying this debate, and I think we will learn so much from it, and are still learning so much.
ARIA:
Also, bringing it back to your field—the podcast is called Possible—can you leave us with a final thought of what is possible to achieve if everything breaks humanity’s way in the next 15 years? And what’s the first step to get there?
SIDDHARTHA MUKHERJEE:
Well, I think what’s possible is, you know, a peaceful coexistence in a diverse, multifaceted community—which retains creativity, empathy, creates art, makes new technologies, cures disease, and not just lives longer, but lives a deeper life. A life of deeper, of deeper satisfaction. I think that’s the limits of possibility. And I think if the vector of any technology is aligned in that direction, that’s the right vector to be aligned towards.
REID:
Possible is produced by Wonder Media Network, hosted by Aria Finger and me, Reid Hoffman. Our showrunner is Shaun Young. Possible is produced by Edie Allard, Sara Schleede, and Paloma Moreno Jiménez. Jenny Kaplan is our executive producer and editor.
ARIA:
Special thanks to Katie Sanders, Surya Yalamanchili, Saida Sapieva, Ian Alas, Greg Beato, and Ben Relles. We’d also like to thank Yan Ma and Little Monster Media Company.