D (3:09)

Yeah, really, really good question. And I think I've been kind of fascinated for a very long time with the not notion of the possibility of actually building biology from first principles. You know, the idea that we could actually synthesize life, it's quite an extraordinary concept. And I think that what's really changed in the last couple of years and even in the last couple of months, actually, to be honest with you, is that all of a sudden this idea that we could design life and build life is becoming reality. And I, I kind of discerned that some years ago that that was going to happen. And I really felt that it was important that somebody actually caught that moment and communicated that moment to the general public because, well, first of all, they need to know that this is happening because the consequences are going to be quite, you know, quite remarkable. And life is going to be changed in ways that some of us, well, most of us can barely imagine in a good way. But. But there are also, you know, things that might happen that may not be so good. Right. And so I, I think that what I wanted to do was to create a kind of pocket guide. You know, I. I went to India in my year off before going to university, and I took the Lonely Planet guide with me. You know, when I wanted to know where to stay, then I, I pulled it out. If I wanted to know where the best restaurants were or how much to tip or the best sites, I always have my little pocket guide with me. So I see my book, here it is again, on the future of species as a kind of pocket guide. A friendly companion, a knowledgeable friend that one can put in one's pocket, carry around as one navigates this new landscape, which isn't a physical place on Earth. It's actually the landscape of the future of life, obviously an uncharted space. And I hope that, that my book will serve as that guide.

C (5:08)

I mean, a Knowledgeable Friend is exactly what it was. I mean, the sort of breadth and depth of different sort of pieces of information, basically, that you've strung together into a narrative in this book is actually genuinely impressive. I mean, it's not very common that you end up with books that talk just as much about this sort of hardcore molecular biology, but also the history, like the deep, like history of science that leads up to those things. I think that, you know, you've already touched on the fact that there's this real question about the future and that it could go in either direction and that we need to be equipped basically to understand how to make our decisions as we go forward. Now, if I understood the sort of basic structure of the book correctly, I felt like the starting few chapters were all about really showcasing the history of how it is that we've come to this moment and this idea about what is a genome and what is its utility, and if you had to explain to somebody in just a few words, how is it that you think about the genome or genomes, and importantly, I think to this conversation is what is a species? What is a species as we define it now? And what will a species be after this sort of revolution?

D (6:32)

Both really good questions. So let's start with the first question. And for the first question, I'll actually reference my own book, which itself references a lecture that I went to when I was doing my PhD in Cambridge, given by probably the smartest biologist that I've ever met on Earth, which was a chap that you probably know called Sydney Brenner. And just before the human Genome project began, Sydney Brenner gave. He was a very brilliant biologist, for those of you who don't know who was in Cambridge at the time. South African chap, very eccentric, quite controversial amongst colleagues because he didn't suffer fools lightly and his brain was like sort of electricity, you know, on steroids. He had the most incredible mind that just moved at the speed of light. Anyway, he gave his lecture and he talked about the human genome as what he called a gene kit. Now when I was a kid, I, I was obsessed with these model airplanes and, and they came in as little plastic pieces joined together and, and with an instruction manual. And you'd basically follow the instruction manual and pull off each piece and, you know, connect them up and eventually you'd end up with a plane and you'd paint it, right? So Sydney Brenner argued that the human genome was a bit like one of those model kits. In other words, within that box, the box, if you like, is the box of the human genome were all the components you needed to build the thing that you were trying to build, right? So if you wanted to build a human, you needed a human gene kit. And the genome is essentially that. It's the kind of all the information, the basic information. It's not all the information, but it's the basic minimal set of information required to kind of describe you, basically. Now the, the, the, the, the awkward thing about, for us though, however, as humans, is that sadly, when you buy a gene Kit from a shop. Toy shop, which toy shops don't really seem to exist anymore these days. But in the old days, I used to go to Toys, Toys, Toys in Swiss Cottage and my mum would buy me a Messerschmitt kit or so I was spitfish fire or something like that. And you'd open it up and the first thing you do is you'd pull out the instruction manual, which you'd unfold, and there were all the instructions to assemble the plane. Now, sadly, we don't come with an instruction manual. So all the problem we have as humans is to try and reverse engineer the missing instruction manual. And if we could do that. And I actually call that instruction manual the Grammar of Life. Right. You know, another way of looking at the gene kit is like a language and the bits come together using a kind of grammar, if you like. Right. And we don't have that instruction manual. And if we did, we'd be able to work out how to cure illnesses and how to live longer and all of those things. So that's the answer to your first question. The genome is that instruction manual. Or it's not quite an instruction on it, but it's. It's the minimalist set of information to make an organism. The second point is really important, right? So historically, a species has been described in a kind of linear way. That's because of heredity. So we talk about a tree of life, and all life, in some way or another, has its origin in the very first ancestral lost common ancestor, which is a hypothetical creature that lived, you know, roughly 4 billion years ago, emerged from the primordial soup and then gave rise to all life on Earth as we know it. And as far as we know, there is only one such ancestor. So all life is connected on this, you know, complex branching tree of life. Right? And that's a perfectly good way to think about heredity. And a species is a kind of compartmentalization of bits of the information on that tree. So it prevents a giraffe from mating with a zebra or a beetle from mating with a crocodile. You know, it's a way of partitioning information so that you don't end up with absurdities, if you see what I mean. Right. So that's a simple definition of a species, a kind of reproductive barrier that keeps gene kits or genomes separate from one another into separate categories. Right. You know, it's like art history is separate to popular science. Right. In a library, right. They don't mingle. Well, actually, in my book, they do, actually, but. But they do, but it's category now. What's happening now though is that the, the very concept of a species is at risk for designed life. It will always be there for naturally evolved life because hopefully that will continue, right? But actually the concept of a species becomes meaningless actually, when you're dealing with life that we design from first principles using AI and then construct, using technologies like the one that we've developed at Genero for kind of writing genomes at scale or printing them out. And the reason for that is that when you can design life history becomes irrelevant. You don't need to reference history. You've got a blank piece of paper and you can just write the narrative of any genome you like and you can take your sources from anywhere you like. So actually we're probably going to have to come up with a new concept for organisms that are, that are built in this way. And in fact, I do address that in my book, fortunately, and I use the word morphera. You know, that in the future we might refer to these human and AI designed and built organisms as morphora because they don't have any reference to history. And actually I also introduce a concept, actually, I don't think I did in my book, but since writing my book, I think I came up with a concept of a forest of life. So we move from the tree of life, which is rooted in history and the last common ancestor, into this infinite forest of life, which is different biologies, which history forgot. It's the world that time forgot.

C (12:54)

I mean, there's so much in that. So I want to just sort of tie the concept of sort of species and genome together and just be really explicit for the listeners that essentially our genomes as they stand, are a reflection of our deep evolutionary past. There is a sort of long narrative that is encoded within our genome, and there are elements within that that mean that we are sort of beholden to how things were done before, in part. And so when you, Adrian, talk about this idea of being able to write completely differently and that the boundaries between species disappears, it's because it means that we can take different components, completely different life histories, different evolutionary histories, and sort of combine them theoretically into one package, right. You also touched upon the sort of importance of acknowledging that that will fundamentally be different in this world that is coming towards us, or at least this possibility that we have now of creating these. What was the word you used more.

D (14:00)

For a Morphora Morphra.

C (14:03)

I haven't been close. Morphora, exactly. And so they are a sort of manifestation or will be presumably a manifestation of what we can imagine. You talk a lot about imagination in the book, I felt. And so I wondered if we could talk a little bit about. Well, let's first talk about the things that we might imagine and why we might want to imagine them. So fair enough. We have this technology now where we can write genomes. Why do we want to write them? What's the, what's the narrative?

D (14:35)

Good question. Right. And fundamentally. Well, there are a number of reasons. Let's divide them into categories. One is to obviously understand how our body works and to get rid of human disease. And I think there's a good reason to believe that disease could be completely eliminated should we wish to do so, and that we could discuss that separately. But that's, that's obviously one use. A second use is to turn biology into an engineering material, a predictive engineering material, much in the same way that steel is an engineering material. And when steel used to be an artisanal material, by that I mean it was a very niche material. It was extremely expensive to make and therefore was only used for a very small number of things. But then a chap called Bessemer produced a process known as the Bessemer process, which reduced the cost of steel production by an order of magnitude. So it became 10 times cheaper. And all of a sudden steel could become democratized. And in fact, Carnegie, as we know, basically rolled out steel all over America. And it was used for railways, it was used to build bridges, and it was used to build buildings. It was also used to build munitions. So there were, there were downsides, but it. But in fact, it became a kind of geopolitical phenomenon. Steel, you know, and the same was true of railways and iron and, you know, many, many. And oil, you know, many other such, such technologies. Now, biology, I believe, is going to be the steel of, of our century and the future. Right. The difference is, is the applicability of biology is far greater than steel. And it can be made naturally because you just need to grow things. So it's a sustainable technology, which is a huge advantage. Another huge advantage of biology is that it generates this phenomenon of intelligence. So you could actually embed intelligence into everyday things. So I believe that in, you know, well within our lifetime, we're going to begin to see things being built from biology. So, for example, biology or DNA is a wonderful substrate for archiving information. Information storage, particularly in the modern world, consumes huge amounts of energy, which, as you know, is destructive, results in the production of greenhouse gases and so on, destroys the Environment. If we stored information, DNA, it's a very, very sustainable way of storing information. Energy production. Instead of burning fossil fuels, we could make biofuels, right? The population of the world, as we know, is skyrocketing. And by 2050, they're going to be literally tens of billions of extra people on Earth, and we need to feed them. And how are we going to do that? Well, I argue that we can make synthetic chromosomes. In fact, we're already doing this in general, that can reprogram, not edit the genomes of crops, but literally reprogram them to be able to withstand droughts and really challenging environmental conditions like high salinity. And that can increase the yield of crops. Biomaterials. Spider silk has the tensile strength of steel. Using AI, one could imagine you could build threads of silk that had five times the tensile strength of steel. Because AI can do that. It can explore variants, right? So these are just some examples of ways in which we can start to embed biology into our everyday life. And I, you know, I, as a sort of humorous example, in a sense, in the book, or certainly on the COVID of the book, I talk about living cell phones. You know, why not. Why couldn't your cell phone be made of biological components? Why couldn't your cell phone naturally divide in two rather than needing to be manufactured? For example, why can't we grow houses rather than build houses? So I'm looking to biology as being a technology that could really help us to preserve nature and to preserve the environment. Now, there's one other example of how, how synthetic life could really change the way we do things. And let me, let me tell you about that. So evolution by natural selection has created all known life, both existing and extinct. But known life represents just the most infinitesimal fraction of all possible life, right? So let's look at a few things that have lived and, and think about how critically important they've been in human civilization. Let's think about maize, let's think about rice. Let's think about potatoes and tomatoes and rubber plants, right? Now, all of these things would. All of these genomes, these species were discovered by evolution, by natural selection. But had things been slightly different, we may well not have had any. We may know there may have been no hens. We may never have known that there was such a thing as a hen's egg, for example, or we may never have eaten roast lamb on a Sunday, right? Now imagine now that within the space of all possible genomes, which is so large that one can barely imagine it that if we were able to chart that space of all possible genomes using AI, we'd be able to say, right over here is a species, a potential cash crop that like a potato or like rice or like maize, which has never existed, which could transform human existence and be of immense value. So I think that's the other kind of discovery aspect. It's like there's this landscape out there, an infinite landscape, and we as humans have always liked to travel. Magellan circumnavigated the globe because we wanted to know what was out there. And you've got this biological universe out there. And I think we're getting into our canoe and we're starting to explore it.

C (21:03)

I mean, this was a wonderful example, again, of the exuberance and hopefulness and optimism that I felt was very much present in the book as well. But I think there's something really important there that I want to pick out, which is the role of. I mean, the technical term is selection. Right. No organism exists in a vacuum on its own, and the possibilities that we see in the world around us are possibilities because of the interaction of, effectively a genome and its environment. Right. The environment is a really critical component of natural selection, if not the critical component of natural selection. Right. I mean, ultimately it's to do with individuals being able to get together and reproduce and produce offspring that are able to exist. Right. And so I wondered. And the environment and the use of this technology to help with environmental causes is something that you have talked about in the book as well. Well. And so I wondered if we could sort of, at this point, maybe we can discuss a little bit about the sort of importance of the naturally occurring organisms. Naturally. Naturally occurring organisms and their dependency on their environment. Right. That's how it is that they come to exist and balance that against a sort of future where we deliberately design and change our environment to our specifications. That seems a little bit, you know, contrary to the needs of the original settlers, as it were. So how do you think about that?

D (22:39)

Just to be clear, you know, that's not what I'm advocating in my book. And I do, in my final chapter, I delineate the sort of manifesto for life, which is actually quite cautious and conservative. And I do highlight all these issues. And you're absolutely right that the ecosystems are incredibly fragile natural ones. Right. And what we've learned, actually, from even very simple interventions like introducing one new species into a new environment, and of course, nature as we know it isn't natural because humans have been fiddling around with nature for Centuries by reintroducing species from one place to another, and so on and so forth. But what we know is that ecosystems, A, they're incredibly fragile, B, they're incredibly complex, and C, it's really difficult to predict how the introduction or modification of a species will impact all the other species within the ecosystem. Right. What that means is that the barrier to doing so is extremely high. Right. So in my view, any, the synthesis of new species for our benefit should, for a very long time, you know, and I can't tell you how long that is, but until we are able to mathematically model ecosystems, and I think I do allude to that somewhere in my book. You know, the idea that you could actually mathematically model how ecosystems are constructed and how they behave. And that's going to be really, really complex, of course, but we're not anywhere near ready to do that just yet. Right. So my view is that we can build new species for particular tasks, but we should firewall them from nature. And the firewalling can be done either by physical separation, but physical separation inevitably breaks down, I think. So I don't think that's particularly safe. I think it can be done genetically. So we can build, we can build in different genetic codes, for example, that prevent the transmission of information from artificial life, if you like, to natural life. And in fact, one of my co founders at Genero, Kai Hang Wang, did exactly that working with Jason Chin at the Laboratory of Molecular Biology in Cambridge. They built a synthetic E. Coli genome and actually changed its genetic code, meaning the way in which particular amino acids are kind of pulled in according to the genetic information in the DNA. And so I think firewalling is essential. And it will also be essential to put in fail safe mechanisms so that if species start misbehaving, you can just delete them very, very easily. Right? But I think the barrier to introducing species in natural ecosystems is very high. And for example, people are already producing some prototypes for organisms, bacterial organisms, that can destroy microplastics, Right? So you can imagine that you, you know, engineering a bacterium that you would release into the environment, it would just go all around the world eventually, and wherever there was microplastic, it would just eat it up. Right? Because we need to get rid of microplastics, right? And that sounds like a good idea, right? But actually, is it, you know, is it really a good idea? Because how would that bacterium, that synthetic bacterium, beyond that particular role, impact other ecosystems? And the chances are it would be perfectly fine. Right? But every time we do that, every time we interfere with nature and try to change it by introducing a new species, particularly a totally synthetic one. One has to be aware that there is tremendous risk. So my view is that the threshold, however intuitive it might feel to do something like that, I think the threshold is really, really quite high.

C (26:41)

You know, in the last section of the book in particular, you do spend a good amount of time sort of talking about the sort of recommended guidelines, maybe, or food for thought. You certainly equip the readers with that kind of information. And I was really struck by the fact that you were talking about biosafety and security, but also the need for dialogue. And so where threshold lies, if I understood you correctly, you're sort of saying it's going to be, you know, dialogue is needing to be had because we all sort of share this boat of planet Earth together. But, you know, the elephant in the room that we haven't addressed directly here is about AI, Right? So you said it at the start and when I asked you why. Now, if I understood you correctly, this sort of. The advent of AI comes at a time where we have learned a good amount about biology and about genomes and about how they work. And the main picture that we've been getting so far is it's extremely complex. And trying to sort of bend one's mind around that complexity and to look for patterns and therefore rules within that is really tricky. And there. Enter scene. AI, if I understood you correctly, you're saying AI gives us, is the tool that allows us to now fully understand what's going on and therefore is also the vehicle through which we will do this design. Is that correct?

D (28:07)

Demis is said this really, really nicely, actually. He basically said that mathematics was the perfect language for describing physics. You know, physical phenomena like the orbiting of planets and enable mathematics. Mathematical modeling enabled us to send man to the moon and, you know, produce these space telescopes and so on and so forth. It turns out that actually mathematics hasn't been that helpful for unpicking the grammar of life, you know, defining the grammar book of life. But it. It turns out that AI seems to be the perfect language to do that. Now, AI as we know it today, these kind of language learning models that run ChatGPT and Claude and so on and so forth, they. They manipulate text, strings of letters that, you know, the 26 letters of the English Alphabet, for example. And actually the. The actual Alphabet of DNA is much simpler. There are only four letters, A, C, T and G. So it's perhaps not surprising that the AI that had been developed to manipulate the 26 letters of the English Alphabet and the alphabets of other languages turns out to be really effective at looking at combinations of the four chemical letters of DNA and abstracting information from them. Now my colleague Brian, he my, my third co founder and he's at the ARC Institute at Stanford and in my view is undoubtedly the leader in this field of using AI driven genome design. He developed two AI models, one called Evo1, one called Evo2, and we refer to those as genome language models. So these are, you know, think of them as being just like ChatGPT, but they, they manipulate, you know, combinations of the, the four nucleotides in it that make up the DNA that forms genomes. And what he showed is that if he trained his AI on huge databases comprising the genome sequences of many different species and indeed many different humans as well, that the AI in a way that human minds are completely incapable of doing because we just can't store that much information. Our brains are just too simple, sadly. But that's just a fact, right? AI is able to go way beyond what the poultry human mind can do and can look at all of these patterns and pull out principles that we just unable to discern. And doing that, Brian was able to design the world's first ever totally artificial species. Now it happened to be a virus, and a virus isn't technically living, it's a biological entity, but it has a genome and it becomes living when it infects a bacterium. So it's kind of at the interface of a living thing, if you like. But he showed that the virus that he could build met the technical definition for a new species. Now he can also do the same with the bacterium. He hasn't made a new bacterial species yet, but he can design bacterial genomes. Now clearly there's a big difference between a virus and an elephant, or a virus and a penguin, or even a human. But what it shows is that we're on the route to designing complex life. We're at the very beginnings of doing so, but the process has begun. So sometimes people, when they read my book or hear about my book, because when they read it, they understand, right? But they say, oh, but this is all in the future, you know, what relevance does it have today? No, it's happening today. This is, this is happening. It has happened, right? We are building new species as we speak, right? So people do need to know about this. And most importantly, and again, the reason I wrote my book is that everybody needs to participate in this debate. Everybody needs to have an opinion about this and so I wrote my book to make this quite complex subject accessible to everybody, whoever they are, whatever their scientific background, and in particular, people who have no scientific background. And as you mentioned, you know, I've put in a lot of history and philosophy and art and you know, all types of different facts drawn from many different areas to make the whole subject fun and interesting and lively and entertaining. But it deals with the core principles. And at the end of the day, what's most important is that we align on a manifesto for life that is sensible, safe, responsible, transparent, equitable, and that configures the pathway, if you like, in a way that is beneficial to society and to all society and to humankind, and that respects nature and values nature for what it is and also values human nature for what it is and doesn't try to undermine human nature as it is.

C (33:43)

I mean, the last few words in the book are in fact about human nature being paradoxical and imperfect. If I may quote you in this moment. And I just want to ask one last question. I think coming from a place of probably selfishness, actually being a scientist, being a human being in this moment. Now this comes up a lot when I talk about science in my line of work. But there's a difference between knowing how something works and understanding its meaning, right? Meaning making is not something that AI does, right? And we know that also from models like ChatGPT. They can simulate something that looks like it's meaning meaningful, but they themselves do not hold the sort of the morals, the values, the principles. And so I wondered, what role do you see for human based meaning making in this? What is the meaning that we are pursuing in this future?

D (34:44)

I think it's absolutely critical that humans maintain the upper and final hand in all decision making when it comes to design of life. I don't think there's anything more frightening than the idea that AI kind of takes the key role in determining life's future. Because eventually, as many people, you know, several people have expressed concerns about AI's ultimate, you know, the ultimate direction that AI might be heading in. And it is quite worrisome to think that AI might design types of life that could undermine us and, and serve its own ends, you know, and, and so we, we must ensure that, that all editorial ownership, if you like, remains in the domain of humans. However, I'm not entirely convinced that, that we'll have a good enough editorial eye to spot things that AI puts in into new genomes, you know, because we simply don't have the mental capacity to. Most of us Don't There might be one or two savants who do. And I've met some people like that, actually, who can look at crystal structures and just immediately visualize them. And, you know, but I think even. Even those people, when it comes to visualizing whole genomes and the impacts of small changes, be very challenging. So I think it's important. I love your question, by the way, and I think that you're right that AI isn't human, you know, and it doesn't have our moral sensibilities, it doesn't have our aesthetic. I don't think it has. Will ever have an imagination quite like ours. And remember also that, you know, I argue that one of the beautiful things about human nature is it is imperfect and paradoxical. Right? That's why we're human. You know, if we were, if we weren't irrational, if we didn't upset our partners every day, if we didn't make mistakes and do things we regret, you know, almost on a daily basis, you know, and suffer, that we. We wouldn't be human, you know, and so there is something beautiful in our imperfection. And I'm not sure that AI would appreciate that. Ultimately, you know, AI would be striving for the, you know, the logically perfect system, but that isn't consistent with our view of. Of beauty and nature, you know, so you're absolutely right to. To, you know, raise this issue.

C (37:14)

The thing that I'm pushing towards here is I think that humans have a tendency, when faced with something that seems so immovable and inevitable, to sort of lose their own sense of agency a little bit. And as we reach the end of our conversation, I wanted to let you know that what I thought this book, your book here did such a fantastic job of, is basically allowing readers, furnishing readers with the knowledge and context necessary to sort of help restore a sense of agency in the face of this potential future. Because to be the head of the ship, to lead the way that we want to lead, we actually have to maintain a sort of sense that we are, you know, we have agency, we have control, and we have morality. And it is a tool that we use. Right? AI and these sorts of biological design concepts, they're all in your vision, as I understand it, at least tools, materials. And we must not necessarily roll over just because our tools appear to us to be like. Like us, to just simply accept that they are right.

D (38:22)

No, that's absolutely right. And the moment we've reached, we've reached in history, and not just the history of humankind, but the History of all life on earth is totally unprecedented. You know, it's a phase transition, because all of a sudden, you know, rather than nature being something that is received and that we observe and that has no real agency or design or intentionality or foresight, suddenly there's a new, a new show in town. You know, there's this, a new author which is us in conjunction with AI, and suddenly all biology which we took for granted, including our own nature, is negotiable, you know, because eventually we will understand the nuances of how human genomes are constructed and we will understand the basis of morality and of free will. And eventually we might find ourselves in a world where an authoritarian state or regime decides that such things are undesirable. Why would you want to have a population with free will? How inconvenient. Why would you want people who challenge your ideas? And, and, and free will is only lost once. So I think eventually those types of issues will become an issue. They're, they're far away, fortunately, but, but they will, you know, they, you know, somebody watches. If this, if this interview survives another hundred years, and if you're watching this in 100 years, you, you will probably appreciate what I'm saying now. And I have no doubt that, you know, those conversations, if we're lucky, those, There will be dialogue and issues like that. If we're unlucky, there won't be. And things will, you know, will be done without consent, you know, so my manifesto for life is, I call it a sketch for a manifesto. And it's a sketch because I don't believe that any individual is able to or has the right even to, to write a manifesto for such an important thing. So, you know, it's simply a sketch. And other people, including everybody who reads my book, should have a voice and be at the table helping to make those important decisions about how life will navigate its own future.

C (40:44)

That's a wonderful note to wrap up on. I think that if I was to summarize this whole conversation, you are saying to us that we have unlocked a new medium in which humans can communicate, can write, can write a narrative of sorts, be authors, and we know from AI based models that are able to consume vast amounts of the things that we've written in written language, right? That there are patterns and trends and things within that that we have to confront, that that is what we have done to date. And now this time is coming where we have a new medium, a new format in which we can write. And the real question is, what is it that we want to write, right? What are we going to narrate at this point? It's what's written and who does that exactly. And you know, your book is a, is a call to arms to sort of think about that. And as I said, I think also a moment to reinforce our sense of agency as we enter into this new dominion so that we have, you know, a sense of what we will narrate as we go. And so on that note, Adrian, thank you so much. That was Adrian Wilson, the author of on the Future of Species, which is now available online and in stores. Look out for that bright yellow cover with a finch on it. I've been Ganesh Taylor, you've been listening to Intelligence Squared. Thank you for joining us.

B (42:14)

Thanks for listening to Intelligence Squared. This episode was produced by me, Mia Sorrenti, and it was edited by by Mark Roberts.