One of the defining characteristics of complex systems science is the shift in emphasis from objects to relationships and processes. How is information related to matter and energy, and how do the distinct formulations of different scientific lineages braid together in a unifying pattern? This search for a more fundamental understanding drives directly into some of the biggest questions science has to ask about the living world — namely, what is life, what is alive, and when did life begin? The Santa Fe Institute has drawn from the deep wells of these questions since the 1980s. In our second episode, Complexity Podcast dove in to explore the origins of life, but even that in-depth conversation left a lot unsaid.
Welcome to Complexity, the official podcast of the Santa Fe Institute. I’m your host, Michael Garfield. While we continue our short summer hiatus, here’s a superb interview with the Santa Fe Institute’s newly announced External Professor, Sara Imari Walker of Arizona State University, by Marty Martin and Art Woods, the hosts of the Big Biology Podcast. In this rapid-fire rap from their ninth episode, Sara talks about how physics — and in particular information theory — refocuses the lens through which researchers ask about the nature of living systems and look for signs of life elsewhere in the cosmos. We hope that you enjoy and — after subscribing to Complexity and Big Biology wherever you go for podcasts — follow up with their equally illuminating conversation with SFI External Professor Andy Dobson on disease ecology.
Visit our website for more information or to support our science and communication efforts.
Sara Walker’s Website
Sara Walker's Google Scholar Page
Complexity Podcast Episode 2 on The Origins of Life at InterPlanetary Festival
Big Biology Podcast Website
Join our Facebook discussion group to meet like minds and talk about each episode.
Podcast Theme Music by Mitch Mignano.
Follow us on social media:
Twitter • YouTube • Facebook • Instagram • LinkedIn
If you are interested in volunteering to help us edit transcripts, please email Michael Garfield.
Sara Walker: We really don't understand how information operates in the physical world. And if we understood the laws of information, they would really be the laws of life. So being a physicist, and again, this was another one of my analogies, but I like these because they're illuminating for me about how I'm thinking about things is that, like, if you want to understand the laws of gravitation, then you go and study black hole because you have a very intense gravitational field there and you can understand, and pro gravity very well. If you want to understand a lot of information and how information operates in the physical world, then you should go and study a living thing because that's where information actually has its most, you know, it's like densest or most intense. And so I think the origin of life process is really this transition where you have systems that where information is not really a prominent part of the physics of those systems to ones where it really is. And so we really have to understand how is it that information actually starts to gain control over the dynamics of the system.
Michael Garfield: One of the defining characteristics of complex system science is the shift in emphasis from objects to relationships and processes. How is information related to matter and energy and how do the distinct formulations of different scientific lineages braid together in a unifying pattern? The search for a more fundamental understanding drives directly into some of the biggest questions science has to ask about the living world. Namely, what is life? What is alive? And when did life began? The Santa Fe Institute has drawn from the deep Wells of these questions since the 1980s. In our second episode, Complexity Podcast dove in to explore the origins of life. But even that in-depth conversation left a lot unsaid.
Welcome to Complexity, the official podcast of SFI. I'm your host, Michael Garfield. While we continue our short summer hiatus. Here's a superb interview with SFS, newly announced, external professor, Sara Imari Walker of Arizona State University by Marty Martin and Art Woods. The hosts of the big biology podcast and this rapid-fire rap from their ninth episode.
Sara talks about how physics and in particular information theory, refocuses the lens through which researchers ask about the nature of living systems and look for signs of life elsewhere in the cosmos. We hope that you enjoy and after subscribing to Complexity and Big Biology, wherever you go for podcasts, follow up with their equally illuminating conversation with SFI external professor Andy Dobson on disease ecology. If you value our research and communication efforts, please consider making a recurring monthly donation at santafe.edu/give or joining our Applied Complexity Network at santafe.edu/action. Thank you for listening.
Sara Walker Episode Transcript
AW: So, we're here with Sara Walker today, talking about origin of life problems. We're meeting up at a Gordon Conference in Biddeford, Maine. Those devoted to, in the grand phrasing, "unifying life across scales and ecology," and hearing a great set of talks. Sara's going to talk tomorrow morning, and we got her to sit down with us and to talk about her work. She is a fellow at the ASUSFI Center for Biosocial Complex Systems, and also an assistant professor at Arizona State University, and we're just super delighted that you're sitting down to talk with us!
SW: I'm excited too, it's going to be a lot of fun.
AW: So, you've written a lot about the origin of life on Earth and possible origin of life elsewhere in the universe, and we just wanted to dig into the details of that, and I want to start by just reading a quote that you had in one of your papers: "The origin of life is among the greatest open problems in science. How is it that life can emerge from non-living matter?" We totally agree. Like, fabulously interesting question.
SW: I was going to say, it's a bad start if you disagree with that.
AW: And I guess, so maybe take us through what we do know, just very briefly, about the history of life on Earth. So, you know, at what point did it arise, and then what's the sort of big, open set of problems that are associated with that?
SW: Yeah, so it's interesting because you're asking about the historical origins of life on Earth, and I think there's something interesting about asking it as a historical question, or asking it as a testable question, so we can get to that later, but so there is this historical question about origins, like what was the origin of life on Earth, and then there's another question that people ask in the origins community is, is the origin of life like a testable thing, like can we do it in the lab, and how do we actually understand that, or can we look for origins of life on other planets. So, there's actually a lot of ways of thinking about origins that aren't just the historical, what happened on early Earth. The idea in current thinking is that life arose on our planet around 4 billion years ago, but the dates are very difficult to constrain. The earliest fossil evidence of life is fairly complex, around 3.8 billion years ago people have found evidence...
AW: Those are the stromatolites?
SW: Yeah, those are the stromatolites, and so those are like microbial mats, they're very complex ecological communities, and if we try to push the understanding of the history of life further back than that, we don't have a fossil record, and it's very unclear and speculative exactly when life emerged. But a lot of people make arguments that since life was so complex at that stage, it had emerged very early, but we don't actually know the pace of evolution early on either, so it could have happened over a couple 100 million time scale, it could've been longer, and so that part of it is very speculative. So, there's a big sort of black box if you go any earlier that we don't know exactly when life emerged, we don't know in what environment, we don't know what it looked like when it first emerged, if it had the same chemistry that modern life has, and so there's a whole host of questions there.
MM: Are we, we're fairly confident though that it arose once?
SW: Presumably. So, this gets into some interesting things, because there is this well, sort of, discussed attribute on life that all of life on Earth shares the same biochemistry, and so what's meant by that statement is every organism we've ever characterized has DNA, every living thing, you know, uses ribosomes as part of its translation machinery, has proteins. So, all of the biochemical architecture that's at the core of life is universal. And so, people will talk often about this idea of a last universal common ancestor, so previously I talked about the rock record, the fossil evidence with stromatolites for life, but we also have evidence of early life from the molecular record, and that's actually a really fascinating record in its own rite, because, you know, there's some molecules like the ribosome that have been around for 3.8 billion years, but they're not a physical structure, it's a thing that keeps reconstructing itself, so there's some kind of information associated with that, right? There's this informational pattern that keeps reconstructing itself on this planet. But if you're talking about like, tracing all of that history back in time, it seems that all of the lineages that we study now have this common ancestor that we call the last universal common ancestor, and that being singular is kind of, I think misleading to a lot of people, because they think it was a single cell, and in reality it was a community of cells, and a lot of people that study this early phase of evolution recognize that there was a lot of horizontal gene transfer happening in things, so you can't even really talk about an individual with how these kind of populations were evolving. And so what I think gets really fascinating about thinking about this early stage is that a lot of our things that we think about, like with individuality or the way we think about evolution being Darwinian and not necessarily a collective kind of breakdown when you get to that last universal, yeah. So the last universal common ancestor, I think, so that's the sort, so one way to think about it like, because I'm a physicist and my brain always goes there, is like, the last universal common ancestor is sort of like what the surface of last scattering is for cosmology. It's like, the surface where like, you know, photons and matter decoupled and we actually can see the early universe, but we can't see much earlier than that, because it's just like a thermal shield. If you think about the last universal common ancestor, we had rapid genetic exchange of information, and we don't really, you know, before that there was no genetic machinery, like we don't know when genes evolved, and we can't go back earlier than genetics, because that's how we trace the history of things. So, it's sort of, it's sort of like this, this surface that we hit all the lineages go back and they kind of unify in this kind of amorphous, you know, messy chemistry of, you know, rapid gene exchange, and then we don't know what preceded that. And so there's a lot of hypotheses about life and early life, and how life originated, and so you'll probably hear all kinds of different ones that are more popular, but the one I'm really kind of fascinated with most recently is this idea that the emergence of life was actually a planetary scale process. I think this is really an intriguing way of thinking about it, because a lot of people think about life as a localized transition, so they think, you know, some little protocell formed in some pocket hydrothermal vent, and some RNA got in it and that RNA started copying itself, or you had some RNA forming on a beach somewhere and that started copying itself, but you always have this kind of, singular...
AW: Gradual takeover of the entire world?
SW: Yeah, the sort of singular event, evolution got started and it spread out. And that's a very biological narrative, right? But then you have like, and so, then you'll have other people talking about sort of like, metabolism, self-organizing, and these kinds of things. And a lot of times that kind of gets also talked about as a localized event, that you had some, you know, collection of molecules that formed what's called an autocatalytic set, where each molecule can produce the next, and it forms a closed cycle so you get a reproducing system. So again, it's trying to get at this idea of a self-reproducing system that can evolve. But if, if you think, so, but there's this whole set of other ideas circling around, how biochemistry might emerge from geochemistry, which I think is really interesting, because you're going to think about like the Earth, and the emergence of life, you know, is emerging from the Earth, and that there actually might have been geochemical cycles that started organizing into biochemical cycles. And so Harold Morowitz is really...
AW: So, what kinds of geochemical cycles?
SW: Well, so, Harold Morowitz was one of the first people that really started talking about this and has been, sort of, promoted more recently by Eric Smith, and Edward [UNKNOWN] and all these kind of people that are thinking about these things, but the idea is that central metabolism, in particular the [what is this word] citric acid cycles, like the primitive core of metabolism, actually might have been something that was thermodynamically favored geochemically. So, you have this self-organizing cycle that actually became a product, was a product of geochemistry, and then out of that emerged all of these biochemical architectures, because once you get the citric acid cycle going, like that basically makes a lot of the things that are building blocks for life.
AW: So, do you mean that the citric acid cycle got going in some prehistory in the absence of all of the enzymes that we now know are catalyzing?
SW: Yes. And so, yeah, so there's a lab in, led by Joseph Moran that's been basically showing a lot of those steps can be catalyzed without enzymes, and just by minerals. So, there are testable hypotheses there, which is really interesting. So, one of the reasons I think this is really intriguing, and it's a relevant thing actually to talk about at an ecology conference, is I think there's a lot of misconceptions when we're talking about life, about there being a fundamental unit of life, or a particular scale that life is like, preferred to exist at. So, I think ecologists appreciate this because they recognize, you know, ecological communities are very alive in a sense, you know, that the individuals aren't, or the, it doesn't make sense to talk about the individuals sometimes because they're so amorphous. And so, for thinking about origins of life, you know, you can think about the origins of life itself actually as an ecological process. There's some organizing cycles, and maybe like individuality emerged later, and one of the reasons that I find this really intriguing is because when I think about life and the kind of, sort of, informational perspective, if you want to call it that, that I think about it, it's really sort of like the hierarchy, and that these processes exist across many scales, and the only natural boundary for that kind of process is actually at the planetary scale. And so, it seems very natural to think about life at a planetary scale, and it's actually very illuminating to do that. So, one of the things I do in my science a lot is try to find places where the thinking is so different, or so skewed from the way we're used to it that you actually have, like potential for a lot of conceptual breakthroughs. They may be, they may not be right, but at least thinking about the problem differently opens up lots of new questions, which is what you want to do in science.
MM: Yeah, that's a really interesting approach to things, you know.
AW: Yeah, amazing. As you're talking about these cycles and sort of geochemistry, I was thinking about, how do you graft on all of these much more familiar biological parts, so in your conception of that, how did DNA arise? How did we get RNAs? You know, sort of transfer mechanism. How do proteins as catalysts arise, and then take over those functions that might have been formerly done by minerals, say, in a geochemical cycle?
SW: Yeah, so I think a lot of these questions are hard because we're so used to thinking about the details in biology and not the underlying processes. So, as you're saying all of these things, I'm thinking, I'm like, how am I going to explain that, how am I going to explain that from the theory? But I think the real thing is that like, for example, when you're talking about DNA, there's an interesting thing with DNA, because DNA instantiates information that's relevant to the organism, and so what I think is important about DNA is that it's a molecule that allows you to copy information contained in the molecule to another molecule, and do so reliably. And so, I think, I think, I don't necessarily have answers to all of the steps that you asked, but I think a way of thinking about it is that we talk about a lot of things that are really essential to biology, like metabolism, and reproduction, and compartmentalization, and so a lot of times when people are talking about life they'll try to make these like list definitions of life, like it has to have all of these attributes. I think if there really is an underlying theory that explains life, that all of those properties are probably derivative of that theory. And so, this is what appeals to me about thinking about information as being the organizing thing for living systems, because all of things really do have sort of a natural description as being derivative processes. So, copying of information is obviously really important, but information flows also can organize metabolic cycles, and if you have a compartment, that means that you actually are excluding some information from the environment and allowing some information in, and actually can structure a system very differently. And there are lots of really interesting, sort of, ways of thinking about what the underlying structure of these systems is, if you translate everything into that kind of language, then you get a much clearer, unified picture of what's actually happening, from my perspective.
MM: That's interesting. That's a really good segue into, when we teach introductory biology, you ask the students, "what's life," and they start giving off the list that you just said, "well, okay... think about this..." Yeah, exactly. So, I mean, to transition into more, sort of, digital-analog, top-down, the types of things that you've been thinking about, you have, to quote you, again if I can get it right, "The two problems of defining life and solving its origin, they can't really disentangle." So, what's that about?
SW: Right. So, I started, so I am an origin of life scientist, and I think of my primary subject matter as being origins of life, but if you look at like, what I do and what my research group does, almost none of it would seem like it's origin of life related, because we look at biology across all scales -- organization, you know, in things like ants or cities, or, you know, whatever it is. But, for me, it's because when I was, in particular when I was a post-doc, because I was working at the Center for Chemical Evolution at Georgia Tech, building models for chemical evolution, there's always this problem with origins of life that you want to talk about this transition from non-life to life, but you have no criteria for what life is. And it would just drive me nuts, because I'm supposed to be building models for chemical evolution and saying they were becoming more lifelike, and I have no quantifiable criteria. And so, this is why I think really that theory is critically important to origins of life, and that theory really must be one where you have some way of objectively saying what it is that you mean by when you say something is alive or not, or something is life or not. And alive and life are interesting also in their own right because they're not exactly the same thing. But, so, most of what I really try to think about is what is life, and how would I actually, and sometimes build a life meter, or make some sort of quantifiable criteria for what life is. And so, for a long time, I think I was thinking that issue was more black and white than I do now, that maybe there was, because you could ask is it continuous, or is there some sort of discontinuity, where you'll go from something is not alive to something is alive. And the way I think about it now is that, when we're talking about life, life is an example of a particular kind of physical phenomenon, and what I really think underlies it is that we really don't understand how information operates in the physical world, and if we understood the laws of information they would really be the laws of life, and so being a physicist, and again this is another one of my analogies, but I like these because they're illuminating for me about how I'm thinking about things, is that, if you want to understand the laws of gravitation, then you go and study a black hole, because you have a very intense gravitational field there, and you can understand and probe gravity very well. If you want to understand the laws of information and how information operates in the physical world, then you should go and study a living thing, because that's where information actually has its most, you know, it's like densest, or like most intense. And so, I think the origin of life process is really this transition where you have systems that, where information is not really a prominent part of the physics of those systems to ones where it really is. So, we really have to understand how is it that information actually starts to gain control over the dynamics of the system and becomes an important part of it. And then most of the unfolding of the biosphere over the last 4 billion years, in my mind, is building increasingly extracted levels of informational structures that are, you know, you build up this hierarchy. So, the kind of information that we manage today is very abstract, not very tied to the physical substrate at all. So, you can compare like, the information in a computer, like you guys have text on a computer I'm looking at right now, and you just read some of that text, so like, the people listening to this podcast are going to hear that and it's going to get translated into their brain. So, you think about all the different physical media that information went through, and somehow it retains its properties of being information, and meaning the same thing in this room as it does to the people that are going to be listening to this podcast. Now, if you look at DNA and like, early information structures, it's really difficult actually to copy the information in DNA to other physical media, other chemistry, and do it so, the information there is really tied to the particular chemical, physical thing. And so, nowadays we can copy DNA to other genetic polymers, but that required 4 billion years of evolution. So, if you give me, yeah that information can be instantiated in other things, but early information was, it became abstracted from those physical things, but the level of abstraction and the number of different things that you could actually instantiate that information in, and what that information is doing as far as its control over the dynamics of the system has just been increasing over biological evolution. But somehow, that started at the origin of life.
AW: Wow, so many ways to go here. I want to ask about, you know, you just sort of gave your reason for trying to define what life is, and I would say that's a hard question for most biologists, because we don't grapple with it very much, weirdly, right? And is that because most biologists sort of take it for granted that we already know what life is, or do they feel like it's too hard of a question, and unapproachable?
SW: Well, I think, I think, it's not necessarily a relevant question to the questions that you're asking. If you're already studying biology, you want to ask questions about the biology we have on Earth, then you don't really need to know what life is universally. So, I think it just depends upon what description, level of description, you're at. I think the place you really need to know is origins of life or looking for life on other planets. But if I just want to know, like, you know, I go in my backyard and I want to understand why the bear is eating the berries, I don't need to know what, you know, what is the fundamental like, physical structure that is a bear. So, I think it just, and I actually was thinking about this for my talk tomorrow, because I talk about something like, something called a hard problem of life, and people usually talk about like, the hard problem of consciousness. But most neuroscientists don't care about what consciousness is, right, because they care about the functioning of the brain, and they want to make your brain healthy, and these kinds of things, and consciousness is a really hard problem. We have like no handle on it. And I think most people studying biology, you know, they care about the mechanistic details of their particular question, but life is a very abstract concept, and it sort of, it exists at a different ontological level if you want,or something, but it's just sort of in a different ballpark of questions. It's also like planetary formation and like gravitation, you know, yeah.
AW: I mean, I guess I totally agree with that, and I, you know, I have my own sort of set of biological levels that I work at, and yet, when you pose it like that, I really want to know, like what is life?
SW: Well I think everybody wants to know! It doesn't mean...
MM: You've already, I think you've suggested, I may be overstating or misrepresenting what you had said, but some of the points that you've made have been about, well let's think of evolution again. You've sort of just made the comment that there were maybe things like communities, which is different than LUCA (last universal common ancestor) usually conceived, so, is there, I mean, have you thought about the potential value of biologists, no matter your level, no matter your interests, what types of things it might do for biologists?
SW: Yeah, yeah, definitely! I mean, so, my quest is that there is like...quest, it's a quest, it's an epic quest, like you know, like I don't know, some video game or something. No, but like, obviously I'm really passionate about the problem I work on, so like, I really care about this problem being solved. I don't care if I solve it, I just want someone to solve it, and so, I really do view myself as just trying to like, make enough of a conceptual shift so people can think about it differently, so like, we as a scientific community can be part of some of these kind of questions, because they're not ones that any individual could possibly answer. But, the, but what I really care about is whether there is some unified explanation for all of life, and what I mean, like, from the origin, to understanding what the heck it is we're doing as a technological civilization, because it's a very bizarre state for, to actually like, be a thing that exists, like most people don't think about it, but like, especially being a scientist, it's really weird that we do science. Like what is science? It's such a weird thing, like there's like these physical things on the surface of our planet that are thinking about how the world works, and then they come up with theories, and then those theories actually describe the world, and then they can do new crazy things with that. An example I like to give is like launching satellites into space, which seems kind of like a mundane thing in our society now, we do it all the time, but in order for that physical process to exist requires knowledge of the laws of gravitation, which means that you have to have a civilization or something like a technological civilization with sophisticated knowledge of how its world works that it's understood the regularities associated with gravity, and then built technology to cause these transformations. That can only happen if you have that knowledge, so that's a kind of information that allows this new thing to happen. And so, I think that's a, that's very fundamental and very deep, and so I think like, if we had the proper theory for what these things are, what life is, that it would inform everything, from what's happening in the chemistry inside cells, to understanding, you know, the future of AI, because it's just about information and what information does, and so... So, I would, I kind of would consider all of these different levels that we look at biology as examples of that physics, and we don't think about it that way, but that's actually the way I've organized my research group, is they all work on like different, totally different biological systems, but the whole point is that looking at them from this unified perspective, hopefully we can get insights from one into another to try to figure out what that theory might even look like, and we don't have a clue what that theory looks like, so, but somebody has to jump in the deep end, pretend it exists, and then see if it does.
AW: So, you've talked about digital versus analog and how you sort of, the solutions to that problem. Do you want to say something about how that fits with the definition of life?
SW: Oh sure. So, I think the reason for proposing talking about digital and analog information processing systems is that in particular with the origins of life, there's a huge dichotomy in how we think about that problem. So, people tend to be very disciplined, specific in how they approach it, which I find fascinating, but like the leading hypotheses for origins of life have traditionally been genetics first or metabolism first, and the genetics first theory assumes that some molecule, presumably RNA, or something that can copy itself, emerged on early Earth, from some prebiotic chemistry, and started copying itself and undergoing an evolutionary process and therefore it was a genetic system. And then the metabolism first says instead, oh well, what life really does is harness free energy and so the first thing that life did wasn't necessarily evolution or copying information, it was harnessing energy from its environment and so the first living entities were metabolic cycles. Or, you can...
AW: And just to be clear, so, which of those is genetic and which of those is analog?
SW: So, yeah, so, yeah, I'm going to get, yeah, I'm going to get there. So, the, so, the point I want to make about the disciplines is that most people that think about the genetics tend to be like molecular biologists or biologists, most people that think about metabolism come from physics or geochemistry, or like these kinds of backgrounds. So, you can like, you can almost clearly see the disciplinary divide there, and I think it's just because we're struggling to make inroads to like, what are the relevant questions to ask for origins of life, how much of life is still there at the origin, and how do we actually even get there? And I think probably all of these things are in part right, and what we really need is a more unified approach of thinking about life, but that has to be an appropriate level of abstraction, and so, from this sort of informational perspective, you can recast the genetics as being like a digital component was important early on, and metabolism as sort of an analog or more, you know, continuous kind of system. And then you can actually cast both in a common language and start asking kind of questions about them from that perspective. So, so that was really part of the motivation, and I have become increasingly convinced that the origin of life problem is actually a problem of unification more than anything else, and usually we have major conceptual breakthroughs in physics when we have unifications of very different things. And you can say that unification in a lot of different ways, you might say, well, genetics and metabolism have to be unified in some kind of, like, way of understanding both at the same time, or the way I talk about it in sort of like a much more abstract way is thinking about information and matter have to be unified, or information and causation is another way to think about it, but that there is some problem where we understand how matter and energy work, and we understand information in the abstract, but we need to understand those two things as a unified concept.
AW: Another trait of life that you've written about is that you can recognize life as the acquisition of top-down control over networks of parts that contribute to life, so can you talk about that transition?
SW: Yeah, yeah, yeah. So, I'll give a little bit of background to understand, motivate why I'm thinking about it that way, but there's this kind of interesting thing... So, I'm trained in physics, and in physics, like the way that we're taught to think about the world is to try to reduce the world to like the most fundamental description of reality and the simplest, and it's usually a sort of very reductionist approach because you want to go all the way down to like, elementary particles, and understand like, all the component parts. And that's great and beautiful and elegant, and I absolutely, I love physics and what it's been able to accomplish, but that description usually tends to be one where you'll take the, what we call the micro-state, or the lowest level of description of a system, and then once you know what the micro-state is doing you basically say everything about any kind of level of description you might want to do above that, and usually like, the way we talk about levels is we'll say, well there's a macroscopic description, and it just means you don't know like, the position and velocity of every particle. And so, thermodynamics is sort of a classic theory that gets at that kind of thing, because temperature is a very good way of describing, you know, like gas of particles, but you don't have to know the position of every particle or anything to talk about something like temperature. When you get to life it gets a lot harder, because if you talk about those macro-scale level properties, those tend to be things that we associate like, with function, for example is like a relevant macro-scale property in biology. And when you're talking about those kinds of things, they actually influence the system. So, it's not like it's just some meta-level description where you don't have complete knowledge of what the micro-states are doing, so you're using it as an effective description. But I think it's actually fundamental to what that system is doing. And so, I can even just like, I mean, as a scientist, we're like, we're describing a thermodynamic theory, but then we use thermodynamics to go and do something in the physical world, so somehow that abstraction actually, you know, matters. So, that was my example of gravity too, is that like, these sorts of higher-level things actually become what I would call causally efficacious in their own right, and that's the top-down aspect. It's that, it's something that's not physically instantiated in any particular substrate -- it can exist in many, it can be copied between many, but somehow it actually can, in part, control the dynamics of the system, or, yeah. So, the top-down causation is actually that part of it.
AW: And if I just try to like, apply that idea to the origin of life on Earth, and I'm imagining, you know, sort of networks that are physically instantiated, and then I'm trying to imagine like, how does this top-down control, like what, how does that crystallize out of this stuff? So, how does it?
SW: Not sure yet! Yeah, ask in a few years. I actually like, I just got a project funded where like, we're going to try to get at that a little bit more concretely with like, actual experimental data and try to measure what information is doing in these like, intermediate stages between non-life and life, so I have a couple experimental collaborators -- Kate Adamala and Lee Cronin, that work on like, sort of opposite ends of the original spectrum for their systems, so the idea there is that we're actually going to go through and build a bunch of chemical systems and look at how information is structuring those systems to try to understand what's happening.
AW: When you say build a bunch of chemical systems, you mean like in the lab?
SW: In the lab. So, Kate, for example, works with synthetic cells, so she can build cells with biological parts but they're very minimal, so it might be like, you know, a couple of genes instead of like, hundreds of genes that we have in real cells, and you can actually track, you know, like what is information doing in that system, and then, Lee takes these complex chemical mixtures and he tries to start from like simple building blocks actually, they're not complex to start, but he evolves complexity into them by varying environments and using, he has like these robotic algorithms that he does like, these mini iterations of ensemble experiments to generate statistics over how chemistry is actually increasing its complexity, but like, going down these many different paths for chemical space. And so, so the idea is like, can we actually bridge those two things that try to look at them in this abstract level of like, what is the information actually doing. But I think that's the critical question about all of this, and actually, the top-down causation aspect of it is a very deep philosophical question, because a lot of people think that can't even work in practice, because there's no limit at the bottom for extra causal forces. And what I, I think part of the misnomer of that is people think there's a micro-state and a macro-state, and that macro-state is exerting control over its own micro-state, but that's not true at all. What's actually happening is, you know, I'm an information processing system, and when I look at the external world, I don't have access to all, to every degree of freedom in the external world. I have certain ones that I care about and pay attention to, and those actually become causal in my dynamics, because those are the ones that are, I'm internalizing and interpreting about my environment. And that's actually a kind of top-down causation -- it just means that I'm not coupled to every single degree of freedom. I'm coupled to a reduced set of degrees of freedom, and that's what I call information. And so, I don't think it's anything mystical, I just think it's just that we haven't really gotten to the bottom of what that is and how that starts happening. But I think it happens in chemistry at a certain scale of chemical complexity, it's just that the chemical systems become so rich in the possibility space, and so rich, just like if you think about like, you know, a folded protein or something, there's no way that like, the function is a small component of the system, and all the rest of the degrees of freedom don't really matter to the function, right? So, when you start getting those kinds of systems, then you have the possibility of information actually starting to matter. Now, how it can start to matter and why on Earth it actually did this weird thing that it doesn't seem to have done on Mars or other planets in our solar system, is a really hard question, but I think the first step is to get into this kind of conceptual framework where we can start thinking quantitatively about it and asking these kinds of quantitative questions, and then we're in this like, kind of, space where we're feeling out, you know, what does that even look like and mean and you know, starting to try to connect to experiment, and hopefully like, the theory and experiment will start feeding back on one another and we'll actually get some more concrete ways of talking about that.
MM: So, the, we weren't going to bring it up, but I have to bring it up in light of what you've been talking about... Where do viruses fit into the way we think about the...
SW: Everyone always asks that question!
MM: Well, the way that you, we had no plans to talk about it, but it's...
SW: Yeah, yeah, yeah, no, it's great! Yeah, no, everybody always asks. So, so, I think one of the biggest misnomers with trying to find life is to say a thing is alive, and we can draw a boundary around that thing. Like, that's just not what life is. In my mind, life is a dynamical process, and it's one where you have particular informational patterns that are like, structuring physical systems across space and time. And that's what life is. So, you can't draw like, a hard boundary around that process. And from that perspective, of course viruses are part of a living architecture, they're part of life, because they contribute to that kind of process. And so this goes, and so I mentioned there was a distinction between life and alive, and in my mind life is all of the things that are generated by this kind of process. So, like, people get mad that I think a computer is life, or a screwdriver is life, but those things literally would not be created without information.
AW: Yeah, well it's consistent! That's okay.
SW: Yeah. And then alive is the systems that are actively constructing things, they're the ones doing the information processing to actually build those things, and like, use internalized information to actually do the construction.
AW: That's a good distinction.
SW: Yeah, so I think there's, and both are kind of necessary for understanding life and what life's doing. They're very fundamental to like, trying to figure that all out.
MM: So, sort of like von Neumann's programmable constructor type of...
SW: Yeah. I love von Neumann's stuff on universal constructors. I think he was way ahead of his time. It's brilliant.
AW: So, let's dig into that. Yeah, so, what is a universal constructor?
SW: Yeah, so, von Neumann, I think, he was really inspired by Turing, but his whole line of reasoning was, you know, Turing was interested whether you could compute any computable function, and von Neumann was like, well that's interesting, but I'm interested in like, you know, real physical things, so, could you build a machine that could build any physical thing? Right? So, this is a really interesting question. Is it possible to make, have one device, or one thing, that could construct, build, any possible thing that could exist?
AW: And just remind our listeners when this thinking was happening, when was von Neumann doing his work?
SW: Yeah, so this was like 1940s, I think he died in 1950, and his book on self-reproducing automata was unfinished, and then it was finished posthumously by Burkes. But, yeah. So, this was very early, yeah, very early. And he inspired a lot of people to think about things very deeply. But yeah! So, this gets into...his real question was about building a machine that could reproduce itself, and so that's why he went into this idea of a machine that could build anything, because if it could build anything, it could build itself. And so, what he recognized though, was in order to build itself, it had to specify itself, which means it had to have a program, or some kind of information to tell it to build itself. And it's really interesting because there's some paradoxes associated with self-reference there, where like, if you're talking about building yourself you have to have an image of yourself, and then you end up getting into this sort of hierarchical loop.
AW: So, that imaging includes the program itself, and yeah.
SW: Yeah, yeah. So, he kind of quickly recognized that you would, you know, you need an infinite storage space to actually specify all of the future generations. And so, what his resolution was was his idea of a universal constructor actually contains three parts. It contains the constructor, which is the thing that does the building. It contains the instructions, which specify how you make the thing, and it contains a third component, which he called a supervisory unit, which tells the system when to copy the instructions. And the copying of the instructions basically means to just make a second copy of it, but it doesn't, it's completely blind to knowing what the actual instructions tell the system to do. And that was actually very. It’s interesting because Schrodinger was able to predict the structure of DNA, talking about the constraints of the laws of physics for being able to have some kind of genetic heredity. And von Neumann kind of did similar logical arguments about what are the necessary conditions in order to have something that could self-reproduce. And both of them ended up predicting really fundamental things about biological structure before they were discovered. So, Schrodinger basically predicted that you need to have an aperiodic crystal, and DNA is an aperiodic crystal, and von Neumann predicted that you need to have something that can take these instructions and be able to build things, and that's sort of the architecture of having DNA be the instructions. When you have DNA copied by a polymerase, it doesn't care what the information in the DNA is contained, and then when you actually read out the DNA, you have something called a ribosome that can make, in principle, any possible protein. So, it's not exactly a universal constructor, but it's universal over the set of proteins. Right? So, it has that kind of logical architecture. So, it's really fascinating that that actually ends up being the case. And then, interesting to me, so, I'm a big fan of something called constructor theory that David Deutsch and Chiara Marletto have been working on, but Chiara wrote this really nice paper, basically doing what I think is kind of an argument like what Schrodinger did, where he talked about the laws of physics and what's necessary for properties of systems to exist, and she made the argument based on von Neumann's things that in order to have evolution with like, and the appearance of design with no design laws actually requires that kind of architecture and having digital information. And so, there's really this interesting thing, like the way physics is structured is actually telling us a lot about what life can and can't do, and what information can and can't do. It's just that people don't usually like, try to do these logical arguments to reason into it.
AW: Yeah, neat. So, just want to make sure I understand myself. So, you're not claiming that life itself is a universal constructor, you're claiming that this system of DNA and ribosomes and... yeah okay.
SW: It's universal over that. But there is an interesting... So, the way I think about evolution, which is very abstract, and what evolution is doing, is that if you look at the history over life on Earth, I think there has been an increasing number of things that can happen. So, if you think about what a constructor is -- a constructor mediates a particular task or a particular thing to be transformed into another thing, and it might be that it takes some materials and makes itself, or it takes the materials and make something else. And that's actually what a constructor is defined as in constructor theory. But if you look at the evolution of life on Earth, what I think is happening is that you're getting more and more constructors, or more and more things that can be constructed, because there's more and more information instantiated in those systems. There's actually an arrow of increasing number of transformations that the information on this planet can mediate, and there's no better example of that than our civilization and our imagination. Like, if you think about it, like science fiction, it's like, you know, actually, this is my favorite example, this is getting really abstract. Yeah, I talk to my students about this kind of stuff. But something that really fascinates me is, so through most of biological evolution, we increase the amount of things that can be done by recording history, historical information, right? So, we have a knowledge of past environments, and that allows systems to be more adaptable, and then they evolve the capacity to control those environments and make new chemistries happen and these kinds of things. But the thing about humans and what we do, we have something that we call imagination, and that's actually a really interesting kind of physical process, because we can actually imagine things that never existed, and then sometimes we can like, actually generate those things, right? So, when you think about like science fiction and like Jules Verne writing about rockets and things, like nobody thought those things could exist, but then because somebody imagined them, like, they actually like, they came into existence! And that's a really interesting kind of thing to be happening, and so, I think all of biological evolution has been this progression of information increasing what's possible, and you could actually talk about a physics that's sort of like an entropy over [UNKNOWN] in some sense, rather than in states that there's you know, more and more things that can happen in the future because of the particular information structure, and that, you know, us and our ability to have abstract thought and all of these kind of things just fit naturally in that framework.
AW: So, would you consider the human brain to be a kind of universal constructor? In the sense that we can imagine anything, really, right?
SW: But we can't build anything.
AW: We can't build it, okay. But we can build the thought of it in our minds, right? But that's not the same thing?
SW: No, it's not the same thing, it's actually physically instantiating it. So, I, where, I'm glad you asked that though, because where I was going with this with the universal constructor is I think we think of that as an individual object, but the way I think about it is that the biosphere as a whole has become a better and better approximation to a universal constructor, because the entire planet can do more things. You know, human civilization can do more things, but you as an individual really can't do that much. And so, but it goes back to this idea, information is really hard to tie to a physical object, and so, it's not like I can point to this thing and say, this is a universal constructor, because you have to talk about like, where the information is in the system and things. So, I think there's like, maybe biology and why it has this arrow (?) of complexity is that once you get the system going and you transition into this kind of physics, it's this unfolding of this structure that's trying to become, it's like becoming a better and better physical instantiation of something like a universal constructor, but it's, it probably, it would never get... It's an... So, David Deutsch has written about this but like, whether it's actually possible for a universal constructor to exist is a really interesting question, and we don't really have a good segue into that, because like, that says a lot about the laws of physics and what's possible, and we just like, I'm, approaching that question is really hard.
AW: So, there's a related set of things about the history of life on Earth, and those are very major transitions that we've seen, and the complexity of life, so, those would be things like the acquisition of eukaryotic organelles, or multicellularity, or, you know, transitions to land and more complex ecosystems. So, do each of those transitions involve some kind of acquisition of energy, or a new use of information and a sort of new way of using universal constructors? Or...
SW: Yeah, something like that, yeah. I definitely think so. So, there's that really famous paper by Maynard Smith and Szathmáryabout the major transitions and associating them with changes in information processing and storage, and I found that work really inspirational from this kind of abstract perspective, and talking about, but I think, if I wanted to quantify a major transition, I would say that you had a new scale on the system emerge that actually became causally efficacious over the lower scales. So, and so, you should be able to actually quantify that by the information flows in the system. If you start looking at some macro-scale properties, that they become more predictive of the micro-scale properties. And so, I did a toy model in an A-lifepaper years ago showing that that kind of transition actually can occur. But, so, I think yes, it definitely has to do with information -- everything in biology does, apparently everything. I mean, that's the goal, right? Is like, to try to unify everything. But those are particularly interesting cases, because you have like entirely... And like, so, a good example is like, right now we seem to be undergoing a major transition where you know, like, most people are living in cities, and there's sort of like this globalization, and we have the internet and things, and so, it's really like the transition of like, you know, local societies as global society, and that's a very kind of top-down transition too, because now, like my actions as an individual are influenced by something about the global state of the planet, right? So, it's like, you can have an affect like you know, a particular emotional response, and billions of people at the same time on this planet. Like, that's never been able to happen in history, but it's because, like, we now have these collective structures, and they're difficult to point to, because they're not like a thing, and so that's why they're really difficult to think about, but like, society exists, it is a physical thing, it's just not, you know, an individual.
MM: So, based on this sort of this, well is there any way to get to the number of transitions that have happened on the planet, from sort of, you know, the basic ideas about top-down control?
SW: Yeah, yeah, I think so. So, the way I think, so, let me see if I can articulate this well. But I think, like, if you want to, I think about life as like a bunch of partitions in a system, like, and those partitions are interacting. So, presumably, like, if I wanted to characterize you, I should be able to like, see how many partition, like, if I could look at you from this abstract perspective, how many partitions you have and how they're hierarchically embedded, like how many are like sub-partitions of another partition, and be able to say something about how many of those kind of transitions occurred. And I think that's sort of like, part of the origin of life process is that you know, the systems kind of had this partitioning of their like, in physically, like save their state space and also like their physical space, and those partitions became coupled and then you've got like, higher order partitions and yeah, now I'm getting very abstract and like, I, it's difficult to explain because there's a concept there, but like, and I guess this is useful also for listeners that like, when you're in these conceptual spaces, articulating things is really hard, but you have to push yourself there and like think about it to try to actually formalize it, and so, it's always a moving target, but that's fun.
MM: Maybe you want to pull back to the 20,000 feet? We often do a 20,000 feet pullback. So, physicists, chemists, have been incredibly good coming up with theory, and biologists just take a different approach, like we sort of start it. Do you think that there'll be theories in biology one day, as there are?
SW: I hope so! Yeah, I think so! I mean, so, I really think that biology is, well in particular astrobiology, because I think you have to talk about observations of other planets, but biology is the next frontier in physics. So, people think, if people think about physics as being subject specific -- so it has to do with particles, it has to do with gravitational systems or like these kinds of things, and biology has nothing to do with physics. But biological systems physically exist, and I think about physics as a particular way of thinking about the world. It's just when you're a physicist and you're thinking about things, you're trying to abstract them to a very universal and powerful explanation, and those tend to be very simple because you want something that can describe many systems. But getting to that level is really hard, and I think we just haven't gotten there with biology yet, but my hope is that we will.
MM: Do you have any sense of what they're going to look like?
SW: Um, yeah... I do, I mean I do in the context of the whole conversation, because like, from my perspective, since I've been thinking about it for a while, like I've gone down this particular road thinking that the way I've been describing these systems is whatever that level of abstraction is. That has something to do with information in systems, like you can think about constructors, number of transformations that can happen in physical systems. If you want to count that, that would be maybe where like the laws of biology are is one way I say, what biology is doing is increasing the number of transformations that can happen as a function of time. So, and that's interesting because of this point that it actually literally makes it possible for states to exist that wouldn't exist otherwise, like satellites orbiting our planet, or quantum technology, or whatever you want to...
AW: So, the flip side of that question I think is, are there going to be new physics that come out of, out of biology? And you know, Schrodinger himself suggested that in what is life, and I think earlier in the conversation you sort of alluded to that, that you as a physicist are going to biology because there's like this density of information that just doesn't exist anywhere else. So, I mean, what are the new physics that are going to come out of this?
SW: Right, so, so my reason for working, so, I was inspired by like, the founders of quantum mechanics or like, Einstein as a student. Like, they opened up entirely new fields of physics. And so, for me, I feel like we have quantum mechanics and we have general relativity, and there must be some theory that's equivalently fundamental that explains life and what life is. Um, and I think that, like going back to things I said earlier, I think that's information. And so we don't have a theory of information and what information is in the physical world, and if we did it would be explanatory of what life is.
AW: Has an implication that there's going to be some portion of biology and physics are going to fuse in the future and that they're going to become sort of the same field, really?
SW: Yeah, yeah!
MM: So, I guess, let's make it the kind of question, this is just what, like the virus one I know you're getting a thousand times...
SW: Sure! No, it's okay, I like them! I answer them differently.
MM: How common is life in the universe?
SW: Oh yeah, that's a great question! Um, yeah, so I always answer this in two ways. I answer the like, optimistic, like Sara, and the one that like you know, the reason I'm an astrobiologist is yes, we will find life somewhere someday. And then I have sort of agnostic scientist Sara that has absolutely no clue, which I think, you know, you have to admit as a scientist. But I actually admit that from sort of logical reasoning that we have evidence for only one inhabited planet, and we wouldn't be on this planet observing that planet being inhabited unless there happened to be an origin of life on our planet. And so, if you actually...
AW: This is the anthropic bias argument?
SW: Yeah, the anthropic bias argument. Um, but um, and so, um, so if you do like proper statistical analysis of those facts, one origin of life, our existence is contingent on it, then it's, um, equally likely that life is really rare and we're the only life in the universe or that life is very common. Like you can't, you actually can't distinguish those hypotheses. So, so, I am, but this also motivates me, I just, I think we have no idea if there's other life out there, and we should just state it as such, because I think we would reason about the problem better, because what I see in my field a lot is a lot of astrobiologists will say life is common, and they'll make arguments like, oh look how many exoplanets that are earth-like we found, so there's so many environments, life must be common, because there's lots of places for it to form, or life emerged rapidly on Earth, therefor it must be common because it must be an easy thing to happen. Um, and both those arguments are false for the reasons I stated previously, is that there's, there's literally no evidence, and we don't know the mechanisms for the origin of life. If we knew the mechanisms, we could extrapolate to how those mechanisms would operate on other planets, but we don't know the mechanisms. And so, without having a theory or an observation of another living thing, we literally can't say anything about the likelihood right now.
AW: So, so how much would it change the calculus if we find life on Mars in the next 10 or 20 years?
SW: It would change it significantly! I mean like, if there's two planets in our solar system that are inhabited, that means that life is probably really likely.
AW: But it, uh...
SW: I mean, unless it had an origin from Earth.
AW: I mean, it seems like...exactly, so if they're related to one another, right? It depends on like, how closely related they are and whether we can, sort of, foresee a common LUCA. (Last universal common ancestor)
SW: Yeah, yeah. So, that's why like, sometimes I say my biggest disappointment is if we found life on Mars and it was just Earth life on Mars.
SW: Yeah, exactly. It would just, I mean it would just ruin it for me. Other people might be excited, ooh life can live on Mars, which is exciting, but it's not the same scale.
AW: So, given how far away stars are, talk about how are we trying to figure out whether there is life on other planets?
SW: Yeah! So, so there's a lot of approaches to that, and I work a lot, um, in the exoplanet community now, so I find their discussion fascinating about how to look for life. But the sort of popular, like kind of consensus thinking right now is that there is no particular smoking gun biosignature. Like a lot of times people used to think like, if we found oxygen in the atmosphere it would just be totally, you know...
AW: Methane's another one.
SW: Methane's another one, and then there's this idea that oxygen and methane together, because they're a particular kind of disequilibria would be evidence of life. But the problem turns out to be much more complicated than that because we only have the example of Earth and we're trying to use Earth biology on other planets. But even if we take Earth, you know, and we do the metabolisms we know on Earth, or atmospheric gases we know biology produces and we put them in models of planets around other stars or with other compositions, the space is so huge that it's just littered with false positives of things that look like Earth-like life but are totally not biological. And so, the exoplanet community is working really hard to try to build models where like, you can actually definitively say this is life or not. But my personal thinking on it - and we actually just, I just led a paper on future directions in exoplanet biosignatures where we make this argument, it's one of many arguments in the paper - but my feeling is that we really have to move from this idea that we're going to look at a single planet and be able to identify its atmosphere and characterize that life is on that planet, to thinking about looking for life more statistically, and that like, and when we think about modeling planets or the observational... We can get really little data from exoplanets, like, like, almost nothing, like a couple molecules in the atmosphere and then like, you know, you might know a few features about the planet, but it's, it's very minimal. Um, and then, um, so we don't even actually have a tight constraint on like, the composition of these planets, let alone the biology on them -- like, we can't even like, we don't have models for the planets. And then we don't know what life is on top of those planets. So, it's a really hard problem. So, but, but, a lot of times when people do like, planetary evolution models, they can build like, a statistical distribution of what they think planetary composition would be like. And so my feeling is if we started doing statistical searches for life, and we had some models for like what we would expect for the distribution of atmospheric compositions, for example, based on no life on those planets, and then we like, see something different, we might be able to say something about like, and constrain likelihoods of life existing in an ensemble, and that we wouldn't have to rely on detecting life on a single planet and then knowing exactly what we were looking for.
AW: So, it makes sort of more probabilistic arguments about...
SW: Yeah, basically. And a lot of people that are like, doing exoplanet stuff now are moving more in like basing inference approaches for inferring properties of exoplanets, and I think that we're going to have to do some kind of inference and statistical type analyses to actually detect life, and people haven't been thinking about it like that because there's always this implicit assumption that we know what life is and we'll know it when we see it, and to really think about it as a scientific problem, you have to completely change all of your conceptions about being able to recognize it, and you have to think about, you know, actually doing inference on a system and like, you know, trying to build up these models for like how likely life is on this planet as the valid hypothesis, and so, so I think the community is starting to move in that direction. I find that really exciting, because, um, like, you know, I always make these analogies with physics with astrobiology, but like, I think about it as, you know like, I want to know the distribution of life out there, not just whether we have life on one planet, so if you do these kind of like statistical searches, you actually can constrain with an ensemble, like how likely is it that, you know, a certain percentage of them have life. And that's more informative, actually, from this anthropic or these statistical arguments, for theory building or for like, you know, actually constraining what we think this process is, and how often it happens on planets and things.
AW: So, if there is life out there on other planets, how likely is it that there's also other intelligent...
SW: Yeah, so, yeah, I don't know how coupled those questions are. From my perspective, like, I just have a gut feeling, and usually like, most of my science is initially based on gut feeling anyway, because you have to be passionate enough about something to like, try to push it and see if it's true or not. But um, but just my gut feeling is that once the origin of life, so the origin of life is the hard question. How do you get this process started in the first place? And I almost feel like it's the quantum to classical transition where you're talking about two different, totally different physics, and somehow you have to go between them. And so there's like, information doesn't really matter and suddenly it's like, everything. And how you go between those is really hard, but I think once that process gets started, it's like an unfolding process and it keeps building on itself, so I think, I think most planets would evolve toward intelligence, but...
MM: So, does that hold for consciousness too?
SW: Um, I think so, but I, like, that's a really hard question too. And I, and I almost think, so, so, I think consciousness is a different problem from life, um, but I do think that they're kind of, they're related in some way, and so the way I think about it is like, so the hard problem with consciousness is why we have internal experience, and that's really hard because that you can't explain in any kind of substrate level narrative. There's absolutely no reason that we have to have an inside, or be thinking about the world. And from my perspective, the life problem is like, why is it that your experience would actually matter to the world, or any kind of internal information processing? So, they're kind of dual in some sense, one is like the internalization of like, you know, information and what it's like to be on the inside of that information, and then the other one is like, why does that actually have a physical influence on the world? So, I don't think they're wholly unrelated, but I do think that, that, right now, unification like, so, I think life will be solved before consciousness at this point. Consciousness might be in the next century. I think AI will help a lot with that. And then I think down the road, like a couple hundred years, the unification problem people will be talking about is consciousness and life, from physics. Like people were worried about GUTs right now, grand unified theories, like, pssh wait a few hundred years and get way more interested... Physicists, physics will get so interesting in the next century.
MM: So, you did, you brought up AI a couple of times, but not really said much about it. So, the way that you're thinking about life, what is it, does it say anything for how we might design AI, especially to avoid big problems with AI?
SW: Oh sure, yeah, so, so, I am, I'm interested in the AI control problem, but I don't know how to approach that problem yet from my thinking. So, I've talked to a lot of people in that community, and I'm intrigued by the problems that they have, but I also am thinking about it from a totally different perspective, so bridging that right now is difficult for me, but um, but I do think, I think... So, I think there's a lot of things that we think are unnatural about what humans are doing right now, like climate change - we think it's unnatural. But, if you look at the history of life on the planet, you know, life has changed the climate dramatically at a planetary scale many times. The only difference now is that we are cognitively aware of it and we actually have the ability to do something about it. And that's a very different system, but the climate change itself is not unnatural. It's a natural byproduct of biological activity. And so, I think AI also, people think it's so unnatural or our technology is so unnatural, but if I look at the evolution of life on Earth, all life has been doing is building increasingly sophisticated informational architectures. And so, I think AI is just a continuation of that process. And I also think it's a particularly interesting one, because I think, I think AI will be critical to like, actually identifying the laws of life in the sense that it's very difficult for us to see ourselves, but AI is not us, it's something we created. And so, they can actually see us in ways we can't see us. It's sort of like building like a, we built microscopes to see cells and telescopes to see the universe. I think AI are going to see life in a more fundamental way.
AW: And we just got to hope that they don't gain top-down causal control over us, right?
SW: Yeah, and so, I'm very much more thinking that it's going to be kind of some kind of hybrid system. And I actually also don't have any qualms about like, our progeny being AI, or...
AW: Despite my negative comment, I don't either.
SW: Yeah, or technology...yeah. Because I think, I think people, people think, oh we're going to be cyborgs in the future, we're going to be robots. But I'm like, but we created them, it's like, you know, they're still part of our lineage.
MM: Just another place to live, right?
SW: Yeah, and I think, yeah, and to think that biology like, I mean humans are going to continue to change. There's no species on this planet that has existed the entire existence of the planet. All we're doing is increasing the longevity of the systems by building these. So, so, I actually like Nick Bostrom's, he made some argument at one point about, you know, all species go extinct, so AI is either going to make us immortal, or it's going to kill us, and those are like the options, right? But, but we're not going to last forever anyway, and so, so, I think it is critically important, and I really like all of the stuff like Future of Life Institute and stuff are doing to try to make sure that AI is positive for humanity, but, but, that's also hard to understand what that means. And so, so, one of the, so, I'm actually working with one of my students, and she's in Japan right now presenting on "Major transitions in planetary evolutions" is the name of her paper, but the idea there being that the planet as a whole has undergone like major transitions to informational processing structures, and we're undergoing a major one now, and AI is a part of that, and we need to think about AI as a planetary scale system, because our technology is a planetary scale system, and so we have to think about it how it fits in a planetary context and how that fits in the continuation of... And so, I guess from the control perspective, like, going back to my earlier comment, my philosophy of how I do science is always to try to find a question no one is asking to make an inroad so that we can ask things a little bit differently, and AI I think there's a lot of really great questions, a lot of really great intellects working on that, so I think what we're trying to do is make this inroad from the planetary science perspective, and sort of the astrobiology perspective. But that's very early right now, so maybe in a couple of years I'll have more to say about that.
AW: And I got to ask one more thing, which is, you know, at what point, how soon do you think we're going to consider AI to be alive?
SW: Yeah. I already think they are. Yeah.
MM: A screwdriver's alive, then.
AW: Yeah, well, well, a screwdriver's not alive, but, but, it's life, yeah, because it's not doing anything, but, I, you know, sometimes, so people are like worried when AI is going to take over and I'm like, it already took over, I mean like, like, so much of your life is dominated by technology. It's like, I don't think it's a bad thing, it's just a different thing. Like, it can be bad, just like anything can be bad. You can eat too much chocolate and it's bad for you, but everybody loves chocolate!
AW: So, when is AI going to have consciousness, in a sense?
SW: Well, that's a different question. That's a totally different question. And, like I said, I don't think we understand consciousness. I don't, I'm not entirely sure if... I think you have to have the physical media right for consciousness, and I'm not, I don't know, I don't know.
AW: Maybe computers aren't designed in quite the right way yet in order to facilitate that?
SW: Yeah, yeah.
AW: Because of connections, and there's not enough of a simulation, so the neurons, or...?
SW: Yeah, so for example, like, it might be that like, people building, you know, more complex chemical systems actually build conscious systems first before silicon chips, just because of the dynamic possibilities available in those systems. So, so, I think, I think, yeah, we don't know enough about consciousness for, and I think about consciousness a lot, but it's a periphery problem for me, because I'm mostly focused on the life problem. So, I don't think I have something super concrete to say about that except that we don't know what consciousness is, and people often convolute intelligence and artificial intelligence with consciousness, and AI is not about consciousness. And, you know, we have a difficult time assessing consciousness in people, so how are we going to do it in machines? I think that's, it's a really hard problem, and I think people kind of mix the issues there a lo
Michael Garfield: Thank you for listening. Complexity is produced by the Santa Fe Institute, a nonprofit hub for complex system science located in the high desert of New Mexico. For more information, including transcripts research links and educational resources, or to support our science and communication efforts, visit santafe.edu/podcast.