Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Episode 8: Artificial Life
9 Plays5 years ago
Can we create life artificially? What would that even mean? Rolf and Joe talk about the field of Artificial Life, or "A-Life", which has worked toward the goal from a number of academic disciplines for the past thirty or so years. They think about different approaches, such as software, hardware, and biological artificial life, and consider what it might take to convince us that we have created something that would be considered life.
Recommended
Transcript
Introduction to Cognation Podcast
00:00:06
Speaker
This is Cognation, the podcast about cognitive psychology, neuroscience, philosophy, technology, the future of the human experience and other stuff we like. It's hosted by me, Joe Hardy. And by me, Rolf Nelson. Welcome to the show. All right.
Artificial Life: Past, Present, and Future
00:00:25
Speaker
Ready to talk about artificial life.
00:00:30
Speaker
Article that we are looking at today or that we're basing our discussion today on is called the past, present and future of artificial life. This is a review article from a couple years ago, 2014, and it takes a look at the state of the field for artificial life. What is artificial life? What is artificial life? Basically, it's like life made by
00:01:00
Speaker
a person rather than by nature. I found this article in this field kind of fascinating. I have to say jumping into it, I didn't know a whole lot about it.
Is Life Necessary for Consciousness?
00:01:12
Speaker
So I had to do a little background checking. I don't know what your background is on this show. Yeah, I didn't know anything about it either. I stumbled on this article because I was thinking about the question we were talking about. What would it be like to add consciousness to a machine? Would that be something that
00:01:29
Speaker
would want to do, if you did want to do it, how would you do it? And I was looking for stuff around that, like how would you build consciousness? And that's how I stumbled into this topic area. I think it's an interesting, maybe that can be a section that we sort of devote to the relation between life and consciousness, because I think that's an interesting one, is that
00:01:52
Speaker
To what extent is the foundation of some sort of life necessary for consciousness and how much you need to know about the way that life works? So maybe we can talk about that in a bit.
Origins and Evolution of Artificial Life
00:02:05
Speaker
But first, I found the background of this field of study of artificial life, or they call it a life, as in get, I suppose, as in get. That's what we need to do.
00:02:21
Speaker
So the field itself, it's sort of a separate interdisciplinary topic that kind of has its own own journal culture to it. Yeah. It has its own journal. I had to look at the journal a little bit and see what was going on with that. So as far as I could understand, so it basically kicked off in the mid eighties or so. Right. Right. And Langton.
00:02:49
Speaker
Christopher Langton, he came up with all of these brilliant ideas around complicated systems and how to conceptualize the idea of artificial life, what its goals should be, all that kind of stuff. And he did research for a while. He went to the Santa Fe Institute, which is kind of a high level think tank for really super smart people who think about issues of complexity.
00:03:17
Speaker
not just complex issues, but literally the issues of complexity. Then he stopped doing research. Presumably, he's still alive. Maybe he's even listening to this. Late 1990s just basically dropped off the face of the earth, does not seem to exist on the interwebs anymore. I could find no mention of him.
00:03:46
Speaker
You know, what happens to a field when you kind of lose its founder? It continues on, but it seems as though the field itself as a coherent. Culture has has sort of it's dissipated a little bit. So to me, I thought this is interesting. Just kind of thinking about the.
00:04:04
Speaker
the evolution of a field itself and how it was formed around this particular individual and had all these great ideas and continues to, I mean, there's a conference every year, although it sounds like maybe a little smaller than it used to be, not as much in the spotlight, but has sort of tapered off a bit. That's really, yeah, that is really interesting. I think he might've been destroyed by his own robot. Oh.
00:04:31
Speaker
Yeah, what did what? Maybe he maybe he came to know too much. Ah, maybe that's it. That must be that must be probably it was one of those research findings that was too mind blowing that he just he couldn't keep going. It couldn't be introduced to the world. Or he somehow actually entered another plane or dimension of reality and he's actually living there. He's actually the same age. He hasn't aged. That seems likely.
Cultural Impact of Artificial Life
00:05:00
Speaker
Well, after this, let's go find him. Yeah, I bet he's I bet he's impossible to find. Sounds like I looked at I found some Reddit speculation about him that he had hobbies in boating and maybe he's out making boats now. Christopher Langton lost at sea. Yeah. Or maybe he maybe he created an Atlantis, some kind of under underwater underground layer.
00:05:30
Speaker
created, he created life and he's living in that underground Atlantis with his whole artificial life society. Yeah. I think it's, I think that is what it's going on. That's because if you, I mean, if you did that, would you want to invite other, I mean, you just created your own whole world of artificial life that's made by you and exactly what you want it to be. Are you going to invite anybody else or let anybody know about it?
00:06:00
Speaker
This is right for a movie. This is a movie making opportunity. Yes. Or maybe even a book opportunity finding Christopher Langton. All the publishers and producers out there. Send us an email tweet tweet us. So.
Defining Types of Artificial Life
00:06:16
Speaker
the concept of artificial life or thinking about artificial life, they talk about, you can do a Google search, basically a Google search on text for artificial life. And it was super popular in the 80s and 90s, which I guess is, you know, around this time Christopher Langton was doing his work and the field was sort of had its most influential ideas. And then
00:06:44
Speaker
but it was even more influential right around 1818. See if you can look at text from 1818 right around when Frankenstein came out by Mary Shelley. There was this sudden spurt and interest in artificial life at that moment. Frankenstein, he gets life by just shooting a jolt of electricity through him.
00:07:12
Speaker
Hopefully the ideas of the 1980s and 1990s are a little more complex than that, but I think it has the same sort of appeal to it, right? That you can you can create life where you have this ability to. Well, I guess there's not a field of reanimation studies, but. There should be. Yeah, there should be. Well, I mean, Frankenstein does get into some interesting philosophical questions about yeah, absolutely right? What is the nature of? Of a person?
00:07:42
Speaker
Yes, it is. Is it the same? Well, we should refer to him as Frankenstein's monster, since, of course, Frank is a doctor, right? Right. So the monster, you know, does the monster have consciousness? Mm hmm. Does the monster feel certainly in the book? I think it makes it out that it comes across that the monster does does feel right. Mm hmm. And then then the question becomes becomes is the monsters feeling most like the brain that was, you know,
00:08:11
Speaker
Right, because it was a patchwork of different brain and body parts. Right. Revivified parts from different dug up. Or was it just cadavers? I don't know. I don't remember. I never read the original book. I think most of what I remember about that is from Young Frankenstein. Right. I did read it, but it's lost now.
00:08:39
Speaker
So the culture of this field of a life or artificial life. So again, it does seem to be this kind of distinct group of ideas. So it's an interdisciplinary topic.
00:08:55
Speaker
I found it sort of interesting to look at some, I know this is just my way of trying to go out and understand it, looking at some of the recent papers in this area from the conference, the most recent conference, and the ones that appear in the journal. Just to get a size of the field.
00:09:17
Speaker
Conferences now consolidate. There are a couple different conferences and now it's consolidated into a single one that sold once a year. About a hundred papers slash posters in it. Moderate to small. And I did find that a lot of the titles of the papers sounded as though they were a little tangential to the idea. So it doesn't appear as though there are that many researchers out there who are specifically
00:09:44
Speaker
getting into a field where they're trying to create life of some sort. In that context, it makes sense to dig in, I think, a little bit into what is life, what is artificial life, and what are the different types of artificial life? Yeah. Maybe a little bit.
Characteristics and Philosophy of Life
00:10:02
Speaker
They define a few different types of artificial life, where there's soft artificial life, which would be like simulations in a computer, basically, so software.
00:10:13
Speaker
hard artificial life, we have hardware like robots and wet artificial life, which synthesizes living systems from biochemical systems, biochemical substances. So the idea that any kind of like wet lab stuff would be like wet a life, any kind of robots or or any other hardware implementations would be
00:10:40
Speaker
hard a life and then software implementations would be soft a life. So basically like a computer simulations, robots and Frankenstein's. That's right, that's right. I think the Frankenstein's there they're talking about are really kind of much smaller systems where you're trying to basically figure out how to create properties of lifelike systems.
00:11:06
Speaker
So they're really working with molecular, at a molecular level rather than plugging arms into sockets and stuff like that. Right. They're not working with string, with like sutures and that kind of thing. Yeah.
00:11:22
Speaker
It's probably worth thinking a little bit about what we mean by life and what are some of the characteristics because it helps frame some of the discussion. One of the concepts is that life is self-reproducing and self-sustaining. I wonder, I was trying to get a sense of, it's a really interesting question. What exactly do we mean when we say that something's alive?
00:11:52
Speaker
maybe one of the goals of a life is to kind of sharpen the ideas and our intuitions about what life actually is. Right. It's hard to think of a set of necessary and sufficient conditions for life because all of the ideas that we might come up with like that it has to be self replicating. We could imagine a system that is self replicating but is not
00:12:19
Speaker
It doesn't really feel to us like it's alive, but I guess we kind of have a cluster of conditions that we describe that sort of apply to life. And maybe I think the definitions that are used here are intentionally a little bit vague because we don't want to get pinned down too much in saying that life is one thing or life is not another thing.
00:12:43
Speaker
Right. I mean, the idea of self replicating already is, is, is problematic because it is neither necessary nor sufficient for life. To your point, you have a self replicating system that is not alive. You could have an a lot of something that is alive that could not replicate itself. Exactly. Billions and millions of examples. Just take a mule, for example. So a mule is a hybrid of what a donkey and a horse. That's right. Cannot have offspring, but yet is clearly alive.
00:13:14
Speaker
It is absolutely alive, no question about it. A person who doesn't have the ability to reproduce is clearly alive. Just an old person. Right, right, yeah, past their prime. Yeah, exactly. So that's already not very helpful. Well, I thought about this. It's clearly a part of the concept of life, that there's some kind of self-replication that in order to compete in
00:13:42
Speaker
the environment for evolutionarily, you have to be able to reproduce yourself. I think in computational models, it's easy to get something that can reproduce itself. Within a computer program, you can have something reproduce easily. But it's not obvious to me that it's easy to make something in the real world right now that can reproduce itself and propagate. I wonder if there are any examples of that even.
00:14:11
Speaker
an artifactual tweak of something that was already alive, right? Something that was not alive, but now you made it alive, and it's truly reproducing itself. A robot that can make another robot. Right. And that's the kind of thing that you would need to really say that you've created. I mean, you can have a robot that was alive, but wouldn't be able to replicate itself. I think. You can imagine that.
Embodied Cognition and Enactivism
00:14:40
Speaker
You can certainly imagine that.
00:14:42
Speaker
But then I guess what they're saying is their goals are just a little bit more aggressive. Sure, that would be a lie, but I guess what they're saying, that's not the same as artificial life. We want to create a whole society or we want to create not just one organism, but a whole set of organisms that can evolve and learn, act autonomously. Well, and certainly all of the creatures that we think of as alive now came from
00:15:12
Speaker
an environment that was self-replicating. Even a mule comes from the same kind of environment that produces a horse and it produces a donkey and stretches way back into original evolution of unicellular organisms and all that. It's along the same continuum. No, exactly. I think that raises a good point, which is when we talk about artificial intelligence, we're always talking about this, or not always.
00:15:42
Speaker
So often we're spending a lot of time talking about this very disembodied process. I think the flavor of what this a life thought processes is getting at is actually life is by its very nature, something that exists in an environment and interacts with that environment to really feel like something is alive. It somehow needs to interact and evolve with within its environment.
00:16:10
Speaker
I think that makes a lot of sense. It seems as though this field came about as people are having these sorts of realizations about artificial intelligence. I know that Rodney Brooks was part of this structure. Rodney Brooks, are you familiar with him? I'm not, no. He's an MIT professor. I think he's retired, but he's done a lot of robotic stuff.
00:16:36
Speaker
There's this really cool movie a couple of years ago called Fast, Cheap, and Out of Control. It's called Fast, Cheap, and Out of Control because that was the name of one of his papers that thinks about robots as swarms that are behaving in the environment rather than programmed symbolic systems. That it's really about, yeah, just like you say, being embodied and being out there in the world.
00:17:06
Speaker
interacting with the world, and that's how complex behaviors are formed, not by programming them explicitly. That relates to this idea of embodied cognition, that cognition thinking is a huge group of people now that self-identify as
00:17:25
Speaker
embodied cognition folks that I think is really important to take a look at the environment and consider the importance of the world when we're thinking about how we think rather than just thinking of our brains as this separate kind of disembodied abstract kind of idea. The abstract disembodied brain can solve a lot of problems and I think a lot of the problems that we're
00:17:52
Speaker
thinking about when we think about AI systems today in computers, for example, natural language processing, I think is a great example of, can be really disembodied and work pretty well and do some interesting stuff. But it's also quite limited and not at all like what our brains mostly do, right? Most of what our brains do has nothing to do
00:18:20
Speaker
with solving symbolic logic problems. It's funny and it feels like that's so important to being human, but it turns out that you can do it entirely differently. What do you mean? Well, like chess, right? It feels like this
00:18:42
Speaker
human ability and I think we've talked about this before that you can play chess that's nothing like how a human would play it and do much better even though it seems like you need rationality and reason same with language like you're talking about language processing that you can do this fairly well in an abstract sort of way that's nothing like the way that human beings actually do it and it still works right right absolutely
00:19:08
Speaker
Some of the stuff that I find interesting in this topic is based on emergent kinds of behaviors for creatures like ants or bees or swarms of birds, so that you get this kind of emergent organization that comes out of individuals. Seems to be a popular topic in this area. Yeah, very much so.
00:19:35
Speaker
One of the terms related to this area that came up, and I mentioned it to you in this email that I sent earlier, was the idea of inactive or inactivism. And so I just came up again in this paper, and I think it relates to this whole thing, and I'm just looking at the Wikipedia page right now. It says, inactivism argues that cognition arises through a dynamic interaction between an acting organism and its environment.
00:20:00
Speaker
It says it claims our environment is one which we selectively create through our capacities to interact with the world. So it's really emphasizing this closed loop sensory, cognitive, motor interaction between the agent and the environment and how cognition itself doesn't make sense outside of that environment.
00:20:25
Speaker
Yeah, that's a tough one to jump right into. Personally, I identify with some of the goals and some of the ideas behind this, but I don't understand how this can be the full framework for understanding how cognition works. Not to digress too much here, but people in this area make really strong claims about how much this is true, and particularly that if we
00:20:53
Speaker
could not interact with the environment if we couldn't move around and sort of poke things and move around in the world that we literally would not be able to perceive or understand anything about the world itself. I think this makes weird predictions that
00:21:11
Speaker
Someone with locked-in syndrome, someone that's completely paralyzed wouldn't be able to actually perceive anything at all because they're not able to move. And it takes motion in order to perceive in the first place. So I feel like some of the strong claims of this are, to me, obviously false. In perception, we're familiar with J.J. Gibson and the idea of ecological optics that the way that perception works is
00:21:41
Speaker
hugely informed by the environment. So if you want to understand how perception operates, you should not necessarily be looking at the brain, but you should be looking at the kinds of information that come to you in the environment. And that makes a lot of sense and goes a long way. But I do feel like it can't be the only story. And maybe this is partially my intellectual background. I don't sympathize with that view entirely.
00:22:12
Speaker
Direct perception is the term that we use to talk about the JJ Gibson worldview of the idea that the perception itself is given to you by the affordances of the information about objects and motions in the world. So the chair is perceived somehow directly as having the affordance of being sit-on-able. And I think it has a similar characteristic. I agree with you.
00:22:42
Speaker
100 percent in this world. I was thinking about Jay Gibson as we were talking about in activism as well. I sympathize with your view and I agree with you in the sense that it doesn't feel wholly satisfying. But I also think that it is useful and I think what was good about direct perception and thinking about the world that way is that it emphasized
00:23:08
Speaker
how important the structure of the world is and how our brains evolved to process that information and that structure. As you're thinking about artificial intelligence or even artificial life, thinking about the nature and structure of the world and how building a system that can be robust and evolve within that world, it's super important to think about the characteristics and the nature of that world.
00:23:39
Speaker
This is an important central topic in some of this stuff. I think this is probably something that was important to the thinking in the 80s and 90s too, probably central to it. Maybe we can jump. What we were originally saying, we're going to make a connection there at the beginning we were talking about.
00:24:01
Speaker
Oh, I don't remember. I don't either. I don't know. It was important, but that's OK. That's totally fine. I don't want to skip ahead too much. So one of the one of the things that I think is that springs up when I think about this stuff is the game of life.
Emergent Properties in Simulations
00:24:23
Speaker
Is this something that you're familiar with? So Conway's game of life. I was thinking about the board game.
00:24:29
Speaker
Ah, yeah, that's what I would initially think of. The game of life is this super simple simulation that's meant to show some kinds of emergent properties. And it's been a while, I guess, since I've seen instantiations of it. But this is something from the 70s or 80s, I guess, where
00:24:53
Speaker
It's sort of like a just a pixel array where each pixel affects the one next to it and you can have things that appear and they'll propagate and move and you can make simple kinds of behaviors so you can create something that will just kind of creep across the page.
00:25:15
Speaker
Just from this really simple set of rules, you can create some interesting emergent kinds of properties and little bits will interact with each other and sort of zip around the screen in different sorts of ways. You can create things that will destroy other things that they'll come in contact with. So that's one of the things I initially think about when
00:25:40
Speaker
I consider this idea of artificial life is that kind of a simulation where you're trying to create initial starting conditions and an environment that you can sort of stir up and watch, watch things that are interesting happen. That one is a really basic one. And I know that if you, you know, if you look up Conway's game of life, you'll be able to find versions of it that you could play around with or, or sort of understand what I'm talking about.
00:26:08
Speaker
it gets across this basic idea where you get an evolution from certain starting conditions and watching emergent behavior appear. You can think about this as the emergent stuff that you get from brains being wired up in interesting ways or from social animals living in groups, whether it's ants or bats or people that display interesting kinds of emergent behavior that
00:26:37
Speaker
might affect the way that they evolve. So anyway, that's something worth checking out. Yeah, and I think that's exactly the kind of thing that we're getting at here with trying to figure out if we can create a little microcosm where life-like properties can happen. We can learn something about the nature of life, what it is, how it works, and then maybe we can create something
00:27:04
Speaker
that's useful also. Or something that will destroy us all, you know? Or something that will destroy us all. Did you see that news about the scientist in China who edited the genes of that twins so they wouldn't have HIV or be able to get- No, I did not see that. I heard something really vague about that. So what exactly happened? So yeah, you use CRISPR. CRISPR, of course.
00:27:32
Speaker
Of course, everything's all about CRISPR these days. All CRISPR. All CRISPR all the time. But yeah, so he created the first gene-edited humans. So these are alive existing human beings? And he was like, yeah, this is cool because now they're not going to get AIDS. I mean, they're little babies. I mean, sure, they might be exposed to AIDS, but they also might not. There's lots of other ways to avoid it. But then it turns out, oh, sorry, I missed.
00:28:03
Speaker
He missed. Do it quite right. They did get and they did get AIDS. No, I know it wasn't even that. It's just like they could they could be exposed to other genetic anomalies, basically. Well, if there's one thing that Jeff Goldblum taught me, it's that you just never know what's going to happen when you start playing God. That's true. That's true. You could end up like some crazy looking fly situation. Yes, the fly and then Jurassic Park. Right.
00:28:31
Speaker
Actually, some of this stuff coincides with right around that time the fly in Jurassic, the original Jurassic Park, too. Yeah, Jurassic Park is a great example.
Artificial vs. Reanimated Life
00:28:41
Speaker
I mean, it's not artificial. Well, I mean, that begs the question, is that a life or is that not a life? I guess they did do some modifications, right? They didn't, they took these fossils, basically pulled out DNA out of the fossils and then kind of replicated it. So they didn't create the DNA, but
00:29:01
Speaker
It certainly created the environment that allowed the whole thing to happen. Well, certainly, if it's not artificial life, I would say it's probably not considered artificial life because it's so similar, but it fits within this idea of an organism existing within that larger environment and that larger ecosphere. So things are pretty well tuned right now. Well, not with people around, I suppose, because we're killing off all the rest of the life on the planet.
00:29:29
Speaker
Things are otherwise pretty well tuned so that things survive in harmony fairly well, and then you pop something else in the environment and results are going to be pretty unpredictable. Rise of chaos theory and popularity of chaos theory was around the time that A-Life was popular too, so I think this is all pretty closely related.
00:29:53
Speaker
And these are these these most of these people that are doing this are computer scientists, too. Other fields got into the game. Biologists get into it. But a lot of this stuff was computer scientists thinking about this and computer science for
Strong vs. Weak Artificial Life
00:30:09
Speaker
sure. I mean, when you talk about soft A.I., especially and even hard A.I., where I mean, sorry, sorry, soft a life and hard a life. So robots, too. Yeah, robots, software.
00:30:22
Speaker
And then that's where you get the question of, at what point is the simulation life? And at what point is it just a simulation that has lifelike properties? And they have this distinction in the paper between strong a life, which is saying, well, life itself is a property of these autonomous agents. Therefore, if you create systems that have life
00:30:53
Speaker
like properties, then that is also life. You just have to call him alive, right? And then versus week a life, which is saying no, it's just a simulation. You can make a simulation that operates very, very similarly to water molecules. You can simulate water molecules and has all the relational properties of water molecules, but the computer doesn't get wet.
00:31:18
Speaker
It's not actually wet and this is directly. This is directly lifted from strong AI and weak AI. So that concept that.
00:31:27
Speaker
If you make something that's functionally equivalent to a brain and acts functionally the same as a person, then you've got to call it conscious. That's the whole concept of strong AI. If you can program a computer to behave exactly the same as a person, then you've got to call it a person and grant it consciousness. So I think this AI life follows in that. That same thread. Yeah, absolutely.
00:31:55
Speaker
How do we feel about that? I feel like there's something that seems much more tractable about life. It's funny to say that life is an easy problem or relatively easy, but it certainly feels more tractable than problems of consciousness and awareness. We have no problem saying that animals are alive. We have zero problems.
00:32:25
Speaker
it's just easier for us to accept that something else is alive than for us to accept that something else is conscious. It feels like there's so much progress in understanding how life works that the question itself doesn't seem as important. So the concept of Elon Vital, the idea that this is what people used to think about life is that there was some
00:32:51
Speaker
substance that inhabited a living body and then when you died it just kind of was no longer there and life was, life just consisted of this substance, you know, sort of maybe a kind of a spirit that disappears and then you're dead. That concept itself just goes away once you, once you start to understand more about metabolic processes, once you understand more cellular mechanisms of how life works, once you can explain all this stuff in
00:33:22
Speaker
reductionist terms, you can figure out how the organs of your body work, how the cells that make up these organs work, and you understand it from a molecular level, it all fits together fairly well. And you can explain it without having to resort to some sort of mystical process or some mystical distinction between something that's alive and no longer alive.
00:33:48
Speaker
And so it's not easy work. There's a lot to it. It's complicated process. Obviously there's a lot going on in something that's alive, but it all seems tractable even though we can't fully mimic a living system. Yeah, well a lot about how life works. Yeah, we understand a lot about how life works.
Complexity in Biological Systems
00:34:06
Speaker
It's also living systems can be pretty simple. Doesn't feel to us like conscious systems could be simple, right?
00:34:15
Speaker
That may be just a fundamental misunderstanding on our parts, but it doesn't feel like that. But to your point, I mean, maybe it's the case that all living things have consciousness. Some people make that argument, in which case it's not really different. Right. That's one way of resolving things, saying that, well, OK, that's just a cut off. It's alive. It's conscious.
00:34:39
Speaker
To some degree, even if it's a tiny amount of conscious, a bacteria is still conscious because it's alive. It's interesting to think about even simple systems can exhibit a lot of complex kinds of behavior. Let's name it that worm C. elegans. Neurons in the order of a couple hundred, nothing compared to what we have, but still difficult to understand.
00:35:05
Speaker
The behavior that emerges out of it is still nearly impossible to fully model correctly. Right. Apparently it's a nematode. Nematode. Nematode. Small little guy. Yeah, no, I mean, these things are complex and they're difficult to capture models. So even like a computational model that would even take into account all of what's going on in C. elegans is beyond what we can do today. I mean, and then the question is,
00:35:35
Speaker
Okay, so let's say you modeled it. All right, then let's put this together. So let's say you modeled C. elegans perfectly. The software would not be alive, would it? So if you've got a worm that is squiggling around on your screen and it's specified to the level of neurons, you may be even smaller. Yeah, I mean, to the level of every single... You can do every atom if you want.
00:36:05
Speaker
Every atom, let's do every atom. Yeah. Good at every atom is fully specified, but it's just hanging out by itself.
Challenges of Simulating Life
00:36:12
Speaker
It's even hard to even think about what that even means because right. What is the simulation doing? I mean, you could say, well, we simulate it to move as it would normally move, but it would just be then random. This is where you get into the embodiment thing because you have to put it into some sort of environment for it to react to. Yeah.
00:36:32
Speaker
It does start to get out this idea of like, it's interesting. Okay. So yeah, you have to put it into some sort of environment that it has to react to something. Let's say you put it into an, you know, and they simulated the entire environment of say, like a small, small puddle of water. Wherever nematodes hang out. Yeah. I feel like they hang out maybe in like water. I don't know. Let's start. Maybe mud, something like this. They definitely don't, they don't fly, I don't think. No, they don't, they're not flying. So let's, let's call it mud for the, and so like,
00:37:01
Speaker
It's got a little mud in there, but it's by itself isn't it? Just with mud. So maybe whatever it eats. So let's say it could simulate it eating all these artificial. You have simulated software things that it eats. And it like metabolizes them and it would get diseases and and live and die. Is that alive? Oh, you mean so if it's still a simulation on your computer, it's in the computer. Yes, in the computer can't leave the computer.
00:37:31
Speaker
So how far do the horizons of its world go? I mean, let's just say that we gave it a small puddle of mud. And right now it's just all by itself. It doesn't have any other creatures in there. So is it possible that the puddle of mud could be struck by lightning from a cloud that happens to be passing by? Sure. Maybe like a cow or something could step on it.
00:38:00
Speaker
I think if it's just by itself and what you see is that as soon as you put it by itself, you have to describe what by itself means. If it's not with other creatures, you have to think about what other creatures mean in this world. If it's eating things, it's probably eating something that was at some point alive. The food source has to come from maybe somewhere outside of the mud.
00:38:30
Speaker
I feel like as long as you're constraining yourself in that way, it doesn't feel like it's alive. I guess what you're trying to mimic is all of the conditions under which it has evolved, right? Anything relevant to how it can reproduce. And also interact with other creatures feels important somehow. Right, behavior is so avoiding getting eaten. Yeah.
00:38:57
Speaker
A lot of its interesting behaviors are going to be related to other autonomous agents. Parasiting off other stuff. Eating stuff. So then if you had like a mini C. elegans society. I don't know. It's starting to feel a little more. Now it feels like it's alive. Not quite. Not quite. I don't know. What do you think? Let's say they could like replicate and they could replicate in ways that
00:39:27
Speaker
that was unpredictable. This is a tough one because I have to admit that I do have a little bit of biological bias here, and I don't know why that is. So if you have it on a computer, it's the information about a creature that's important, purely functionally all the stuff that it can do. But I do have to say that it's still hard to imagine it as being alive if it's a simulation on a computer.
00:39:55
Speaker
Even though all of the relevant stuff should be there. Well, I guess that's the question, right? I think I think that what what I'm thinking is that all the relevant stuff cannot be there because there's all kinds of things that are not there.
00:40:11
Speaker
You have ones and zeros that represent three-dimensional, four-dimensional structures in space and time, but they're not the actual structures themselves. It's not the same thing. It's just not possible to actually perfectly replicate. It doesn't mean anything to say that you perfectly replicated. I think that's just your 4D bias. Yours is biased towards things that are in four dimensions. Right. Exactly, exactly.
00:40:40
Speaker
Yeah, but doesn't feel like I mean, maybe someday you could something would come along would make you feel differently about it, but I agree. It doesn't feel like. It's even possible to put all the quote unquote important stuff in there.
00:40:53
Speaker
I think that our bias towards feeling like these simulations are our intelligence or our life has to do more with just how our own consciousness tells us that these symbolic things are super important. Right. We may judge it by how similar it is to us and it's difficult to, um, a little difficult to move past that point. Right. Right. And then then, so then, okay, that's, that's soft, that's soft artificial life. So we're having a hard time getting there. Yeah.
00:41:21
Speaker
We're seeing some of them. I'm imagining these tiny little C elegans that are swarming around on a computer. And I'm imagining giving them a tiny little blue pill or a tiny little red pill to see if they can escape the matrix. Right. Which you choose.
00:41:50
Speaker
So that's like soft artificial life. That's soft artificial life. What if we started to build them out using non-organic stuff?
Creating Life-like Robots
00:42:01
Speaker
So robots, however we can imagine some artificially constructed non-biological thing that's out there, but out there crucially interacting with the environment. Yeah. And we've got a little, let's say we put a little Petri dish full of mud or whatever they like to hang out in.
00:42:20
Speaker
Uh-huh. Does it have to be, see, that's the question. Should it be mud or should it be artificial mud? I think that the robot version, I just think of some kind of robot with a camera and some way of moving around in the environment. So we're putting the robot and it's in our environment, not in the robot environment. I assume that
00:42:45
Speaker
Because that's the that seems like the whole advantage to this is that you can just pop it in the actual real world and see how it interacts with the rest of the world and learn and learn from it. Let's say if we can we can build a little little thing that looks a little bit like a nematode and kind of squiggles around and eats whatever nematodes eat just like made out of so made out of robot parts or a robot part. So it just knows it doesn't really get energy from the things that's eating. It's just kind of like
00:43:14
Speaker
filtering it through, but it can move around, still needs a battery. I mean, it's a Roomba at that point. Right. That's not alive. What would we have to do to make that thing be alive? I don't think I have a good intuition about that. Yeah. Except that it just needs a whole lot of behaviors that look like the kind of behaviors that living things exhibit. That's where it starts to become the thing about self-replication.
00:43:41
Speaker
Yeah, I thought that's where I was. That's where I was thinking, too, that, you know, you can mimic a few of the kinds of behaviors of living things, but self-replication is an incredibly complex process. And I mean, I guess you can't imagine thinking of a room, but just OK, so you can it can vacuum and it can move around and it can find its way back to the charger. Oh, and also it can replicate itself because that's that takes a lot of work.
00:44:11
Speaker
Just in the dust that it vacuums up or something. It's got a lot of extra parts that it doesn't currently have. A lot of parts that does not have right now. But it would feel way more convincing if it could as being alive. It becomes a question, how would the thing replicate itself if it needed a battery? It would have to be a pretty simple robot in order
00:44:41
Speaker
The more complicated you make the robot, the more complicated it has to be to replicate it. So it has to be a fairly simple kind of thing that can replicate itself. When I think of the first self-replicating robots, I always think about one that would just simply build another one like itself out of the same kind of parts that it already has using a formula, like a program, like a recipe that it has in its own memory stores.
00:45:09
Speaker
Yeah, I just build another one like itself and then just send it off into the world. And then it could mess around with it, make it a little different. And then they would just keep multiplying and fill up as much space as it could out there in the world. It wouldn't necessarily replicate itself from its own parts. But it certainly could download its own memory easily enough. That would be easy to do. That would start convincing me that it's alive. It's getting there.
00:45:40
Speaker
And we're starting to see what the Robopocalypse would look like. Yes. You start programming. You just start programming creatures that can replicate
Self-Replication and Origin Models
00:45:51
Speaker
themselves. And they just keep doing that until they take up all the space. That's just like a Robopocalypse of just being squished by all of the robots that are filling up all the space. The Squishpocalypse. The Squishpocalypse.
00:46:10
Speaker
All right, so it's easy enough to imagine being able to build a robot that can build another one like itself and randomly vary whatever aspects of design. That's super easy to imagine. No problem. That doesn't feel necessarily alive. The thing that made it feel alive before is that we were kind of talking about it like it was intentionally doing it. Varying itself. Right.
00:46:40
Speaker
or even had the intention somehow of replicating itself. The thing that gets at my intuition about what is it that would feel alive and not feel alive. I think one of the reasons why this is difficult is just the huge leap in complexity from something that's not alive to something that is alive. This is probably one of the things that the field of A-Life is trying to understand is how
00:47:08
Speaker
life itself originally came from non-life. How did self-replicating DNA come about in the first place? It's an immensely complicated structure. There had to be some process by which it's gradually formed. It didn't just spontaneously appear out of nowhere.
00:47:32
Speaker
I think this may be one of the fundamental understandings that a field is trying to get at is how does this go from really simple processes progressively build up to something that's more and more complicated and then you get to something like DNA and single cell organisms and just straightforward path from there to us and all kinds of complicated life.
00:47:58
Speaker
Right, these guys were talking about a distinction between metabolism first or replicator first approaches to the origins of life. So did you first have like a thing that was able to move around and like, we're not probably have to move around, but would somehow be able to metabolize or would you have to have something that would be able to replicate itself first?
00:48:22
Speaker
the first living thing or proto-life was able to sustain itself essentially, right? But couldn't reproduce. Or you do start with a thing that could reproduce itself, but didn't really have any other lifelike characteristics. So the classic, the easier one of these to understand, I think is the replicator first bottle, which is the RNA world hypothesis. So life sort of starts with just a bunch of like random little strings of RNA.
00:48:51
Speaker
They're just, you know, they're molecules. They just kind of came together through the way that chemical things kind of come together. And then they start just getting together in all different ways. And some of these ways, some of these ways are more robust than others. Right. So it's like an evolution kind of thing starts kicking off. This idea versus metabolism first and replicator first, it seems it has to be replicator first that
00:49:22
Speaker
You just kind of start rolling by replicating and maybe some successive changes. It seems difficult for me to think of metabolism first as being realistic, but maybe I just don't know this well enough to understand what that means. Right. I don't either. OK, so this this other idea, this is another term in here that I don't know, is this worth knowing? Autopoiesis. Autopoiesis. I had to look it up.
00:49:52
Speaker
I had to look it up too, I have to admit. It's not a term that I was familiar with. I feel like I've heard it somewhere before, but I don't think, I think when I previously heard it, I didn't choose to look it up. So autopoiesis just means a maintenance of organization. So I guess the idea that life is composed of self-organizing structures that are autonomous in some sort of way.
00:50:22
Speaker
That's what I get out of it. Yeah, I got the key point. Right. I got self-reproducing and self-maintaining. So both self-reproducing and self-maintaining. So there's some sort of self to it. In other words, there has to be individuals. Right. It can't just be a big gigantic goop of life. There's some sort of differentiation. It implies that it has some membrane
00:50:50
Speaker
that separates it from other stuff, right? Well, these sorts of things are interesting, I guess, when you're a sci-fi writer and you're thinking about what life is like on other planets.
Speculating Life on Other Planets
00:51:05
Speaker
And I bet you there are a bunch of A-Life researchers that think about this quite a bit. What sorts of features of life are more or less inevitable? And what sorts of features are changeable?
00:51:21
Speaker
So if you've got an alien on planet Meatzorp that evolves in some kind of way, how much can we constrain what we would know about how this evolution would take place? This is the whole thing where they talk about silicon-based versus carbon-based life forms. How would that make things look different? We have a hard time getting creative about this topic. I mean, all of our aliens always have heads.
00:51:50
Speaker
It's true. And there's more. And eyes and abs and ears and all this stuff. Something is like a bigger head or something. Oh, yeah. Or like three heads or six arms. They look like an octopus, but that's still something on Earth, right? Right, exactly. I mean, if you think about like the Boston Dynamic robots that they make that walk around, it's like that's amazing and awesome. Also, like probably the hardest fucking way you could ever
00:52:17
Speaker
build something to do any of the tasks that it's supposed to do. Right. Which you could do really more straightforwardly in a particular way. With a more purpose-built system. But if you want a creature that has, just imagine that alien on planet Meepzorp that has evolved by some process of evolution by natural selection, what are the
00:52:46
Speaker
What are the possible ways that it can get around? I mean, it can fly, or it can walk. I don't know. It can. Roll, square. Oh, yeah, roll. It could roll. There's only so many ways you can move around that engineers haven't thought about, though. Exactly. There's only so many physical possibilities. And I guess one of the things that a lot of this research is in mathematical modeling, you can imagine some constraints that environments are going to give mathematically that only
00:53:16
Speaker
There are only so many optimal ways that you can move around and so many ways that you can reproduce effectively. Right. I mean, I think a lot of it, though, has to do with if you're constraining yourself to four dimensions. Oh. Right. So if you start hanging out in like 10 dimensions, a whole different situation. Yeah, I'm having a hard time imagining that. Yeah, no, I don't have that ability at all.
00:53:46
Speaker
Yeah, so I think.
Conclusion: The Interdisciplinary Nature of A-Life
00:53:48
Speaker
Alright, so we have not solved all of the problems of life, but I think it's a really interesting field and there's a lot of cool stuff in there and just one point to wrap up with is the idea that a life is something that's kind of dispersed as a specific interdisciplinary topic and something that exists more within biology, computer science and all over the place.
00:54:16
Speaker
It's probably had a nice life cycle as with some core ideas to it. It's probably dispersing a bit. We can't make artificial life yet.