WEBVTT - Ep111 "Might we be surrounded with undetected minds?" (with Michael Levin) 0:00:05.120 --> 0:00:09.520 What is intelligence And if we look hard, might we 0:00:09.760 --> 0:00:13.600 find it in very weird places, not just in brains, 0:00:13.720 --> 0:00:17.080 but in all kinds of structures in our universe? And 0:00:17.239 --> 0:00:20.360 how would we even recognize it? And what does any 0:00:20.440 --> 0:00:23.240 of this have to do with a dog born without 0:00:23.280 --> 0:00:27.920 front legs learning how to walk bipedally, or making new 0:00:28.000 --> 0:00:33.120 little organisms out of single cells, or how Wikipedia might 0:00:33.200 --> 0:00:37.159 be like an axilautel and why we are so blind 0:00:37.400 --> 0:00:41.680 to the vast variety of minds that might surround us. 0:00:44.440 --> 0:00:48.000 Welcome to Intercosmos with me, David Eagleman. I'm a neuroscientist 0:00:48.000 --> 0:00:50.919 and an author at Stanford, and in these episodes we 0:00:51.000 --> 0:00:55.440 sail deeply into our three pound universe to understand some 0:00:55.520 --> 0:01:13.600 of the most surprising aspects of the world around us. 0:01:15.200 --> 0:01:18.840 Today's episode is about intelligence, not in the way that 0:01:18.880 --> 0:01:22.680 I've talked about in earlier episodes, about how the structure 0:01:22.720 --> 0:01:25.520 of the human brain gives rise to intelligence and how 0:01:25.520 --> 0:01:29.839 we can measure whether AI's intelligence. Today's episode is way 0:01:29.880 --> 0:01:32.600 beyond that. Today I'm going to talk with one of 0:01:32.640 --> 0:01:37.000 my most brilliant and creative colleagues, Michael Levin, about how 0:01:37.040 --> 0:01:41.720 we might find intelligence all around us in ways that 0:01:41.760 --> 0:01:45.240 we don't typically into it. So let's start at the beginning. 0:01:45.840 --> 0:01:49.520 What is intelligence. It's a word that we usually reserve 0:01:49.920 --> 0:01:54.480 for something abstract and cerebral, something associated with problem solving 0:01:54.560 --> 0:01:58.680 and planning and passing IQ tests. We tend to picture 0:01:58.760 --> 0:02:03.680 intelligence as a property of brains, and especially big human brains. 0:02:04.160 --> 0:02:08.480 We're generally willing to grant some intelligence to dolphins and 0:02:08.560 --> 0:02:12.280 chimps and clever birds like ravens, but it's hard to 0:02:12.400 --> 0:02:15.680 know how to think about so many other things happening 0:02:15.720 --> 0:02:20.560 in the world. For example, my skin cells heal a wound. 0:02:21.080 --> 0:02:25.080 Is that intelligence or is that just biochemical cascades? A 0:02:25.160 --> 0:02:30.560 plant grows towards sunlight intelligent? A worm gets its head 0:02:30.600 --> 0:02:33.920 cut off and it regrows it. That's amazing, But we 0:02:34.000 --> 0:02:37.239 don't tend to call that cognition. But what if we've 0:02:37.280 --> 0:02:41.480 been looking at the whole notion of intelligence too narrowly. 0:02:41.800 --> 0:02:45.760 What if intelligence isn't just about neurons and genes, but 0:02:45.840 --> 0:02:49.560 it's about goals, and specifically, it's about the ability of 0:02:49.560 --> 0:02:54.920 a system to navigate towards an objective, to adapt to 0:02:54.960 --> 0:02:57.840 its circumstances, to make decisions. 0:02:57.200 --> 0:02:57.960 Along the way. 0:02:58.400 --> 0:03:02.560 That's a broader definition of intelligence. And if we apply it, 0:03:02.600 --> 0:03:07.600 suddenly intelligence doesn't just belong to creatures with brains. It 0:03:07.680 --> 0:03:12.000 becomes something that shows up in places we didn't expect. 0:03:12.120 --> 0:03:16.359 Think about really simple creatures like a tadpole. It's millions 0:03:16.400 --> 0:03:20.919 of cells collaborate and communicate and organize into an eye 0:03:21.000 --> 0:03:25.720 and a spine and a heart without anybody orchestrating the 0:03:25.760 --> 0:03:28.880 whole thing. There's no central planning, it's just a kind 0:03:28.880 --> 0:03:33.519 of emergent intelligence at work. Or think about a flatworm 0:03:33.560 --> 0:03:36.600 that can be cut into pieces and each piece regenerates 0:03:36.640 --> 0:03:40.760 a complete, properly shaped body. How does each fragment know 0:03:41.240 --> 0:03:45.320 what's missing? Where exactly is that information stored? What is 0:03:45.400 --> 0:03:49.600 guiding the process? And as we ask these questions, that 0:03:49.720 --> 0:03:52.640 leads us to ask how we can learn to talk 0:03:52.680 --> 0:03:55.400 to these systems in the language that they understand, like 0:03:55.800 --> 0:03:59.840 voltage gradients or bioelectric circuits or chemical signals. Can we 0:03:59.840 --> 0:04:04.760 start reprogramming the goals of tissues? Can we tell a 0:04:04.880 --> 0:04:08.560 clump of cells to build something new? And can we 0:04:08.720 --> 0:04:13.600 use this kind of knowledge to regenerate organs, or repair 0:04:13.920 --> 0:04:15.200 birth defects or. 0:04:15.560 --> 0:04:17.840 Create entirely new forms of life. 0:04:17.920 --> 0:04:21.200 These are the kinds of questions that today's guest has 0:04:21.240 --> 0:04:24.760 spent his career exploring and his work leads us to 0:04:24.800 --> 0:04:29.680 the conclusion that we're probably surrounded by minds, almost all 0:04:29.680 --> 0:04:33.120 of which we don't recognize. Minds are pervasive, and they're 0:04:33.160 --> 0:04:37.440 not restricted to brains, but spread across all kinds of 0:04:37.560 --> 0:04:45.040 levels of organization, from single cells to societies. My guest 0:04:45.400 --> 0:04:49.320 is Michael Levin. He's a distinguished professor of developmental and 0:04:49.360 --> 0:04:52.760 Synthetic biology at Tufts University, and I've had him on 0:04:52.800 --> 0:04:55.920 the podcast before because he's really one of my favorite 0:04:55.960 --> 0:04:59.279 thinkers in the field. He's massively creative and always pulling 0:04:59.279 --> 0:05:04.039 off amazing results that the frontier where biology meets information 0:05:04.120 --> 0:05:07.400 theory or computation or philosophy, and as you're going to see, 0:05:07.400 --> 0:05:11.960 his work always challenges our deepest intuitions about agency and 0:05:12.040 --> 0:05:16.240 memory and selfhood. So you've heard of SETI, the Search 0:05:16.320 --> 0:05:21.359 for Extraterrestrial Intelligence. Recently, Levin proposed SUTI, the search for 0:05:21.800 --> 0:05:27.000 unconventional terrestrial intelligence. As we're about to hear, his position 0:05:27.160 --> 0:05:30.880 is that right here on Earth, there are already aliens 0:05:31.040 --> 0:05:35.440 among us that stretch and sometimes break. Are typical ways 0:05:35.839 --> 0:05:40.520 of thinking about minds. So if you've ever wondered where 0:05:40.560 --> 0:05:44.120 intelligence begins, how far it reaches, or whether you might 0:05:44.200 --> 0:05:47.719 share more in common with blobs of cells than you think. 0:05:48.000 --> 0:05:56.640 This episode is for you. Here's my interview with Mike Levin. So, Mike, 0:05:56.800 --> 0:06:00.839 let's start by telling us how you define intelligence. 0:06:01.240 --> 0:06:04.520 Okay, what I use is a definition that helps us 0:06:04.640 --> 0:06:06.520 move forward in the lab. I do not claim that 0:06:06.520 --> 0:06:09.000 it's the only definition or that it captures everything there 0:06:09.080 --> 0:06:13.080 is to capture about intelligence. But I like William James's definition, 0:06:13.120 --> 0:06:15.440 which is some degree of the ability to reach the 0:06:15.480 --> 0:06:18.279 same goal by different means. So it's some level of 0:06:18.440 --> 0:06:21.280 ingenuity to get your goals met when things change. That 0:06:21.320 --> 0:06:24.640 doesn't capture play. It doesn't necessarily capture creativity, things other 0:06:24.680 --> 0:06:25.479 than problem solving. 0:06:25.520 --> 0:06:27.560 But this is what we focus on experimentally. 0:06:27.720 --> 0:06:30.360 So typically when we think about intelligence, we think about 0:06:30.440 --> 0:06:33.279 brains and nervous systems. But you think it doesn't even 0:06:33.279 --> 0:06:34.719 necessarily require that. 0:06:35.160 --> 0:06:38.400 Correct, Because if you're looking at it in this way, 0:06:38.400 --> 0:06:41.280 that it's basically a functional capacity to navigate some kind 0:06:41.279 --> 0:06:45.440 of problem space and meet specific goals under changing circumstances. 0:06:45.680 --> 0:06:48.400 There are apparently a wide range of architectures that can 0:06:48.400 --> 0:06:50.680 do this, and in order to see that what you 0:06:50.760 --> 0:06:53.960 need to do is to relax some really constraining assumptions 0:06:54.000 --> 0:06:56.280 that we often have about the problem space that we're 0:06:56.279 --> 0:06:56.680 working in. 0:06:56.839 --> 0:07:00.360 And so you often describe intelligence as scale free. So 0:07:00.440 --> 0:07:01.880 just give us a sense what you mean by that. 0:07:02.040 --> 0:07:04.920 Yeah, I mean that you know, as humans, we are 0:07:05.000 --> 0:07:07.000 because of our own evolutionary firm where we are so 0:07:07.120 --> 0:07:10.040 obsessed with three dimensional space and moving around in three 0:07:10.040 --> 0:07:12.360 dimensional space, to the point where if people see some 0:07:12.400 --> 0:07:14.520 sort of AI that isn't rolling around in some sort 0:07:14.560 --> 0:07:16.920 of robotic body, they're going to say, this is not embody, right, 0:07:16.920 --> 0:07:19.920 because they're expecting a body. People are expecting a body 0:07:19.920 --> 0:07:23.520 that moves through three dimensional space. But actually the biology, 0:07:23.560 --> 0:07:26.400 for example, has been solving problems and navigating all kinds 0:07:26.440 --> 0:07:28.560 of spaces that are hard for us to visualize. So 0:07:28.600 --> 0:07:31.600 the space of gene expression states, the space of physiological states, 0:07:31.600 --> 0:07:34.720 the space of anatomical possible outcomes, and so on, and 0:07:34.760 --> 0:07:38.000 so in order to understand how we navigate those spaces, 0:07:38.440 --> 0:07:40.200 you have to think in other scales. Some of these 0:07:40.200 --> 0:07:42.880 things happen very slowly, some of these things happen incredibly quickly. 0:07:43.400 --> 0:07:45.800 Some of these things are very small, some of them 0:07:45.800 --> 0:07:49.160 are very large, and we are just you know, focused 0:07:49.200 --> 0:07:52.400 on medium sized objects moving at medium speeds. But you know, 0:07:52.560 --> 0:07:55.080 I'm not claiming it's entirely scale free, but I'm saying 0:07:55.080 --> 0:07:59.320 that there are very deep scale invariant principles that operate 0:07:59.360 --> 0:08:01.520 at many different scales besides the ones we're used to. 0:08:01.760 --> 0:08:05.720 So how have you gone about looking for intelligence at 0:08:05.720 --> 0:08:06.440 other scales? 0:08:06.800 --> 0:08:09.640 So one thing that you that you can do once 0:08:09.640 --> 0:08:13.000 you've bought into the fact that we can't assume how 0:08:13.040 --> 0:08:15.240 intelligent anything is or what kind of intelligence it has, 0:08:15.280 --> 0:08:17.680 but you have to do experiments. Then then what it 0:08:17.720 --> 0:08:19.600 turns out you have to do is you have to 0:08:19.920 --> 0:08:22.720 posit some sort of problem space you have that the 0:08:22.760 --> 0:08:25.080 system is working in. You have to posit some sort 0:08:25.080 --> 0:08:27.200 of goal. That is, these are all hypotheses on some 0:08:27.240 --> 0:08:29.600 sort of goal that it's it's trying to reach. And 0:08:29.640 --> 0:08:32.280 then what you have to do is perturbational experiments to 0:08:32.920 --> 0:08:34.840 prevent it from reaching its goal. And then you see 0:08:34.840 --> 0:08:36.880 what how you know how how smart the system is 0:08:36.920 --> 0:08:38.640 in getting around whatever you did to it, So you. 0:08:38.640 --> 0:08:40.480 Knock it off track and then you see how it 0:08:40.520 --> 0:08:42.319 comes back, or if it comes back. 0:08:42.320 --> 0:08:45.240 Knock it off track is a good one add barriers 0:08:45.360 --> 0:08:47.319 of some sort in whatever space you're working doesn't have 0:08:47.360 --> 0:08:49.280 to be a physical barrier, but whatever space you're working in, 0:08:49.320 --> 0:08:51.680 add a barrier, or in fact manipulate. 0:08:51.200 --> 0:08:52.000 The system itself. 0:08:52.120 --> 0:08:54.240 So change the system, right, It's not all about the environment, 0:08:54.360 --> 0:08:56.120 so it can be about the system itself. And so 0:08:56.559 --> 0:08:59.160 we've done this now in many different context Here are 0:08:59.160 --> 0:09:01.280 a couple of favorites of ours. The biggest one that 0:09:01.320 --> 0:09:04.120 we do most of our work in is morphogenesis. So 0:09:04.440 --> 0:09:07.120 we all make a journey from a single cell to whatever. 0:09:07.360 --> 0:09:09.760 You know, we're going to be a human, a giraffe, plant, whatever, 0:09:09.840 --> 0:09:12.440 And that journey, as it turns out, as a matter 0:09:12.480 --> 0:09:14.760 of experimental fact, it turns out that journey is not 0:09:15.000 --> 0:09:17.640 a kind of open loop. What you know, the way 0:09:17.640 --> 0:09:20.400 people model the emergence and complexity, lots of simple things 0:09:20.440 --> 0:09:23.199 happening again and again, and ultimately some sort of complex 0:09:23.240 --> 0:09:23.920 event happens. 0:09:24.080 --> 0:09:27.160 That isn't how it works. It's very contact sensitive. 0:09:27.320 --> 0:09:30.800 If you try to prevent let's say, embryos or regenerating limbs, 0:09:31.120 --> 0:09:33.840 or or you know, in any you know, metamorphosis, any 0:09:33.840 --> 0:09:36.240 of these processes, you try to prevent them from reaching 0:09:36.280 --> 0:09:39.240 their goal. They often have extremely ingenious ways to get 0:09:39.240 --> 0:09:42.520 there anyway, okay, And you can quantify this, and you 0:09:42.559 --> 0:09:46.120 can say, what are kinds of perturbations that it's able 0:09:46.160 --> 0:09:48.760 to deal with? And does it have delayed gratification? Does 0:09:48.800 --> 0:09:51.680 it have planning, does it have a representation of the 0:09:52.200 --> 0:09:55.000 of counterfactual states? Does it have learning and memory? Does 0:09:55.040 --> 0:09:57.520 it store you know, a map of its environment? You 0:09:57.559 --> 0:09:59.400 can you can test all of these things empirically. 0:09:59.559 --> 0:10:02.720 So I mean in terms of let's say an embryo developing, 0:10:03.040 --> 0:10:06.559 what we think traditionally in textbooks is that the genetics 0:10:06.559 --> 0:10:09.720 somehow gives a blueprint and the whole thing just donepacks. 0:10:10.040 --> 0:10:13.280 But you're asking is how is the system intelligent if 0:10:13.280 --> 0:10:15.640 we knock it off track or put barriers in the way, 0:10:15.840 --> 0:10:19.000 how does it figure out how to come together correctly? So, 0:10:19.200 --> 0:10:21.920 what's the specific example of something you've done here? Here 0:10:22.080 --> 0:10:24.400 there are some of my favorites. These first two are 0:10:24.440 --> 0:10:26.640 not my work. These this is like classic, classic work 0:10:26.640 --> 0:10:27.120 in the field. 0:10:27.480 --> 0:10:31.440 So if you take normally, imagine cutting a cross section 0:10:31.559 --> 0:10:34.600 through the kidney tubule of a newt. Normally what you'd 0:10:34.640 --> 0:10:36.400 find is like eight to ten cells and they work 0:10:36.440 --> 0:10:40.760 together to build to build this tube like structure. So 0:10:41.000 --> 0:10:42.840 what you can do is you can make polyploid neutes 0:10:42.880 --> 0:10:45.760 that have multiple copies of their chromosomes, which means their 0:10:45.760 --> 0:10:47.920 cells have to get bigger. Those newts are still the 0:10:47.960 --> 0:10:50.840 same correct size. So that's the first interesting thing. Wow, 0:10:50.840 --> 0:10:53.120 well the cells get bigger, the thing scales down. How 0:10:53.120 --> 0:10:56.080 does it do it by using fewer but bigger cells 0:10:56.120 --> 0:10:57.480 to make the exact same structure. 0:10:57.679 --> 0:11:00.000 So that's an adjustment, right, never mind the environment. 0:11:00.080 --> 0:11:01.920 Your own parts are changing, and this thing is figuring 0:11:01.920 --> 0:11:03.480 out how to get to the exact same goal, the 0:11:03.480 --> 0:11:06.840 same neud same shape, same size, fewer of these bigger cells. 0:11:06.960 --> 0:11:08.360 So let me ask you a quick question. 0:11:08.480 --> 0:11:12.480 Is this analogous to the fact that a mouse's heart 0:11:12.520 --> 0:11:15.720 and an elephant's heart are doing the same thing, but 0:11:15.760 --> 0:11:19.160 they're made of a completely different number of cells. It's 0:11:19.200 --> 0:11:21.560 a massive heart in an elephant, very tiny mouse, but 0:11:21.760 --> 0:11:23.760 it's doing the same function. 0:11:24.160 --> 0:11:27.240 It's similar, but there's only but there's one major difference 0:11:27.280 --> 0:11:29.480 both in a mouse and in an elephant. What people 0:11:29.520 --> 0:11:31.960 will say is, well, both of those have had long, 0:11:32.280 --> 0:11:34.320 long history of being what they are, and so this 0:11:34.400 --> 0:11:35.040 is just mechanical. 0:11:35.120 --> 0:11:36.080 It just does what it does. 0:11:36.760 --> 0:11:39.480 My example is different because you've done something to this 0:11:39.600 --> 0:11:40.480 new that it does. 0:11:40.360 --> 0:11:41.040 Not normally do. 0:11:41.240 --> 0:11:45.400 You've given it a completely novel circumstance, and then it 0:11:45.440 --> 0:11:48.040 adjusts in a new way. And the craziest thing happens 0:11:48.200 --> 0:11:51.440 when you make the cells really gigantic. Okay, these are 0:11:51.440 --> 0:11:53.680 like I think six or eight and the newts so 0:11:53.760 --> 0:11:56.920 massive polyplaty. What happens is the cells are so big 0:11:57.040 --> 0:11:58.560 there's no room for more than one cell. 0:11:58.679 --> 0:12:01.440 One cell will wrap around und itself and give you 0:12:01.520 --> 0:12:03.120 the lumen of the tubule in the middle. 0:12:03.280 --> 0:12:06.280 Now, now this is crazy because because that is that's 0:12:06.320 --> 0:12:08.960 a different molecular mechanism that side of skeletalle bending before 0:12:09.000 --> 0:12:12.000 it was sell to sell communication and tubulo genesis. So 0:12:12.160 --> 0:12:13.880 think about what this means. If you're a nude coming 0:12:13.880 --> 0:12:16.360 into the world. Never mind the environment. You don't really 0:12:16.360 --> 0:12:17.840 know what your environment's going to be. You don't know 0:12:17.840 --> 0:12:19.800 how many copies of your chromosomes you're going to have, 0:12:20.160 --> 0:12:22.080 you don't know how big your cells are going to be. 0:12:22.160 --> 0:12:25.200 You don't know which of your many genetic affordances you 0:12:25.240 --> 0:12:27.800 can use. Right, you have different molecular mechanisms you can use. 0:12:28.040 --> 0:12:29.320 You have to figure out what to do in a 0:12:29.320 --> 0:12:31.920 novel circumstance and still get the job done. I mean 0:12:32.000 --> 0:12:34.679 this sounds like every IQ test you've ever heard of 0:12:34.720 --> 0:12:36.640 when people show you, here's a little box, some tax 0:12:36.679 --> 0:12:38.280 and a candle, and I want you to, you know, 0:12:38.360 --> 0:12:41.880 solve this this particular problem. Right, Yeah, you have genetic affordances, 0:12:42.200 --> 0:12:45.840 and then that morphogenetic process is not just doing the 0:12:45.840 --> 0:12:47.000 same thing every single time. 0:12:47.040 --> 0:12:49.000 You have to solve these problems. That's one of my 0:12:49.000 --> 0:12:49.840 favorite examples. 0:12:50.000 --> 0:12:53.440 Another one that we discovered is tadpoles become frogs, and 0:12:53.480 --> 0:12:54.960 in order to do that, they have to rearrange their 0:12:55.000 --> 0:12:57.800 face because their face looks actually quite different from and 0:12:58.120 --> 0:12:59.679 you know, so the eyes, the jaws of reading has 0:12:59.679 --> 0:13:00.440 to has to move. 0:13:00.520 --> 0:13:02.880 What people thought was that this is a hardwired process. 0:13:02.920 --> 0:13:05.559 Basically, somehow the genetics just tells every organ how far 0:13:05.600 --> 0:13:07.800 to moving, what direction, and then you get from a 0:13:07.840 --> 0:13:09.240 normal tadpole to a normal frog. 0:13:09.440 --> 0:13:10.439 So we decided to test that. 0:13:10.520 --> 0:13:12.400 Because you can't assume these things, you have to test 0:13:12.679 --> 0:13:16.120 and so what we created was something called Picasso tadpoles, 0:13:16.400 --> 0:13:20.720 so basically scrambled all the initial positions, so the mouth 0:13:20.800 --> 0:13:22.320 is off to the side, the eyes on the back 0:13:22.360 --> 0:13:25.760 of the head like everything is completely scramble. And what 0:13:25.800 --> 0:13:28.480 we find is that they make pretty normal frogs because 0:13:28.520 --> 0:13:31.120 all of these things will move in novel paths to 0:13:31.400 --> 0:13:33.600 reach the correct goal, and then they stop. Actually, sometimes 0:13:33.640 --> 0:13:35.040 they go too far and they have to double back 0:13:35.040 --> 0:13:36.920 and stop. This is another example. You start them off 0:13:36.920 --> 0:13:38.760 in the wrong position, they don't just blindly go the 0:13:38.800 --> 0:13:41.720 same distance. They actually go until they meet their goal. 0:13:41.760 --> 0:13:42.640 And you know, it's a goal. 0:13:42.840 --> 0:13:44.960 And when I say goal, I don't mean it's a human. 0:13:44.800 --> 0:13:46.520 Level, like I know what my goal is. That's a 0:13:46.559 --> 0:13:48.880 kind of metacognition that I'm not claiming here. I'm saying 0:13:48.960 --> 0:13:51.640 it's in the cybernetic sets like your thermostat has goals. 0:13:51.800 --> 0:13:54.000 It's a set point, and now how clever are you 0:13:54.080 --> 0:13:56.520 to be able to reach that set point when things change? 0:13:58.120 --> 0:14:00.920 As an interesting analogy, what's going on at the level 0:14:00.960 --> 0:14:04.000 of brain plasticity. We tend to think that, let's say, 0:14:04.040 --> 0:14:08.880 a dog's brain is pre wired to drive a dog's body. 0:14:09.240 --> 0:14:11.800 But one of the examples that I talked about in 0:14:11.800 --> 0:14:14.880 my book Live Wired was this dog who was born 0:14:14.960 --> 0:14:18.480 without front legs, and so she just walks bipedally and 0:14:18.520 --> 0:14:20.400 she moves all around and. 0:14:20.400 --> 0:14:21.240 Gets by that way. 0:14:21.280 --> 0:14:25.760 Why, because she needed to get to her her dog 0:14:25.800 --> 0:14:28.360 bowl and her water and other dogs and so on, 0:14:28.600 --> 0:14:30.320 and so she just figured out. It turns out it's 0:14:30.360 --> 0:14:32.840 not that hard for a dog to walk on back legs. 0:14:32.840 --> 0:14:35.160 And the question is could all dogs walk on their 0:14:35.200 --> 0:14:39.000 back legs? Presumably, but they don't have the proper motivation 0:14:39.920 --> 0:14:43.440 to do so. But the point is that the dog's 0:14:43.520 --> 0:14:44.760 body is very flexible. 0:14:44.760 --> 0:14:46.400 It meets the goals of the world. 0:14:46.720 --> 0:14:51.080 Another analogy is the world's best archer, as in he's 0:14:51.120 --> 0:14:55.240 got the world record for the longest accurate shot in archery. 0:14:55.680 --> 0:14:57.920 Is a guy named Matt Stutsman who happens to have 0:14:58.000 --> 0:15:01.440 no arms, and he got interest in archery and figured 0:15:01.480 --> 0:15:05.160 out how to pull the bow with his legs, and 0:15:05.240 --> 0:15:07.840 so he shoots with his legs and became a great 0:15:07.920 --> 0:15:08.600 archer that way. 0:15:09.000 --> 0:15:12.720 Amazing, amazing. Yeah, yeah, the plasticity is incredible. 0:15:12.800 --> 0:15:15.640 And you know, the earlier the earliest example that I 0:15:15.680 --> 0:15:17.520 know of the of this hind leg thing is is 0:15:17.720 --> 0:15:20.080 Slipper's Goat, which was this I think it was in 0:15:20.120 --> 0:15:23.560 the forties. This guy Slipper, a published study of a 0:15:23.600 --> 0:15:26.120 goat who, again born without four legs, learned to walk 0:15:26.160 --> 0:15:28.720 on its hind legs. When they dissected the goat, they 0:15:28.720 --> 0:15:31.440 found out that a lot of the adjustments that you 0:15:31.480 --> 0:15:33.840 need for bipedal locomotion, right, So things about the hips, 0:15:34.160 --> 0:15:36.480 you know, the spine, all kind of stuff, we're all 0:15:36.520 --> 0:15:38.680 there right as opposed to what you normally think of 0:15:38.720 --> 0:15:40.640 for the evolution of modern humans as you know how 0:15:40.880 --> 0:15:42.840 I mean, many hundreds of thousands of years you need 0:15:42.880 --> 0:15:45.480 for that. And this is what's really interesting about this 0:15:45.520 --> 0:15:47.920 plasticity is that you can project it into other spaces. 0:15:47.960 --> 0:15:50.600 So so, as you pointed out, you know, can a 0:15:50.640 --> 0:15:52.640 dog brain run an upright body? 0:15:52.720 --> 0:15:52.920 Right? 0:15:53.240 --> 0:15:53.480 Now? 0:15:53.800 --> 0:15:55.520 Look at individual cells? 0:15:55.520 --> 0:15:58.600 Can the same genome run a completely different anatomy and 0:15:58.640 --> 0:16:01.200 set of behaviors? And this is what we've I mean, 0:16:01.240 --> 0:16:03.920 other people have shown other examples of this, but for example, 0:16:03.920 --> 0:16:07.680 in our lab, xenobots anthwrobots, right, these living constructs that 0:16:07.760 --> 0:16:10.440 have a completely different body than what they normally do. 0:16:10.440 --> 0:16:13.720 They have a different behavioral repertoire, no genetic change, same 0:16:13.840 --> 0:16:17.000 gene regulatory networks are running a completely different body. 0:16:17.080 --> 0:16:20.280 For the listenership, can you define anthrobots and xenobots which 0:16:20.280 --> 0:16:20.760 you've built? 0:16:20.920 --> 0:16:22.520 Sure, let's start with the zenobots. 0:16:22.520 --> 0:16:25.800 So in the cases of zennabots, what our team did, 0:16:25.800 --> 0:16:28.160 and this is in collaboration with Josh bond Guard's lab 0:16:28.160 --> 0:16:30.320 at University the e Vermont and this is Doug Blackiston 0:16:30.360 --> 0:16:32.560 in my group and Sam Kregman did a lot of 0:16:32.560 --> 0:16:35.040 the computational work for it. What happens is that when 0:16:35.040 --> 0:16:37.640 you liberate some epithelial cells from an early frog embryo, 0:16:37.840 --> 0:16:40.160 normally what they do is they form this like two 0:16:40.240 --> 0:16:43.920 dimensional outer covering of an embryo and the outer skin layer, 0:16:43.960 --> 0:16:46.680 and they do that because they're induced to do that by. 0:16:46.560 --> 0:16:47.280 The other cells. 0:16:47.360 --> 0:16:49.000 Well, if you get them away from the other cells, 0:16:49.000 --> 0:16:50.840 you sort of liberate, then find out what they really 0:16:50.840 --> 0:16:52.600 want to do on their own and what they do. 0:16:52.920 --> 0:16:53.840 They could do many things. 0:16:53.880 --> 0:16:55.560 They could crawl away from each other, they could die, 0:16:55.560 --> 0:16:58.280 they could make a flat layer like cell culture. What 0:16:58.320 --> 0:17:00.320 to actually do is they form this little ball with 0:17:00.720 --> 0:17:02.640 cilia that are on the outside. He's a little moving 0:17:02.680 --> 0:17:05.640 hairs and they organize them so that the thing can 0:17:05.680 --> 0:17:07.160 swim and it starts swimming around. 0:17:07.160 --> 0:17:08.800 It has all sorts of interesting behaviors. 0:17:08.840 --> 0:17:10.159 A couple of years ago we show that they do 0:17:10.200 --> 0:17:12.959 kinematic cell for replication, which is that if you sprinkle 0:17:12.960 --> 0:17:14.760 a bunch of loose skin cells in their environment, they 0:17:14.800 --> 0:17:17.160 will collect them into little balls and guess what, those 0:17:17.200 --> 0:17:19.760 become the next generation of zenobots. Right, So they can 0:17:19.760 --> 0:17:22.440 do this weird kinematic replication that, as far as we know, 0:17:22.520 --> 0:17:25.880 no other creature does. They express hundreds of genes differently 0:17:25.960 --> 0:17:29.320 than they do within the embryo. No genetic change. By 0:17:29.359 --> 0:17:31.320 the way, this is we're not adding anything. There are 0:17:31.320 --> 0:17:34.320 no scaffolds, no, no synthetic circuits. But they use their 0:17:34.960 --> 0:17:38.960 transcriptional affordances differently. They turn to hundreds of new genes 0:17:39.359 --> 0:17:42.119 and among other things, it turns out they're sensitive to 0:17:42.320 --> 0:17:43.440 acoustic vibrations. 0:17:43.560 --> 0:17:45.400 That's the latest thing that just came out a month ago. 0:17:45.440 --> 0:17:47.280 Is that we get because we found they were turning 0:17:47.280 --> 0:17:49.000 out a bunch of genes related to hearing. 0:17:48.760 --> 0:17:50.520 And we said, is it possible at these things we hear? 0:17:50.840 --> 0:17:53.800 And so Vipofpie my group put a speaker under them 0:17:53.880 --> 0:17:55.880 and showed that, yeah, there's actually sounds you can send 0:17:55.880 --> 0:17:57.399 them that they will respond to. 0:17:57.640 --> 0:17:58.600 So that's zenobots. 0:17:58.640 --> 0:18:02.040 Anthrobots are a similar story because when we first did it, 0:18:02.080 --> 0:18:05.280 some people said, well, you know, they're embryonic cells and 0:18:05.400 --> 0:18:07.679 amphibia are plastic. Maybe that's why this is like a 0:18:07.680 --> 0:18:10.280 frog embryology thing. You know, this is specific to and centebat. 0:18:10.440 --> 0:18:12.160 So I said, okay, what's the furthest you can get 0:18:12.160 --> 0:18:13.080 from an embryonic frog. 0:18:13.080 --> 0:18:14.160 Well, I'll be an adult human. 0:18:14.520 --> 0:18:17.879 And so we went and we took trachael epithelial cells 0:18:18.119 --> 0:18:21.040 from adult human patients and we showed and this is 0:18:21.240 --> 0:18:24.040 Gizem Gumushke's work, PhD student in my group. 0:18:23.800 --> 0:18:26.440 Who developed a protocol whereby. 0:18:26.400 --> 0:18:29.600 Again simply by taking the cells out of their normal context, 0:18:29.640 --> 0:18:33.040 you get to release their the various possible outcomes that 0:18:33.080 --> 0:18:35.320 they can do, and they make anthrobots. It's a little 0:18:36.119 --> 0:18:39.360 round little thing that zips around. It has a couple 0:18:39.440 --> 0:18:42.880 of interesting properties. First of all, it can heal neural wounds. 0:18:43.080 --> 0:18:45.000 So if you played a dish of human neurons and 0:18:45.000 --> 0:18:46.840 you put a big scratch through it with a scalpel, 0:18:47.200 --> 0:18:49.600 they will. When they find the scratch, they settle down, 0:18:49.600 --> 0:18:51.359 a bunch of them. We call it a superbot cluster. 0:18:51.440 --> 0:18:54.119 They settle down and within about four days, if you 0:18:54.200 --> 0:18:55.560 lift them up, you see that what they did. For 0:18:55.640 --> 0:18:57.800 what they did meanwhile is they healed across the They 0:18:57.880 --> 0:19:00.200 healed across the gap. Okay, who would have thought that 0:19:00.240 --> 0:19:02.960 your trachular ethelial cells that sit there quietly dealing with 0:19:03.000 --> 0:19:06.199 you know, mucus and and and the air particles have 0:19:06.280 --> 0:19:09.520 the ability to to to to heal neurons. And these 0:19:09.520 --> 0:19:13.840 guys express about nine thousand genes differently than right, so 0:19:13.960 --> 0:19:17.000 what almost half the genome they express differently than than 0:19:17.000 --> 0:19:20.000 they do in the body. They're, by the way, younger 0:19:20.040 --> 0:19:22.240 than the patient, than the than the cells that they 0:19:22.240 --> 0:19:24.679 came from. So so actually that process of becoming an 0:19:24.680 --> 0:19:27.000 answer what actually rolls back the epigenetic clock. 0:19:27.280 --> 0:19:29.000 So so they're they're they're a bit younger. 0:19:29.000 --> 0:19:31.720