WEBVTT 1 00:00:01.199 --> 00:00:06.200 Welcome to the Sentient Code, where intelligence is engineered, autonomy 2 00:00:06.280 --> 00:00:10.439 is emerging, and a line between human and machine grows thinner. 3 00:00:10.800 --> 00:00:15.359 Each episode, we decode the algorithms, explore the robotics, and 4 00:00:15.439 --> 00:00:19.000 examine the ideas shaping the future of artificial minds. 5 00:00:23.839 --> 00:00:26.440 So I want you to imagine a tiny little organism. 6 00:00:26.879 --> 00:00:30.879 It's roughly maybe a millimeter across, just swimming through this 7 00:00:31.359 --> 00:00:34.960 shallow pool of fluid in a petri dish. Right, microscopic, yeah, 8 00:00:35.000 --> 00:00:38.920 totally microscopic. And physically it looks well, it looks exactly 9 00:00:38.960 --> 00:00:39.560 like pac Man. 10 00:00:39.640 --> 00:00:41.479 It really does. It has that exact shape. 11 00:00:41.560 --> 00:00:44.759 It's got this crescent shape with a distinct gaping mouth. 12 00:00:45.240 --> 00:00:48.399 And the craziest part is that its behavior actually matches 13 00:00:48.439 --> 00:00:51.399 the shape. So it propels itself forward, right, sweeping through 14 00:00:51.399 --> 00:00:55.399 a fluid, and it's just gathering up these loose, unattached 15 00:00:55.479 --> 00:00:56.399 cells in its. 16 00:00:56.240 --> 00:00:58.079 Path, like eating dots in the game. 17 00:00:58.200 --> 00:01:01.159 Exactly like the game. It forces the loose cells into 18 00:01:01.200 --> 00:01:04.959 this dense little pocket inside its mouth, it compresses them 19 00:01:04.959 --> 00:01:09.079 into a sphere, and then it just injects that compacted sphere. 20 00:01:08.719 --> 00:01:10.879 Out And that's where things get really weird. 21 00:01:10.879 --> 00:01:14.519 Right, because within a few days that ejected sphere grows 22 00:01:14.519 --> 00:01:17.719 its own little motor structures, wakes up and starts swimming 23 00:01:17.719 --> 00:01:21.680 around to repeat the exact same sweeping and compressing process. 24 00:01:21.840 --> 00:01:24.319 Yeah, it's a method of replication that well, it just 25 00:01:24.359 --> 00:01:27.680 doesn't cleanly fit into any of our existing biological categories, 26 00:01:27.719 --> 00:01:30.000 not at all. I mean, it isn't mitosis, it isn't 27 00:01:30.079 --> 00:01:33.560 viral hijacking where it invades a host, and it certainly isn't, 28 00:01:33.560 --> 00:01:36.319 you know, sexual reproduction. We're looking at what it's called 29 00:01:36.359 --> 00:01:37.519 kinematic self. 30 00:01:37.319 --> 00:01:40.719 Replication, kinematic meaning movement exactly. 31 00:01:40.760 --> 00:01:44.120 It's a purely geometric physical assembly of offspring. 32 00:01:44.280 --> 00:01:47.879 Okay, and here's the element that completely shatters our traditional 33 00:01:47.959 --> 00:01:52.519 understanding of biology, right because this microscopic pac man it 34 00:01:52.560 --> 00:01:56.560 isn't a machine made of silicon or like nanoplastics. 35 00:01:56.640 --> 00:01:57.319 No, not at all. 36 00:01:57.359 --> 00:02:00.200 And it's also not a naturally evolved animal. It is 37 00:02:00.200 --> 00:02:03.400 constructed entirely from amphibian skin cells. 38 00:02:03.439 --> 00:02:04.000 Frog cells. 39 00:02:04.079 --> 00:02:08.520 Yeah, specifically cells that have somehow I don't even know 40 00:02:08.520 --> 00:02:13.960 how to say it, abandoned their genetically predetermined destiny. They 41 00:02:13.960 --> 00:02:16.759 were supposed to become the smooth outer layer of a frog, 42 00:02:16.800 --> 00:02:20.000 but instead they've coordinated to function as an entirely novel, 43 00:02:20.039 --> 00:02:22.159 autonomous reproducing entity. 44 00:02:22.080 --> 00:02:25.439 Which forces us to confront this massive paradigm shifting question. 45 00:02:25.479 --> 00:02:28.599 I mean, if a biological cell can just completely disregard 46 00:02:28.680 --> 00:02:30.879 its evolutionary context to become something. 47 00:02:30.759 --> 00:02:32.039 Entirely, what is the limit? 48 00:02:32.120 --> 00:02:35.479 Right, right? What is the actual limit of its capability? 49 00:02:36.080 --> 00:02:41.280 Are cells merely executing this rigid genetic script or are 50 00:02:41.360 --> 00:02:44.960 they active agents navigating a landscape of potential forms? 51 00:02:45.120 --> 00:02:48.199 And that tension is exactly what we are dissecting today. 52 00:02:48.479 --> 00:02:52.719 We are venturing into the absolute bleeding edge of synthetic biology. 53 00:02:52.800 --> 00:02:57.120 We're looking at zenobots, their human derived successors known as anthrobots, 54 00:02:57.520 --> 00:03:01.599 and really the deeper philosophical earthquake happening underneath all this research. 55 00:03:01.719 --> 00:03:03.759 It's a fundamental challenge to the dogma. 56 00:03:03.879 --> 00:03:07.759 Totally, because from little school onward, we're taught that biological 57 00:03:07.759 --> 00:03:10.759 form is just the result of blind, bottom up emergence, 58 00:03:10.840 --> 00:03:14.840 you know, physics, chemistry, random genetic mutations just thrown into 59 00:03:14.879 --> 00:03:18.000 the crucible of natural selection over millions of years, the 60 00:03:18.039 --> 00:03:22.439 standard Darwinian model, right, But the behavior of these cellular constructs, 61 00:03:22.520 --> 00:03:26.240 it implies a radically different architecture to reality. It suggests 62 00:03:26.280 --> 00:03:30.159 that biological matter might actually be accessing or maybe discovering, 63 00:03:30.199 --> 00:03:32.520 pre existing ideal blueprints. 64 00:03:32.280 --> 00:03:34.120 And that, right there is the crux of the debate 65 00:03:34.120 --> 00:03:36.360 we're going to unpack today. We really need to look 66 00:03:36.400 --> 00:03:40.280 past the sheer engineering marvel of these biological robots, as 67 00:03:40.319 --> 00:03:43.240 cool as they are, and examine the mechanisms at play. 68 00:03:43.599 --> 00:03:45.479 So we're going to break down the physical forces of 69 00:03:45.560 --> 00:03:51.319 kinematic replication, investigate the bioelectric networks that act as cellular software. 70 00:03:50.879 --> 00:03:53.840 The software of life, which is just a wild concept, 71 00:03:54.080 --> 00:03:54.639 it really is. 72 00:03:54.960 --> 00:03:57.560 And ultimately we have to question whether the genome is 73 00:03:57.599 --> 00:04:01.560 actually a rigid architectural blueprint or if it's merely like 74 00:04:01.599 --> 00:04:05.199 an interface pointing to a much larger, unseen dimension of 75 00:04:05.319 --> 00:04:06.560 biological possibility. 76 00:04:06.639 --> 00:04:09.000 Okay, so let's start with the physical genesis of all this. 77 00:04:09.599 --> 00:04:12.960 The timeline kicks off in the early twenty twenties, researchers 78 00:04:13.000 --> 00:04:16.639 at Tufts University, the University of Vermont, and the Wiss 79 00:04:16.639 --> 00:04:20.800 Institute at Harvard managed to create the very first xenobots. 80 00:04:21.360 --> 00:04:24.120 Yeah, and the nomenclature there comes directly from the source material. 81 00:04:24.160 --> 00:04:26.839 It's an office lavist, which is the African Claude frog. 82 00:04:26.959 --> 00:04:27.319 Got it. 83 00:04:27.639 --> 00:04:31.600 So the researchers harvested pluripotent stem cells from the blastulas 84 00:04:31.600 --> 00:04:33.639 stage of these frog embryos. 85 00:04:33.319 --> 00:04:36.079 Blasterless stage meaning like super early development ry. 86 00:04:36.079 --> 00:04:40.480 Yeah, right, exactly. These are rudimentary, undifferentiated cells. They haven't 87 00:04:40.480 --> 00:04:43.319 decided what they're going to be yet. And it is 88 00:04:43.360 --> 00:04:46.560 so critical to state right up front that the genome 89 00:04:46.639 --> 00:04:48.879 of these cells was left entirely intact. 90 00:04:48.959 --> 00:04:52.519 Okay, wait, let's emphasize that no crisper cast nine, no crisper, 91 00:04:52.759 --> 00:04:58.079 no viral vectors delivering modified genes, no synthetic base pairs, no. 92 00:04:58.240 --> 00:05:04.279 That zero modification. They are genetically indistinguishable from the cells 93 00:05:04.319 --> 00:05:06.279 of a standard frog swimming in a pond. 94 00:05:06.360 --> 00:05:09.519 Right now, that's insane. So how did they change them? 95 00:05:09.639 --> 00:05:13.839 The intervention was purely physical. Researchers basically utilize these microscopic 96 00:05:13.879 --> 00:05:17.639 forceps and electrodes to manually separate the cells, strip away 97 00:05:17.639 --> 00:05:20.480 the surrounding embryonic signals, the chemical cues that tell them 98 00:05:20.480 --> 00:05:23.759 to make a tadpole and aggregate them into novel millimeter 99 00:05:23.800 --> 00:05:24.720 scale configuration. 100 00:05:24.959 --> 00:05:26.519 Is literally pushing them into a pile. 101 00:05:26.759 --> 00:05:31.680 Essentially, yes, and in that isolated, reorganized state, the cells 102 00:05:31.720 --> 00:05:34.439 didn't simply undergo apoptosis and die, which is what. 103 00:05:34.600 --> 00:05:36.399 You might expect, right, if you take cells out of 104 00:05:36.399 --> 00:05:39.240 a body, they usually just die exactly. 105 00:05:38.920 --> 00:05:41.959 And they didn't form an amorphous tumor either. They exhibited 106 00:05:42.079 --> 00:05:44.079 highly coordinated emergent properties. 107 00:05:44.120 --> 00:05:46.839 It became autonomous, yes, but I want to drill down 108 00:05:46.879 --> 00:05:50.079 into the mechanics of that autonomy because how exactly were 109 00:05:50.120 --> 00:05:52.480 they moving? I mean, without a central nervous system to 110 00:05:52.600 --> 00:05:56.360 coordinate motor functions, how does a cluster of frog skin 111 00:05:56.720 --> 00:05:58.079 cells propel itself. 112 00:05:58.160 --> 00:06:00.839 Well, it comes down to a repurposed bio logical motor 113 00:06:00.920 --> 00:06:02.199 called the cilium sili. 114 00:06:02.240 --> 00:06:02.720 Okay, yeah. 115 00:06:02.759 --> 00:06:07.000 Silia are these microscopic hair like organelles that protrude from 116 00:06:07.000 --> 00:06:09.879 the surface of many different types of cells. In a 117 00:06:09.959 --> 00:06:13.439 normal frog embryo or in your own respiratory tract for 118 00:06:13.480 --> 00:06:17.319 that matter. Slated cells form a stationary mucosal layer, so 119 00:06:17.360 --> 00:06:20.000 they just sit there, right. Their normal job is to 120 00:06:20.000 --> 00:06:22.680 beat in a synchronized rhythm to move fluid or clear 121 00:06:22.720 --> 00:06:27.759 out debris. They're essentially stationary pumps. But in the zenobot configuration, 122 00:06:28.279 --> 00:06:31.959 the cells self organized so that these cilia function like 123 00:06:32.199 --> 00:06:33.519 hundreds of tiny ores. 124 00:06:33.560 --> 00:06:36.399 But wait, without a brain, how do the ores row 125 00:06:36.439 --> 00:06:38.519 in the same direction. Because of half the cells are 126 00:06:38.519 --> 00:06:41.439 beating left and half are beating right, the bot would 127 00:06:41.480 --> 00:06:44.319 just spin in place or just tear itself apart. 128 00:06:44.879 --> 00:06:47.639 Yeah, and that is where the multi scale competency of 129 00:06:47.639 --> 00:06:49.519 the tissue really reveals itself. 130 00:06:49.680 --> 00:06:51.920 Multi scale competency, yes. 131 00:06:52.000 --> 00:06:55.000 Meaning the tissue can solve problems at a higher level 132 00:06:55.000 --> 00:06:58.600 than just the individual cell. The cells rely on localized 133 00:06:58.639 --> 00:07:02.560 biomechanical and bio electrical communication, so they form what are 134 00:07:02.600 --> 00:07:05.759 called gap junction gap junctions. Yeah, they're physical channels between 135 00:07:05.800 --> 00:07:09.040 the cell membranes. They allow ions and small molecules to 136 00:07:09.079 --> 00:07:11.199 pass directly from one cell cytoplasm to the. 137 00:07:11.199 --> 00:07:14.000 Next, like a little tunnel between the cells exactly. 138 00:07:14.480 --> 00:07:17.920 And through these junctions they synchronize their calcium signaling. It 139 00:07:17.959 --> 00:07:22.079 creates a localized consensus. The cells negotiate a global polarity 140 00:07:22.360 --> 00:07:23.399 for the entire cluster. 141 00:07:23.680 --> 00:07:26.399 Wow, So they literally talk to each other to figure 142 00:07:26.399 --> 00:07:27.920 out which ways forward they do. 143 00:07:28.199 --> 00:07:31.720 They align the beating of their cilia to achieve directional motility. 144 00:07:32.199 --> 00:07:36.079 They can swim linearly, they can navigate mazes, and impressively, 145 00:07:36.120 --> 00:07:38.079 they exhibit remarkable self healing. 146 00:07:38.199 --> 00:07:39.680 Oh right, the self healing. 147 00:07:40.040 --> 00:07:43.800 Yeah, if you last read a zenobot physically tearing it 148 00:07:43.839 --> 00:07:48.000 almost completely into for it, the cells rapidly migrate and 149 00:07:48.079 --> 00:07:52.040 knit the tissue back together, restoring their original spherical or 150 00:07:52.040 --> 00:07:53.000 sculpted shape. 151 00:07:53.120 --> 00:07:55.920 Okay, I want to pause on that localized consensus thing, 152 00:07:56.040 --> 00:07:58.959 because this is where the standard analogies we usually hear 153 00:07:59.120 --> 00:08:01.920 start to fail. I mean, you often hear this described 154 00:08:01.959 --> 00:08:05.120 as taking a pile of lego bricks meant for a castle, 155 00:08:05.519 --> 00:08:08.319 shaking the box, and watching them assemble into a car. 156 00:08:08.480 --> 00:08:09.639 Right, I've heard that one. 157 00:08:09.600 --> 00:08:12.879 But Lato's are inert. A single Lato brick doesn't negotiate 158 00:08:12.879 --> 00:08:13.800 with the brick next to it. 159 00:08:14.120 --> 00:08:17.040 No, and the Lato analogy is fundamentally flawed for exactly 160 00:08:17.079 --> 00:08:21.399 that reason. It treats biological matter as passive hardware. A 161 00:08:21.439 --> 00:08:25.319 much more accurate comparison would be a flock of birds 162 00:08:25.399 --> 00:08:26.399 or a swarm of bees. 163 00:08:26.560 --> 00:08:28.439 How interesting a swarm, right. 164 00:08:28.319 --> 00:08:31.040 Because a single bee has its own localized sensory inputs, 165 00:08:31.040 --> 00:08:34.840 its own behavioral algorithms. But when thousands of bees interact 166 00:08:34.879 --> 00:08:37.759 within a specific physical constraint, they form a hive that 167 00:08:37.840 --> 00:08:41.759 regulates its own internal temperature and collectively decides where to forage. 168 00:08:41.799 --> 00:08:44.720 So the cells are like individual bees exactly. 169 00:08:45.039 --> 00:08:48.360 The xenobod cells are active agents. When you remove the 170 00:08:48.399 --> 00:08:51.480 top down chemical gradients of the embryo, you know, the 171 00:08:51.519 --> 00:08:55.879 signals shouting form a tadpole spine. Here the cells default 172 00:08:55.919 --> 00:08:57.799 to a baseline agential state. 173 00:08:58.200 --> 00:08:59.960 They wake up and realize they're on their own. 174 00:09:00.240 --> 00:09:03.759 Yeah, they assess their new topological boundaries, and they optimize 175 00:09:03.759 --> 00:09:07.600 for survival and equilibrium in that new micro environment. 176 00:09:07.480 --> 00:09:11.240 Which perfectly sets up the twenty twenty one breakthrough. Because 177 00:09:11.240 --> 00:09:15.159 surviving and moving is one thing that's cool, but replication 178 00:09:15.679 --> 00:09:19.799 requires an entirely different level of systemic organization. So let's 179 00:09:19.799 --> 00:09:23.120 dig into the kinematic self replication because the mechanism behind this. 180 00:09:23.240 --> 00:09:25.840 Is just wild, it really is. I mean to understand 181 00:09:25.840 --> 00:09:28.840 why kinematic replication is so disruptive, we have to look 182 00:09:28.879 --> 00:09:32.799 at the deeply entrenched assumptions of molecular biology. Okay, reproduction 183 00:09:32.840 --> 00:09:35.000 as we currently define it is intrinsically tied to the 184 00:09:35.039 --> 00:09:36.480 transmission of genetic. 185 00:09:36.120 --> 00:09:38.200 Information passing on the DNA. 186 00:09:38.039 --> 00:09:42.600 Right, Mitosis copies the genome, myosis recombines it. Even a virus, 187 00:09:42.639 --> 00:09:44.639 which kind of borders on the edge of life anyway, 188 00:09:45.039 --> 00:09:48.240 operates by injecting RNA or DNA into a host to 189 00:09:48.320 --> 00:09:49.559 hijack its ribosomes. 190 00:09:49.879 --> 00:09:51.840 The code is always the star of the show. 191 00:09:52.159 --> 00:09:55.840 Always. The biological imperative is the propagation of the code. 192 00:09:56.360 --> 00:10:02.039 But kinematic self replication completely bypassesmic transmission as the driving 193 00:10:02.080 --> 00:10:02.919 physical mechanism. 194 00:10:03.039 --> 00:10:06.960 Right, It's closer to what's it called a Vonnoumann probe, Yes, 195 00:10:07.000 --> 00:10:09.759 the theoretical spacecraft right, where a spaceship lands on a 196 00:10:09.799 --> 00:10:12.519 dead planet, mines iron and silicon and just builds a 197 00:10:12.519 --> 00:10:15.679 physical clone of itself to launch to the next star system. 198 00:10:15.759 --> 00:10:19.000 Precisely, it is reproduction through the sheer mechanical assembly of 199 00:10:19.120 --> 00:10:23.320 environmental raw materials. Now, the Tufts and Vermont researchers knew 200 00:10:23.639 --> 00:10:26.960 the unmodified frog cells didn't have the biological hardware to 201 00:10:27.039 --> 00:10:30.080 reproduce sexually or even bud like yeast. 202 00:10:29.960 --> 00:10:32.399 Right, They're just skin cells exactly. 203 00:10:32.159 --> 00:10:35.440 So they turned to an in silico evolutionary algorithm to 204 00:10:35.440 --> 00:10:39.519 see if a specific physical geometry could induce a replication behavior. 205 00:10:39.639 --> 00:10:42.320 Okay, I want to get really granular on this algorithm, 206 00:10:42.360 --> 00:10:45.679 because how do you simulate the physical nuance of a 207 00:10:45.759 --> 00:10:48.720 living biological cell inside a supercomputer? 208 00:10:49.399 --> 00:10:52.440 Well, the utilized a specialized physics engine coupled with a 209 00:10:52.440 --> 00:10:57.600 morphogenetic algorithm. The supercomputer modeled the basic biomechanical properties of 210 00:10:57.639 --> 00:10:58.120 the frog. 211 00:10:58.000 --> 00:11:01.279 Stem cells, like what kind of properties their. 212 00:11:01.120 --> 00:11:05.360 Adhesion forces, their fluid dynamics and an aqueous medium, their density, 213 00:11:05.919 --> 00:11:08.000 and the contractile forces of the cilia. 214 00:11:08.120 --> 00:11:11.960 So it's basically a highly advanced physics simulator, right. 215 00:11:12.279 --> 00:11:15.639 And then the AI began this massive combinatorial search. It 216 00:11:15.720 --> 00:11:20.360 generated a random arbitrary three dimensional shape simulated, dropping it 217 00:11:20.399 --> 00:11:23.279 into a virtual Petri dish filled with thousands of loose 218 00:11:23.279 --> 00:11:27.559 simulated stem cells, and just calculated the physical interactions. 219 00:11:27.039 --> 00:11:29.559 Most of which I assume resulted in absolute failure. 220 00:11:29.600 --> 00:11:31.840 Oh almost entirely. I mean, the vast majority of shapes 221 00:11:31.879 --> 00:11:35.159 just tumbled aimlessly, or they ended up pushing the loose cells. 222 00:11:34.919 --> 00:11:37.159 Away like a bad snowplow exactly. 223 00:11:37.679 --> 00:11:41.080 But the algorithm was looking for a specific objective function, 224 00:11:41.519 --> 00:11:44.799 the aggregation of loose material, so if a specific shape 225 00:11:44.840 --> 00:11:48.080 accidentally managed to corral a few cells together, the AI 226 00:11:48.200 --> 00:11:52.279 selected It, introduced slight morphological mutations to the design, and 227 00:11:52.399 --> 00:11:55.639 ran the simulation again. Wow, it was brutally evaluating the 228 00:11:55.639 --> 00:11:58.759 fluid mechanics and the rentals number of these microscopic entities, 229 00:11:59.120 --> 00:12:02.879 optimizing purely for the mechanics of sweeping and compressing. 230 00:12:02.639 --> 00:12:06.759 And after billions of simulated generations, the algorithm converged on 231 00:12:06.799 --> 00:12:10.240 an optimal design. It did the pac Man shape, the 232 00:12:10.279 --> 00:12:11.799 sea shaped crescent. 233 00:12:11.559 --> 00:12:16.159 Yes a semitaorus. And what is truly fascinating is that 234 00:12:16.200 --> 00:12:20.559 when the biologists actually took microscopic surgical tools and manually 235 00:12:20.600 --> 00:12:25.039 sculpted the living frog cells into this AI generated sea shape, 236 00:12:25.159 --> 00:12:28.480 the physical reality perfectly matched the simulation. 237 00:12:28.720 --> 00:12:29.519 That's incredible. 238 00:12:29.600 --> 00:12:32.679 The parent bought would swim in concentric circles using the 239 00:12:32.720 --> 00:12:35.840 interior curve of the sea to catch loose stem cells. 240 00:12:36.519 --> 00:12:39.279 The physical geometry created a fluid vortex that pulled the 241 00:12:39.320 --> 00:12:40.799 loot cells into a tight cluster. 242 00:12:41.159 --> 00:12:44.279 But the replication is an infinite right, I mean, entropies 243 00:12:44.279 --> 00:12:45.240 still exists. 244 00:12:44.879 --> 00:12:48.639 All thermodynamics absolutely still apply. This is a crucial constraint. 245 00:12:48.919 --> 00:12:52.759 Kinematic replication in this specific setup suffers from generational decay. 246 00:12:52.879 --> 00:12:55.720 Generational decay meaning it gets weaker over time. 247 00:12:55.840 --> 00:12:58.919 Basically, the parent bot might gather fifty cells to form 248 00:12:58.919 --> 00:13:02.320 a first generation office, but that offspring being maybe slightly 249 00:13:02.399 --> 00:13:05.600 less structurally perfect or slightly smaller, not only possessed the 250 00:13:05.600 --> 00:13:08.840 physical surfaced area to gather thirty cells, and then the 251 00:13:08.879 --> 00:13:13.440 third generation might only gather fifteen. Eventually, the constructs drop 252 00:13:13.519 --> 00:13:17.159 below the critical mass required to maintain coordinated ciliary movement 253 00:13:17.639 --> 00:13:19.120 and the lineage terminates. 254 00:13:19.399 --> 00:13:22.399 So it's not an uncontrollable gray goose scenario where they 255 00:13:22.399 --> 00:13:23.159 take over the world. 256 00:13:23.320 --> 00:13:24.639 No, no, not at all. 257 00:13:24.679 --> 00:13:29.000 But the philosophical proof of concept, it's just staggering. You 258 00:13:29.120 --> 00:13:33.440 have an artificial intelligence that successfully explored a combinatorial space 259 00:13:33.480 --> 00:13:37.519 of physical interactions and discovered a method of biological reproduction 260 00:13:37.759 --> 00:13:40.559 that Darwinian evolutions simply never bothered to. 261 00:13:40.559 --> 00:13:44.080 Invent and never bothered to invent. Is the perfect phrasing there. 262 00:13:44.360 --> 00:13:49.240 Natural selection only optimizes for fitness within a specific ecological 263 00:13:49.360 --> 00:13:52.200 niche a frog needs to outswim a fish and catch 264 00:13:52.200 --> 00:13:55.720 a fly. It never encountered a selective pressure that required 265 00:13:55.759 --> 00:13:59.440 isolated skin cells to mechanically assemble offspring in a highly 266 00:13:59.480 --> 00:14:01.240 concentrate soup of stem cells. 267 00:14:01.279 --> 00:14:03.360 It just never came up in nature exactly. 268 00:14:03.559 --> 00:14:06.399 The capability was there latent in the physics of the cells, 269 00:14:06.480 --> 00:14:09.799 but natural evolution never accessed that specific attractor state. 270 00:14:10.080 --> 00:14:12.399 Now a skeptic might look at this and say, Okay, 271 00:14:12.840 --> 00:14:17.120 this is a fascinating parlor trick using embryonic stem cells. 272 00:14:17.720 --> 00:14:21.639 Because stem cells are by definition pluripotent. They are the 273 00:14:21.639 --> 00:14:25.240 blank slates of the biological world, programmed to be flexible. 274 00:14:25.519 --> 00:14:28.120 So it's easy to dismiss this as a quirk of 275 00:14:28.360 --> 00:14:29.519 amphibian embryology. 276 00:14:29.639 --> 00:14:31.080 Right, you could argue that, Yeah. 277 00:14:31.000 --> 00:14:32.759 But the research didn't stop there. 278 00:14:33.039 --> 00:14:35.480 No, And this is where we move past the zenobot 279 00:14:35.720 --> 00:14:39.360 and into the mid twenty twenties creation of anthrobots human cells, 280 00:14:39.559 --> 00:14:45.480 not just human cells, adult somatic human cells, specifically ciliated 281 00:14:45.480 --> 00:14:49.559 epithelial cells harvested from the human trachea, the wind pipe, 282 00:14:49.559 --> 00:14:50.080 the windpipe. 283 00:14:50.120 --> 00:14:53.799 Yeah, okay, let's unpack the biological significance of that, because 284 00:14:53.840 --> 00:14:56.480 in the traditional view of cellular development, you have the 285 00:14:56.480 --> 00:15:00.879 epigenetic landscape, right, Like Conrad Waddington's famous model on the hill, Right, 286 00:15:00.919 --> 00:15:02.840 a stem cell is like a marble at the top 287 00:15:02.879 --> 00:15:04.960 of a hill. As it rolls down the hill, it 288 00:15:05.039 --> 00:15:08.360 falls into different valleys, eventually coming to rest at the 289 00:15:08.360 --> 00:15:12.480 bottom as a fully differentiated cell, a neuron, a muscle cell, 290 00:15:12.639 --> 00:15:15.679 or a windpipe cell. And the dogma says once it's 291 00:15:15.679 --> 00:15:17.559 at the bottom of the valley. It is locked in, 292 00:15:17.679 --> 00:15:20.519 fully locked, It has deactivated the genes it doesn't need, 293 00:15:20.840 --> 00:15:23.480 and it is strictly committed to its specialized function. 294 00:15:23.759 --> 00:15:27.399 That has been the central dogma for decades. Differentiated cells 295 00:15:27.399 --> 00:15:32.320 are locked hardware, but anthrobots fundamentally violate that assumption. Oh 296 00:15:32.440 --> 00:15:36.039 so well, the researchers took these fully mature trachial cells 297 00:15:36.200 --> 00:15:39.799 that a completely rolled down Waddington's landscape and settled into 298 00:15:39.840 --> 00:15:42.399 their terminal identities and isolated them. 299 00:15:42.440 --> 00:15:46.360 And without the manual surgical sculpting that the xenobots required, 300 00:15:47.279 --> 00:15:48.919 these cells just self assembled. 301 00:15:49.080 --> 00:15:52.639 Yes, that is a massive leap in multi scale competency. 302 00:15:53.159 --> 00:15:55.960 The researchers didn't need microscopic forceps to carve out a 303 00:15:56.000 --> 00:15:59.679 C shape this time. They just manipulated the surrounding extracellular 304 00:15:59.679 --> 00:16:02.679 matre altering the viscosity of the fluid. 305 00:16:02.360 --> 00:16:04.639 Environment, and the cells just responded to that. 306 00:16:04.759 --> 00:16:10.799 They responded dynamically, They spontaneously aggregated into multicellular spheroids, and 307 00:16:10.960 --> 00:16:14.600 more incredibly, they managed to orient their cilia wait oren them. 308 00:16:14.720 --> 00:16:18.559 How So, in a wind pipe, the cilia face outward 309 00:16:18.720 --> 00:16:21.120 into the lumen to move mucus up and out of 310 00:16:21.120 --> 00:16:25.080 the lungs in the petri dish. The cells actually recognize 311 00:16:25.120 --> 00:16:28.759 their new spherical topology and ensure the cilia were facing 312 00:16:28.799 --> 00:16:30.399 outward on the surface. 313 00:16:30.000 --> 00:16:31.720 Of the sphere to act as ores. 314 00:16:31.399 --> 00:16:33.480 Again exactly enabling them to swim. 315 00:16:33.600 --> 00:16:35.399 Okay, if you're listening to this right now, let that 316 00:16:35.480 --> 00:16:39.039 sink in the cells currently lining your respiratory tract. Right 317 00:16:39.080 --> 00:16:43.759 this second possess the latent dormant capability to disconnect from 318 00:16:43.799 --> 00:16:48.519 their tissue matrix, spontaneously self organized into a motile, spherical 319 00:16:48.600 --> 00:16:51.320 robot and navigate an aqueous environment. 320 00:16:51.639 --> 00:16:54.399 They just require a shift in their contextual boundaries. That 321 00:16:54.600 --> 00:16:58.240 is wild, and their competencies extend far beyond simple motility. 322 00:16:58.600 --> 00:17:02.120 The medical implications became violently clear when researchers tested how 323 00:17:02.120 --> 00:17:04.759 these anthrobots interacted with other human tissues. 324 00:17:04.880 --> 00:17:06.319 Right the paramedic behavior. 325 00:17:06.480 --> 00:17:09.200 Yes, So they engineered a two dimensional layer of human 326 00:17:09.240 --> 00:17:12.559 neural tissue in a dish and mechanically scratched it. 327 00:17:12.599 --> 00:17:13.599 They created a wound. 328 00:17:13.680 --> 00:17:16.400 They created a physical wound that severed the neural connections, 329 00:17:16.480 --> 00:17:19.599 effectively as simulated spinal cord injury. Then they introduced the 330 00:17:19.640 --> 00:17:23.599 anthrobots into this environment, and without any external programming or 331 00:17:23.640 --> 00:17:25.400 chemical breadcrumbs from the researchers. 332 00:17:25.400 --> 00:17:26.559 Nobody told them what to do. 333 00:17:26.759 --> 00:17:30.519 Nobody told them. The anthrobots navigated directly to the site 334 00:17:30.559 --> 00:17:35.000 of the laceration. They settled specifically across the gap in 335 00:17:35.039 --> 00:17:39.160 the tissue, and then they began exerting an influence on 336 00:17:39.200 --> 00:17:39.799 the neurons. 337 00:17:39.960 --> 00:17:41.200 They started healing them. 338 00:17:41.319 --> 00:17:44.279 They didn't just passively sit there, you know. They stimulated 339 00:17:44.279 --> 00:17:47.799 the neural tissue to regenerate, coaxing the neurons to bridge 340 00:17:47.799 --> 00:17:52.599 the gap and restore the network. They functioned as autonomous biological. 341 00:17:51.920 --> 00:17:54.960 Paramedics, which is just mind blowing. But then we hit 342 00:17:54.960 --> 00:17:57.799 the twenty twenty six update, which takes this entirely out 343 00:17:57.799 --> 00:18:01.640 of the realm of mechanical paramedics and introduces the concept 344 00:18:01.839 --> 00:18:02.839 of the neurobot. 345 00:18:03.319 --> 00:18:07.319 Right, so, researchers observe that in certain configurations, these human 346 00:18:07.400 --> 00:18:11.440 constructs began developing spontaneous, primitive neural networks. Okay, hold on, 347 00:18:11.880 --> 00:18:14.119 we aren't just looking at a sphere of cells moving 348 00:18:14.119 --> 00:18:17.519 with silly anymore. We are seeing the emergence of complex, 349 00:18:17.599 --> 00:18:22.000 coordinated movements driven by an internal signaling network that highly 350 00:18:22.039 --> 00:18:24.000 resembles a rudimentary nervous system. 351 00:18:24.160 --> 00:18:28.160 Wait wait, you lost me. If the original tracheal cells 352 00:18:28.240 --> 00:18:32.519 weren't neurons and there is no brain orchestrating this, how 353 00:18:32.559 --> 00:18:36.519 does a neural network spontaneously emerge? Doesn't the DNA have 354 00:18:36.599 --> 00:18:39.440 to trigger a massive shift in cell type to create 355 00:18:39.480 --> 00:18:39.920 a neuron. 356 00:18:40.400 --> 00:18:43.240 That is the exact right quote, and it points to 357 00:18:43.279 --> 00:18:46.480 the profound plasticity of the bioelectric sensidium. 358 00:18:46.640 --> 00:18:48.440 Bioelectric what sinsidium? 359 00:18:48.720 --> 00:18:51.440 It's a network of cells that act together. The cells 360 00:18:51.519 --> 00:18:54.279 don't necessarily need to completely revert to a stem cell 361 00:18:54.319 --> 00:18:57.359 state and rewrite their entire transcriptional profile to become a 362 00:18:57.400 --> 00:19:01.480 neuron in a traditional sense. Remember, all are electrical. What 363 00:19:01.519 --> 00:19:04.480 happens in a neurobot is that the gap junctions between 364 00:19:04.599 --> 00:19:06.920 the cells, those little tunnels we talked about, they become 365 00:19:07.000 --> 00:19:10.759 highly specialized. They start allowing rapid coordinated voltage spikes that 366 00:19:10.799 --> 00:19:12.599 mimic action potentials. 367 00:19:12.160 --> 00:19:12.920 Like brain waves. 368 00:19:12.960 --> 00:19:16.519 Basically, yes, the tissue realizes that in order to coordinate 369 00:19:16.559 --> 00:19:19.440 complex movements in its new environment, it needs a faster 370 00:19:19.480 --> 00:19:23.960 communication network, so it dynamically reallocates its bioelectric resources to 371 00:19:24.000 --> 00:19:27.400 create a primitive neural architecture. It is problem solving at 372 00:19:27.440 --> 00:19:28.160 the tissue. 373 00:19:27.920 --> 00:19:30.400 Level, which brings up the most glaring mystery in all 374 00:19:30.440 --> 00:19:32.799 of this. I mean, if the genome, the DNA sequence 375 00:19:32.839 --> 00:19:36.519 remains completely unchanged throughout this entire process, which it does, right. 376 00:19:36.839 --> 00:19:41.079 If the DNA isn't explicitly dictating build a pac man 377 00:19:41.240 --> 00:19:44.720 or build a neuralbot, where is the memory of these 378 00:19:44.759 --> 00:19:48.920 shapes and behaviors actually stored. If DNA is just the 379 00:19:48.960 --> 00:19:52.359 parts list, what is functioning as the operating system? 380 00:19:52.559 --> 00:19:55.039 And that leads us to the invisible architecture of all 381 00:19:55.119 --> 00:19:57.480 morphogenesis bioelectricity. 382 00:19:57.559 --> 00:20:02.400 Okay, when you hear bioelectricitymediately think of neurology, right, action 383 00:20:02.480 --> 00:20:05.000 potential shooting down an axon, heart taste makers, the. 384 00:20:05.079 --> 00:20:07.960 Rapid transient spikes. Yes, yeah, But what Michael Levin and 385 00:20:08.039 --> 00:20:12.000 his colleagues have championed is the understanding of developmental bioelectricity. 386 00:20:12.079 --> 00:20:13.960 Developmental so slower. 387 00:20:13.680 --> 00:20:18.079 Much slower, slower, persistent electrical networks. Every single cell in 388 00:20:18.079 --> 00:20:20.559 your body has a membrane, and embedded in that membrane 389 00:20:20.599 --> 00:20:24.079 are ion channels. These are proteins that actively pump charged 390 00:20:24.079 --> 00:20:26.920 ions like potassium, sodium, and chloride in and out of 391 00:20:26.920 --> 00:20:27.279 the cell. 392 00:20:27.440 --> 00:20:30.880 And this creates a resting membrane potential, meaning the inside 393 00:20:30.880 --> 00:20:34.000 of the cell has a different electrical charge than the outside. 394 00:20:34.160 --> 00:20:37.839 Exactly, it creates a localized voltage gradient, and because these 395 00:20:37.880 --> 00:20:41.079 cells are connected via those gap junctions, they share this 396 00:20:41.200 --> 00:20:44.519 electrical state with their neighbors. They form an electrical sensidium. 397 00:20:44.599 --> 00:20:47.000 It isn't just a physical clump of cells. It is 398 00:20:47.039 --> 00:20:48.680 a unified electrical network. 399 00:20:48.839 --> 00:20:50.880 It's like a shared battery, almost a shared. 400 00:20:50.680 --> 00:20:54.400 Battery, and a shared information network. Levin's research demonstrates that 401 00:20:54.440 --> 00:20:58.839