WEBVTT - Ep55 "Could a brain plugin instantly teach you to fly a helicopter?" 0:00:03.240 --> 0:00:07.680 Could you learn to fly a helicopter not by practicing, 0:00:08.080 --> 0:00:12.120 but instead by uploading the information directly into your brain? 0:00:12.880 --> 0:00:16.200 What would society do if kids no longer had to 0:00:16.200 --> 0:00:18.560 go to school? And what does any of this have 0:00:18.640 --> 0:00:23.200 to do with suntan booths or nano robots or torking 0:00:23.440 --> 0:00:27.680 over a presidential address or what a cowboy on a 0:00:27.760 --> 0:00:35.000 hill is simply not able to see. Welcome to Inner 0:00:35.120 --> 0:00:38.800 Cosmos with me David Eagleman. I'm a neuroscientist and author 0:00:38.840 --> 0:00:42.840 at Stanford and in these episodes we sail deeply into 0:00:42.880 --> 0:00:47.239 our three pound universe to understand why and how our 0:00:47.320 --> 0:00:59.400 lives look the way they do. Today's episode is about 0:00:59.440 --> 0:01:05.280 the potability of really coming to understand the tangled forest 0:01:05.360 --> 0:01:08.520 of eighty six billion neurons in your head and the 0:01:08.640 --> 0:01:11.960 trillions of connections between them. And if we could do that, 0:01:12.480 --> 0:01:17.200 could we upload information directly into your brain? Could we 0:01:17.360 --> 0:01:21.760 speed up education this way? Now? At the moment, this 0:01:21.880 --> 0:01:25.720 is all pure fantasy because we simply don't have the 0:01:25.800 --> 0:01:29.160 technology to allow us to do that. But the question 0:01:29.200 --> 0:01:32.880 we're going to ask today is whether this is theoretically 0:01:33.000 --> 0:01:36.760 possible and something we can look forward to around the 0:01:36.959 --> 0:01:40.640 corner of the next century, and what are the caveats, 0:01:40.640 --> 0:01:44.800 the things to watch out for, and the unexpected complexities here. 0:01:45.160 --> 0:01:49.080 So let's get started some hundreds of years ago and 0:01:49.240 --> 0:01:53.280 still in many impoverished places in the world, children of 0:01:53.320 --> 0:01:56.960 the species Homo sapiens reproduce by the time they are 0:01:57.080 --> 0:02:01.360 young teens. But this situation is it's totally different in 0:02:01.440 --> 0:02:05.560 modern times and modern societies. Now, young people go to 0:02:05.840 --> 0:02:10.040 school for their first eighteen years or twenty one years, 0:02:10.480 --> 0:02:14.080 and increasingly twenty five or twenty six years for an 0:02:14.120 --> 0:02:17.880 advanced degree, and in fields like medicine, they take another 0:02:18.080 --> 0:02:22.040 several years of internship and residency. And in a field 0:02:22.120 --> 0:02:26.320 like neuroscience research people do a postdoctoral fellowship and then 0:02:26.320 --> 0:02:29.919 they hope to become an assistant professor, and then an 0:02:29.919 --> 0:02:33.760 associate professor and finally a full professor. And most people 0:02:34.080 --> 0:02:37.040 are in their forties by the time they get there. 0:02:37.560 --> 0:02:40.880 So what accounts for this recent historical change. Why do 0:02:40.919 --> 0:02:43.840 we do so much schooling for so much of our 0:02:43.960 --> 0:02:48.359 lives now? Well, it's because we are a runaway species. 0:02:48.400 --> 0:02:50.919 We've gone off in a totally different direction than all 0:02:50.960 --> 0:02:55.840 our animal cousins, and we have made thousands of important 0:02:56.200 --> 0:03:00.880 discoveries about our world and produced so much art invarious forms, 0:03:01.160 --> 0:03:04.800 And as a result, there's so much to learn, and 0:03:04.840 --> 0:03:09.760 so we need to spend decades in institutions of learning, 0:03:09.840 --> 0:03:13.480 not to mention, reading books and listening to podcasts to 0:03:13.720 --> 0:03:18.920 understand what millions of humans have devoted their lives to 0:03:19.160 --> 0:03:23.280 figuring out. But what if there were a way that 0:03:23.360 --> 0:03:24.920 we didn't have to do that? What if there were 0:03:24.960 --> 0:03:28.600 a way to simply upload the information, in other words, 0:03:28.680 --> 0:03:33.320 to put the information directly into your brain. So let's 0:03:33.320 --> 0:03:38.160 harken back to this great scene in The Matrix where 0:03:38.280 --> 0:03:42.240 Neo and Trinity are being hotly pursued by the antagonist, 0:03:42.320 --> 0:03:45.560 agent Smith, and our two heroes end up on top 0:03:45.600 --> 0:03:49.240 of a building, and there they spy a helicopter parked 0:03:49.280 --> 0:03:52.720 on the roof, and Neo asks Trinity do you know 0:03:52.760 --> 0:03:56.960 how to fly that? And she replies not yet, And 0:03:57.000 --> 0:04:00.480 she flips open her phone and she calls Tank, the operator, 0:04:00.560 --> 0:04:03.680 and she says, I need a pilot program for a 0:04:03.720 --> 0:04:07.280 B two twelve helicopter, And we see the operator rotate 0:04:07.320 --> 0:04:09.840 his chair in front of his bank of computers and 0:04:09.880 --> 0:04:12.440 he quickly types out a bunch of commands, and she 0:04:12.600 --> 0:04:17.000 closes her eyes and one second later she turns confidently 0:04:17.080 --> 0:04:21.320 to Neo and says, let's go. So what happened is 0:04:21.360 --> 0:04:26.480 that Tank the operator had taken the expertise, the complicated 0:04:26.640 --> 0:04:30.159 know how of flying a B two twelve helicopter and 0:04:30.360 --> 0:04:34.160 just uploaded it to her brain. So the question we're 0:04:34.200 --> 0:04:37.719 going to ask today is is that theoretically possible from 0:04:37.839 --> 0:04:42.360 a neuroscience perspective, and what will make that straightforward? And 0:04:42.400 --> 0:04:46.679 what will make that not straightforward to accomplish someday. Now, 0:04:46.880 --> 0:04:50.000 in some ways, the whole idea sounds crazy because it 0:04:50.040 --> 0:04:52.800 seems like we always have to earn things if we 0:04:52.880 --> 0:04:55.719 want changes to our brains or body. You can't just 0:04:55.760 --> 0:04:58.839 get something for free. But of course, people for decades 0:04:58.880 --> 0:05:02.120 have been climbing in into suntan booths instead of spending 0:05:02.200 --> 0:05:06.279 days outside, and people get botox, which binds to receptors 0:05:06.360 --> 0:05:09.680 the ends of peripheral nerves and changes the wrinkliness of 0:05:09.720 --> 0:05:13.320 your face. And people are increasingly doing things to not 0:05:13.400 --> 0:05:15.640 have to go to the gym but instead to have 0:05:16.040 --> 0:05:20.919 your abdominal muscles built for you with electrical stimulation. You 0:05:21.040 --> 0:05:24.039 just lie on the table and your muscles contract over 0:05:24.080 --> 0:05:26.480 and over and The idea is that your muscles can 0:05:26.520 --> 0:05:30.279 grow stronger and look better without you having to do 0:05:30.320 --> 0:05:33.160 a single sit up. You just lie there. So what 0:05:33.160 --> 0:05:36.720 would be the equivalent in the realm of education? Can 0:05:36.760 --> 0:05:39.680 we imagine a time when you don't have to bury 0:05:39.720 --> 0:05:42.800 yourself in a book to master some domain, where you 0:05:42.839 --> 0:05:45.400 don't have to spend hundreds of hours sitting in a 0:05:45.440 --> 0:05:49.080 flight simulator, but instead you hook something up to your 0:05:49.120 --> 0:05:52.600 brain and then it is as though you already knew 0:05:53.000 --> 0:05:57.120 quantum mechanics or electrical engineering, or Persian history, or how 0:05:57.160 --> 0:06:00.200 to serve for hang glide or repair that model of 0:06:00.240 --> 0:06:05.479 dishwasher or whatever. Now, how would you push information to 0:06:05.640 --> 0:06:10.120 the brain? We currently do this by sitting down dozens 0:06:10.120 --> 0:06:13.200 of children in front of someone who already has the 0:06:13.200 --> 0:06:17.919 information in their brain, and that person uses words or pictures, 0:06:18.040 --> 0:06:21.320 and the students attend to those stimuli and try to 0:06:21.480 --> 0:06:26.239 translate those words or pictures into changes in their own 0:06:26.400 --> 0:06:30.520 private jungle of billions of neurons. They try to convert 0:06:30.600 --> 0:06:34.040 what they're hearing or seeing into storage in their own 0:06:34.080 --> 0:06:36.760 internal model in a way that makes sense to them. 0:06:37.240 --> 0:06:41.479 What learning means is that you very finely change the 0:06:41.600 --> 0:06:45.719 networks in your head. That's it. That's what we pay 0:06:45.760 --> 0:06:48.159 lots of tuition for and go off to college for 0:06:48.560 --> 0:06:51.799 to get someone who already has information in their network 0:06:52.240 --> 0:06:56.240 to translate it through the low bandwidth channel of language 0:06:56.560 --> 0:06:59.720 over to your network. So, just to be clear on this, 0:07:00.240 --> 0:07:04.240 before you know some factor concept, your network is configured 0:07:04.279 --> 0:07:07.279 in some way, and then I tell you, oh, that 0:07:07.560 --> 0:07:11.480 dog's name is Nebula, and then you encode that information. 0:07:12.040 --> 0:07:14.200 This connection in your brain gets strengthened and this one 0:07:14.240 --> 0:07:17.520 gets weakened, and this synapse unplugs, it replugs over there, 0:07:17.600 --> 0:07:20.280 and this happens over millions of synapses, and then you 0:07:20.960 --> 0:07:24.240 know something that you did not know before. And for 0:07:24.800 --> 0:07:28.800 deeper knowledge, like flying a B two twelve helicopter, this 0:07:28.880 --> 0:07:32.240 requires not just the memory of a fact, but of 0:07:32.280 --> 0:07:36.440 a procedure. And so those changes happen in different brain 0:07:36.480 --> 0:07:39.920 areas and they're more widespread. But what is required in 0:07:40.000 --> 0:07:44.200 all these forms of learning are simply changes in the 0:07:44.240 --> 0:07:48.800 patterns of your network, presumably just the synaptic connections, but 0:07:48.880 --> 0:07:53.280 maybe other details as well, like which neurotransmitter receptors are 0:07:53.280 --> 0:07:56.240 being expressed on the membranes and whatever. But that's it, 0:07:56.760 --> 0:08:01.320 that's what it means. To learn something. So is there 0:08:01.400 --> 0:08:06.000 any way to implement those changes besides the old fashioned 0:08:06.040 --> 0:08:09.480 way of sitting for a semester in a classroom or 0:08:09.600 --> 0:08:14.200 spending hours in the helicopter flight simulator. Well, there's been 0:08:14.200 --> 0:08:18.120 a lot of excitement about brain machine interfaces, such as 0:08:18.200 --> 0:08:22.960 the brain electrodes that are implanted robotically by the company Neurallink. 0:08:23.200 --> 0:08:25.880 So I'll just take a quick moment to clarify the 0:08:26.000 --> 0:08:30.360 landscape of electrodes in the brain. Even though neuralink hit 0:08:30.440 --> 0:08:32.800 the news recently. The first thing to note is that 0:08:32.880 --> 0:08:37.600 these brain machine interfaces have been around for many decades 0:08:37.640 --> 0:08:41.960 since people started inserting electrodes. These are just thin metal 0:08:41.960 --> 0:08:44.920 wires into the brain. The idea is that you just 0:08:45.160 --> 0:08:49.199 insert this electrode into the neural tissue and you listen 0:08:49.679 --> 0:08:53.480 to the electrical activity of the cells. And researchers pretty 0:08:53.559 --> 0:08:56.079 quickly figured out that if you send a little bit 0:08:56.080 --> 0:08:59.480 of electricity down the wire down this electrode, you can 0:08:59.520 --> 0:09:03.360 stimulate the cell to make it active where it pops 0:09:03.400 --> 0:09:07.079 off its own little electrical spikes that travel around. So 0:09:07.400 --> 0:09:10.520 you put in some electricity and it goes And this 0:09:10.600 --> 0:09:14.880 is the technology behind, for example, deep brain stimulation you 0:09:14.920 --> 0:09:18.439 might have heard of this. Take Parkinson's disease. There's a 0:09:18.640 --> 0:09:22.960 tiny brain region called the subthalamic nucleus, and it was 0:09:23.000 --> 0:09:26.280 discovered starting from work in the nineteen seventies that you 0:09:26.320 --> 0:09:29.800 can insert your electrode into this area and zap it 0:09:29.840 --> 0:09:32.640 with a bit of electricity and you get these amazing 0:09:32.679 --> 0:09:39.040 effects of the movement problems of Parkinson's essentially disappearing. And 0:09:39.080 --> 0:09:41.479 by the way, the reason you can stick an electrode 0:09:41.520 --> 0:09:44.200 into the brain is because the brain doesn't have any 0:09:44.280 --> 0:09:47.280 pain receptors, so you can just dunk the little metal 0:09:47.280 --> 0:09:50.640 wire right in there after you've opened a little portal 0:09:50.760 --> 0:09:53.520 in the skull. So what's happening when you put these 0:09:53.520 --> 0:09:57.240 little bursts of electricity in is that the cells fire, 0:09:57.280 --> 0:09:59.560 which has effects on the rest of the network that 0:09:59.600 --> 0:10:02.440 those cells are connected to, and it also changes the 0:10:02.520 --> 0:10:06.080 electrical oscillations. And why this works so well in Parkinson's 0:10:06.120 --> 0:10:08.400 is still a bit of a mystery, but you get 0:10:08.679 --> 0:10:10.720 what you want out of it, and people have been 0:10:10.800 --> 0:10:14.400 using this sort of brain stimulation for all kinds of purposes. 0:10:14.600 --> 0:10:19.080 For example, my colleague Helen Mayberg puts electrodes directly into 0:10:19.120 --> 0:10:23.000 a very specific area near the singulate gyrus, and she 0:10:23.120 --> 0:10:27.439 stimulates and can pull people out of deep clinical depression 0:10:27.520 --> 0:10:30.720 this way. So there are many labs and clinics using 0:10:30.760 --> 0:10:34.440 the technique of stimulating individual cells in the brain, and 0:10:34.520 --> 0:10:37.440 the direction of the technology over the past couple of 0:10:37.480 --> 0:10:41.760 decades has been getting more and more electrodes implanted, so 0:10:41.800 --> 0:10:44.439 that you're not just hitting one or a few cells 0:10:44.440 --> 0:10:47.240 at the tip of the electrode, but you're instead exciting 0:10:47.640 --> 0:10:51.360 tens or hundreds or eventually thousands of cells by using 0:10:51.400 --> 0:10:56.679 a whole specific collection of electrodes. And companies like Neuralink 0:10:56.720 --> 0:10:59.520 have become famous in the public eye because of the 0:10:59.559 --> 0:11:03.440 idea of sewing these electrodes very finely into the brain 0:11:03.800 --> 0:11:06.680 and getting a thousand of them and soon more than that. 0:11:06.920 --> 0:11:10.640 And in all these cases, the electrodes can read and write, 0:11:10.679 --> 0:11:13.040 in other words, they can record the activity in the 0:11:13.040 --> 0:11:16.040 brain cells, but they can also stimulate the brain cells 0:11:16.240 --> 0:11:34.160 to put activity in there. So once you have the 0:11:34.200 --> 0:11:37.640 electrodes in there, could you just send in the right 0:11:37.880 --> 0:11:41.080 zaps of electricity in just the right pattern, spread over 0:11:41.120 --> 0:11:45.040 millions of neurons with precise timing of your patterns in 0:11:45.240 --> 0:11:49.040 such a way that you shape the network so that 0:11:49.120 --> 0:11:53.040 you can fly a helicopter. Now, all that sounds pretty 0:11:53.040 --> 0:11:57.280 exciting as a theoretical possibility, but I think there are 0:11:57.280 --> 0:12:01.160 two major technical hurdles here to be able to stimulate 0:12:01.240 --> 0:12:03.320 lots and lots of cells in the brain in the 0:12:03.360 --> 0:12:07.679 way that you might want to upload helicopter instructions. The 0:12:07.720 --> 0:12:13.240 first is simply a physical challenge. The brain is very delicate, 0:12:13.600 --> 0:12:17.080 and so Mother Nature has surrounded it in the armored 0:12:17.160 --> 0:12:20.280 plating of the skull. So it's very very hard to 0:12:20.520 --> 0:12:24.280 get at this fragile, delicate tissue of the brain, and 0:12:24.320 --> 0:12:26.280 so if you want to insert an electrode, you have 0:12:26.320 --> 0:12:29.440 to actually drill a small hole in the skull to 0:12:29.600 --> 0:12:32.600 expose the brain and then you can put your electrode in. 0:12:32.920 --> 0:12:36.599 The difficulty is that there are eighty six billion neurons, 0:12:36.640 --> 0:12:40.560 and at the moment, even with our fanciest technology, we 0:12:40.600 --> 0:12:43.600 can only get to say a thousand of these at 0:12:43.640 --> 0:12:47.320 any time, and so that is useless in terms of 0:12:47.400 --> 0:12:51.120 actually having access to the whole system. It would be 0:12:51.160 --> 0:12:54.720 equivalent to if you really wanted to say something to 0:12:54.840 --> 0:12:58.199 all eight billion people on the planet, but you only 0:12:58.240 --> 0:13:02.680 had one hundred followers on social media. The huge majority 0:13:02.679 --> 0:13:04.760 of the world will have no idea that you've ever 0:13:05.000 --> 0:13:08.480 said anything, or that you even exist. And that's the situation. 0:13:08.880 --> 0:13:12.160 When you zap a few hundred neurons, the other tens 0:13:12.200 --> 0:13:15.199 of billions of neurons don't even know that you're knocking 0:13:15.240 --> 0:13:19.560 on the door. So to actually insert information into the brain, 0:13:19.640 --> 0:13:23.160 you'd somehow need to access all or at least most 0:13:23.240 --> 0:13:27.240 of the neurons to make any meaningful change. Now, I'm 0:13:27.280 --> 0:13:29.880 not yet addressing how you would know what you want 0:13:29.920 --> 0:13:31.760 to change, where I'll come back to that in a moment. 0:13:32.080 --> 0:13:35.120 Let's just imagine for now that you know exactly what 0:13:35.160 --> 0:13:37.199 you want to tweak in the brain. Now, I do 0:13:37.240 --> 0:13:40.160 think that in the future there may be a very 0:13:40.160 --> 0:13:45.920 different solution besides electrodes to this issue of manipulating the network, 0:13:45.960 --> 0:13:48.880 because I don't think the idea of dunking electrodes in 0:13:48.880 --> 0:13:52.280 there is ever going to be a long term solution. 0:13:52.800 --> 0:13:55.199 When I squint into the future, I think the solution 0:13:55.360 --> 0:13:59.400 is something like nano robots. So what are nano robots. 0:13:59.440 --> 0:14:04.240 The idea is that you use atomically precise three D 0:14:04.440 --> 0:14:09.360 printing to make little molecular machines out of atoms. Essentially, 0:14:09.559 --> 0:14:14.040 you make little robots that carry out some functions, so 0:14:14.080 --> 0:14:18.720 they're like little robots, but they're microscopically small, built out 0:14:18.720 --> 0:14:23.320 of individual atoms, by the way, which is what proteins are. Anyway, 0:14:23.360 --> 0:14:27.000 you could make these super durable, for example by printing 0:14:27.040 --> 0:14:31.920 them out of carbon, making them diamond robots. The idea, 0:14:32.240 --> 0:14:35.120 and this is probably not for several decades. The idea 0:14:35.240 --> 0:14:39.240 is that you swallow a pill with tens of billions 0:14:39.280 --> 0:14:42.320 of these little nano robots in there, and they float 0:14:42.360 --> 0:14:45.800 through your bloodstream and you give them the right FedEx 0:14:45.880 --> 0:14:50.080 labels to pass the blood brain barrier, and once they're 0:14:50.080 --> 0:14:53.360 in there in the brain, they wiggle their way into 0:14:53.400 --> 0:14:56.720 your neurons where they can read the activity and they 0:14:56.720 --> 0:15:01.400 can cause the cell to spike to fire signal whenever 0:15:01.440 --> 0:15:05.920 they need to. So, with proper signaling between the nanobots, 0:15:06.280 --> 0:15:10.640 using for example, mesh networking, you could in theory generate 0:15:10.760 --> 0:15:14.680 whatever patterns you needed to across the entire brain, and 0:15:14.720 --> 0:15:18.600 if your science is really advanced, then you hit the 0:15:18.760 --> 0:15:22.920 correct brain wide patterns that will cement in the knowledge 0:15:23.400 --> 0:15:26.560 of how to fly a B two twelve. Now, although 0:15:26.560 --> 0:15:31.240 this is not happening anytime soon, it certainly seems plausible 0:15:31.480 --> 0:15:34.560 that this could be in our future. But wait, there's 0:15:34.680 --> 0:15:38.760 actually a difficult twist to this story. I said before 0:15:38.840 --> 0:15:42.120 there are two technical hurdles, and here comes the second. 0:15:42.520 --> 0:15:47.040 And that hurdle is that there won't be a single 0:15:47.560 --> 0:15:51.320 program for flying a B two twelve helicopter. Why not, 0:15:52.240 --> 0:15:56.200 because the brain inside each of us is totally unique. 0:15:56.240 --> 0:15:59.640 We each have a massive forest of eighty six billion 0:15:59.680 --> 0:16:04.160 euro on each with ten thousand connection points reaching out 0:16:04.400 --> 0:16:08.840 and interacting with other trees. And it's a living forest 0:16:08.960 --> 0:16:12.880 such as each connection, every twig on every branch finds 0:16:12.920 --> 0:16:16.360 its place in life based on the exact details of 0:16:16.400 --> 0:16:22.120 what you have seen and heard and experienced in your life. 0:16:22.160 --> 0:16:27.400 You born in your hometown, with your family, your neighborhood, 0:16:27.520 --> 0:16:32.320 your culture, your moment in history. All those things determine 0:16:32.360 --> 0:16:35.520 the exact wiring of your brain. And your brain has 0:16:35.560 --> 0:16:37.960 a network that is different from his brain over there, 0:16:38.160 --> 0:16:41.320 and her brain over there, and everyone else's brain on 0:16:41.360 --> 0:16:46.720 the planet. And the exact wiring is what makes you you. 0:16:47.560 --> 0:16:51.400 So in the proposed future of the Matrix, the operator 0:16:51.520 --> 0:16:55.000 Tank would have to specify that he wants a program 0:16:55.400 --> 0:16:59.600 to pilot a B two twelve helicopter that is specified 0:16:59.720 --> 0:17:04.480 exact exactly for Trinity's brain, that is bespoke for her 0:17:04.640 --> 0:17:08.840 neural network only. And if Tank tried to upload the 0:17:08.880 --> 0:17:13.200 same program to Neo's brain or Morpheus's brain, who knows 0:17:13.320 --> 0:17:17.240 what that would result in. Because if the program alters 0:17:17.320 --> 0:17:20.840 the way that neuron nineteen million, three hundred fifty six 0:17:20.840 --> 0:17:23.879 three hundred and two is talking to its neighbors, and 0:17:23.960 --> 0:17:28.399 it does this over a million other neurons with high specificity, 0:17:28.680 --> 0:17:31.679 that might teach Trinity how to fly a helicopter, but 0:17:31.760 --> 0:17:35.080 it certainly would not work for someone else whose brain 0:17:35.520 --> 0:17:38.120 is different. So how do we get around that problem, 0:17:38.160 --> 0:17:42.800 the problem of everyone having a unique neural network. Well, 0:17:43.359 --> 0:17:46.320 the answer will have to rely on what is called 0:17:46.640 --> 0:17:51.320 system identification. This is an engineering approach where you have 0:17:51.440 --> 0:17:56.960 some complicated dynamic system and you measure lots of input 0:17:57.080 --> 0:18:01.840 output pairs, as in, when I put this in, what happens? Okay, 0:18:02.000 --> 0:18:04.520 now it happens if I put that in. So imagine 0:18:04.520 --> 0:18:07.960 you find a really complicated machine and you don't know 0:18:08.000 --> 0:18:10.240 exactly what it does. So you tap one of the 0:18:10.320 --> 0:18:13.640 keys and you see how it moves, and then you 0:18:13.680 --> 0:18:15.760 tap three of the keys at the same time, and 0:18:15.800 --> 0:18:18.320 you look at what it does as its output, and 0:18:18.359 --> 0:18:20.600 then you hit a series of the keys and you 0:18:20.640 --> 0:18:23.520 see what results. And you do this over and over 0:18:23.560 --> 0:18:26.080 and over to try to figure out what is the 0:18:26.160 --> 0:18:31.560 structure under the hood. This system identification approach is used 0:18:31.560 --> 0:18:34.320 in lots of fields. For example, in economics, let's say 0:18:34.320 --> 0:18:37.159 you want to figure out the guts of the stock market. 0:18:37.320 --> 0:18:40.840 So you take lots of inputs like gross domestic product 0:18:40.920 --> 0:18:44.000 and inflation and unemployment and interest rates and blah blah blah, 0:18:44.080 --> 0:18:46.480 and you look at all these as inputs and you 0:18:46.520 --> 0:18:49.080 look at the reaction of the market this way, and 0:18:49.119 --> 0:18:52.200 you develop better and better mathematical models of what the 0:18:52.320 --> 0:18:55.800 machinery of the stock market is doing, even though you 0:18:55.880 --> 0:18:59.359 can't see it. Okay, So the question is, could you 0:18:59.440 --> 0:19:04.760 do system identification on a human brain. No one's ever 0:19:04.920 --> 0:19:07.400 really done this because there's no purpose for it now, 0:19:07.840 --> 0:19:11.399 but someday it might make sense. So the idea is 0:19:11.480 --> 0:19:15.639 you go into a super futuristic brain scanner and you 0:19:15.680 --> 0:19:20.199 get lots of inputs, and this sophisticated brain imaging device 0:19:20.440 --> 0:19:23.800 measures the outputs, in other words, which cells in your 0:19:23.840 --> 0:19:27.959 brain are responding. So you see a rapid series of 0:19:28.040 --> 0:19:30.960 images and you hear words, and you feel touches on 0:19:31.040 --> 0:19:34.119 your body, and you smell smells, and you run through 0:19:34.520 --> 0:19:39.480 thousands or maybe millions of little micro experiences while your 0:19:39.520 --> 0:19:42.760 brain is getting measured. And in theory, this is how 0:19:42.800 --> 0:19:47.600 a scientist could say, Aha, Trinity's brain is organized like this, 0:19:47.760 --> 0:19:51.400 while Neo's brain is laid out like that, and Morpheus's 0:19:51.440 --> 0:19:54.719 brain has a slightly different pattern, And you might find 0:19:54.800 --> 0:19:59.480 that for teaching the operation of a B two twelve helicopter, 0:20:00.119 --> 0:20:04.520 in his brain thinks about it in analogy to riding 0:20:04.560 --> 0:20:07.440 a horse and controlling it, which let's say she grew 0:20:07.560 --> 0:20:11.760 up riding horses, while Neo's brain would learn the helicopter 0:20:11.960 --> 0:20:15.399 in analogy to the way a motorcycle feels, which is, 0:20:15.440 --> 0:20:18.040 let's say how he grew up. And for Morpheus, the 0:20:18.200 --> 0:20:22.520 actions of piloting emerge from his deep knowledge of surfing, 0:20:22.560 --> 0:20:24.359 which is how he grew up and what is stored 0:20:24.400 --> 0:20:27.760 in his brain. Now, it's not clear how many inputs 0:20:27.760 --> 0:20:30.440 you'd have to ping in there to get high enough 0:20:30.520 --> 0:20:33.960 resolution to make all the little changes you need, but 0:20:34.119 --> 0:20:39.160 presumably that would get figured out with enough experimentation. Okay, 0:20:39.480 --> 0:20:43.720 so let's say we as a society grow to a 0:20:43.800 --> 0:20:48.280 point where we can do system identification on an individual's 0:20:48.320 --> 0:20:54.720 brain and then use nanobots to upload knowledge of helicopter piloting. 0:20:55.119 --> 0:20:57.920 I need to emphasize that this is not right around 0:20:57.960 --> 0:21:02.040 the corner, but it certainly seems the theoretically plausible. Another 0:21:02.680 --> 0:21:07.040 century of advancement, and suddenly the network that makes you 0:21:07.840 --> 0:21:12.520 can get directed and shaped in a bespoke manner. And 0:21:12.560 --> 0:21:14.719 if we come to a point where we can do it, 0:21:15.080 --> 0:21:19.760 that's possibly the biggest societal change. I can imagine you 0:21:20.080 --> 0:21:22.000 say to your three year old kid, Okay, we're gonna 0:21:22.040 --> 0:21:25.000 upload first grade now. Great, Now, go play outside for 0:21:25.040 --> 0:21:28.159 an hour, and then we're gonna upload second grade after lunch. 0:21:28.560 --> 0:21:31.400 Imagine that by the end of the week, your three 0:21:31.480 --> 0:21:35.120 year old knows as much as a full professor does. Now, 0:21:35.760 --> 0:21:39.000 so what becomes of society and the way we run 0:21:39.040 --> 0:21:41.639 it now? You may think the analogy here is to 0:21:41.640 --> 0:21:45.000 look at super smart, genius kids in our current world, 0:21:45.000 --> 0:21:48.439 But these kids often go off to attend college at 0:21:48.480 --> 0:21:51.520 twelve years old, and they very often end up lonely 0:21:51.560 --> 0:21:54.199 and socially misplaced, because really what they want is to 0:21:54.200 --> 0:21:57.320 play with their colleagues other kids their age, But they 0:21:57.320 --> 0:22:00.040 get stuck with a bunch of older kids who have 0:22:00.119 --> 0:22:04.600 gone through puberty and are running deeply carved evolutionary programs 0:22:04.920 --> 0:22:07.920 that cause their brains to be taken over by sexuality, 0:22:08.160 --> 0:22:10.560 and that software hasn't yet turned on in the heads 0:22:10.560 --> 0:22:13.439 of these young genuses, and as a result, they can't 0:22:13.480 --> 0:22:15.760 mesh with what is happening around them, and they can 0:22:15.800 --> 0:22:21.159 feel very lonely in these contexts. But the future scenario 0:22:21.240 --> 0:22:25.399 of uploading knowledge is totally different because now every single 0:22:25.480 --> 0:22:30.560 kid can stay among colleagues. But the question is if 0:22:30.680 --> 0:22:33.960 education is uploaded, what do the kids do all day? 0:22:34.440 --> 0:22:38.159 Do they launch startups at the age of six, do 0:22:38.200 --> 0:22:41.600 they write epic novels by the time they're eight years old? 0:22:42.240 --> 0:22:46.520 Do they return to reproducing as teenagers like their distant 0:22:46.560 --> 0:22:50.480 ancestors did? And is it dangerous that they have all 0:22:50.640 --> 0:22:54.800 the knowledge of decades of schooling but without the maturity. 0:22:55.320 --> 0:22:57.840 The most slowly developing part of the brain is the 0:22:57.840 --> 0:23:02.760 prefrontal cortex, and this underlies our ability to simulate possible 0:23:02.800 --> 0:23:06.520 futures and think about consequences. So imagine a kid with 0:23:06.560 --> 0:23:10.359 an undeveloped prefederal cortex who has all the knowledge that 0:23:10.440 --> 0:23:15.040 Albert Einstein commanded at midlife. But this child lacks the 0:23:15.080 --> 0:23:19.280 ability to simulate consequences, so they think something like, wouldn't 0:23:19.280 --> 0:23:22.359 it be hilarious to build a small nuclear bomb and 0:23:22.440 --> 0:23:25.719 blow up my neighbor's porch, Or wouldn't it be a 0:23:25.760 --> 0:23:31.840 crackup to disrupt the presidential broadcast by hijacking the frequency 0:23:31.880 --> 0:23:36.320 and imposing a video of me twerking or whatever? Because 0:23:36.440 --> 0:23:40.240 children don't yet have a fully developed profederal cortex that 0:23:40.359 --> 0:23:44.080 can't simulate consequences the way an adult can, and this 0:23:44.200 --> 0:23:47.840 is why it could be dangerous to inject the knowledge 0:23:47.880 --> 0:24:04.840 of an adult into a child's brain. Now, perhaps I'm 0:24:04.840 --> 0:24:09.240 being shortsighted here, and we could somehow upload maturity as well. 0:24:09.480 --> 0:24:14.240 We could figure out the learning that translates to morally 0:24:14.240 --> 0:24:18.800 complex situations and simulate those over and over do the 0:24:18.840 --> 0:24:25.240 synaptic equivalent of working through the possibilities and feeling the consequences. 0:24:25.520 --> 0:24:29.560 Maybe you could massively speed up emotional learning that way. 0:24:29.680 --> 0:24:32.360 After all, as my father would always tell me, the 0:24:32.400 --> 0:24:37.600 wise person learns from experience, but the wiser person learns 0:24:37.640 --> 0:24:41.280 from the experience of others. So maybe there could be 0:24:41.760 --> 0:24:47.360 enough uploaded knowledge where a kid understands various possible scenarios 0:24:47.400 --> 0:24:51.080 and outcomes, and the good decision making simply results from 0:24:51.080 --> 0:24:55.159 a deep knowledge of previous examples, things that have happened 0:24:55.160 --> 0:24:58.919 to other people, all of which have been uploaded. So 0:24:59.040 --> 0:25:01.919 maybe the maturity problem could be taken care of, but 0:25:02.040 --> 0:25:06.880 still we're looking at massive societal shifts that would render 0:25:07.000 --> 0:25:12.320 our current civilization totally unrecognizable. Now, we all like to 0:25:12.320 --> 0:25:15.600 be very thoughtful about the future, but it doesn't matter 0:25:15.840 --> 0:25:19.359 what we speculate about it, because we are guaranteed to 0:25:19.440 --> 0:25:22.679 be wrong. We can only envision what we're capable of, 0:25:22.760 --> 0:25:26.159 in this case, a cartoonish version of a bunch of 0:25:26.200 --> 0:25:29.200 super intelligent kids running around while their parents go off 0:25:29.200 --> 0:25:31.720 to their jobs. But the world is likely to be 0:25:32.440 --> 0:25:36.720 very different by then. Presuming that everything is massively sped 0:25:36.800 --> 0:25:41.639 up by artificial intelligence, it seems very possible that society 0:25:41.720 --> 0:25:45.600 is going to evolve exponentially faster at a pace that 0:25:45.680 --> 0:25:49.359 we really can't conceive of here in the first third 0:25:49.400 --> 0:25:52.199 of the twenty first century. I mean, just imagine that 0:25:52.400 --> 0:25:57.360 AI knocks down scientific problems rapidly, such that we move 0:25:57.440 --> 0:26:02.480 from our current state of pretty wide spread ignorance to perfect, 0:26:02.640 --> 0:26:06.639 wonderful models of everything in the cosmos. Just think about 0:26:06.680 --> 0:26:11.560 the incredibly slow pace between the Stone Age and the 0:26:11.560 --> 0:26:14.800