1 00:00:03,400 --> 00:00:10,120 Welcome, I hope this is working. Welcome to neural link up live 2 00:00:10,120 --> 00:00:12,440 update. We're going to tell you about 3 00:00:13,240 --> 00:00:18,320 the progress of the first patient with neural link and 4 00:00:19,600 --> 00:00:22,600 sort of your recap of of the progress there. 5 00:00:23,200 --> 00:00:27,320 Then talk about what changes we were making for the second 6 00:00:27,320 --> 00:00:31,440 patient, which we're hoping to do an implant in the next week 7 00:00:31,440 --> 00:00:33,920 or so. And this is for our first 8 00:00:33,920 --> 00:00:36,440 product, which is called Telepathy, which enables you to 9 00:00:36,440 --> 00:00:38,840 control a computer or a phone just by thinking. 10 00:00:39,680 --> 00:00:42,320 So let's in fact. So we'll start off with just 11 00:00:42,320 --> 00:00:44,120 some introductions. DJ you want to start? 12 00:00:44,640 --> 00:00:47,880 Hi everyone, My name is DJ Saul. I'm an electrical engineer and a 13 00:00:47,880 --> 00:00:49,080 chip designer. By training. 14 00:00:49,760 --> 00:00:54,080 I led the design of first several generations of the 15 00:00:54,120 --> 00:00:57,280 Neurolink implant. Currently I was on the founding 16 00:00:57,280 --> 00:01:00,840 team and currently a President. I'm Matthew McDougal. 17 00:01:00,840 --> 00:01:03,920 I'm a practicing neurosurgeon and head of Neurosurgery at 18 00:01:03,920 --> 00:01:06,400 Neurolink. Yeah, go ahead. 19 00:01:06,560 --> 00:01:09,440 Yeah. Head of Branding Interfaces 20 00:01:09,440 --> 00:01:11,160 applications. And I'm bliss. 21 00:01:11,160 --> 00:01:12,960 I'm a software engineer at Neurolink trying to figure out 22 00:01:12,960 --> 00:01:15,400 how to turn brand activity into cool stuff in the world. 23 00:01:16,360 --> 00:01:16,920 All right. Thank you. 24 00:01:16,920 --> 00:01:19,800 Well, let's see. So we'll just get going into the 25 00:01:19,800 --> 00:01:24,720 presentation. So our first product is sort of 26 00:01:25,000 --> 00:01:30,440 like I said, we call telepathy, which is enables the person with 27 00:01:30,440 --> 00:01:33,080 Neuralink implant to control their phone or computer just by 28 00:01:33,080 --> 00:01:35,240 thinking. And once you can control your 29 00:01:35,240 --> 00:01:37,320 phone or computer, you can essentially control almost 30 00:01:37,320 --> 00:01:39,920 anything just and literally just by thinking. 31 00:01:39,920 --> 00:01:45,120 So there's no eye tracking or anything, it is purely, purely 32 00:01:45,120 --> 00:01:50,840 your thoughts. So this is really quite a 33 00:01:50,840 --> 00:01:54,680 profound device that can help a lot of people who have lost the 34 00:01:54,680 --> 00:01:56,520 connection between their brain and body. 35 00:01:56,960 --> 00:02:03,320 So imagine people like Stephen Hawking, who, you know, imagine 36 00:02:03,320 --> 00:02:06,160 if he if he could communicate at the same speed as someone who 37 00:02:06,160 --> 00:02:07,880 had still had the connection to their brain and body. 38 00:02:08,919 --> 00:02:12,960 So it's really something that can help millions of people 39 00:02:12,960 --> 00:02:16,080 around the world. And it's a, it's part of our 40 00:02:16,080 --> 00:02:21,080 overall goal of enabling a very high bandwidth connection 41 00:02:21,080 --> 00:02:25,400 between the brain and, and your and the rest of the world and 42 00:02:25,400 --> 00:02:28,600 your computers. The long term goal to which 43 00:02:28,600 --> 00:02:33,760 sounds a little esoteric, is to mitigate the the risk of, of the 44 00:02:33,760 --> 00:02:39,560 civil civilizational risk of AI by having a sort of closer 45 00:02:40,040 --> 00:02:43,440 biosis between human intelligence and digital 46 00:02:43,440 --> 00:02:46,520 intelligence. But that, that'll take many 47 00:02:46,520 --> 00:02:50,880 years along the way, we're, we're going to help solve a lot 48 00:02:50,880 --> 00:02:54,640 of brain injury or spinal injury issues. 49 00:02:55,280 --> 00:03:00,280 So, and then with our first product apathy, that's, that's 50 00:03:00,440 --> 00:03:04,040 going to be really quite profound that there is also 51 00:03:04,040 --> 00:03:09,960 potential long term for bridging the gaps or if there are damaged 52 00:03:09,960 --> 00:03:14,960 or severed neurons being able to span the gap between the brain's 53 00:03:14,960 --> 00:03:18,680 motor cortex to the spine to enable someone to use their body 54 00:03:18,680 --> 00:03:20,280 again. I think that would be very 55 00:03:20,280 --> 00:03:24,640 exciting. And it's, you know, that that is 56 00:03:24,640 --> 00:03:27,000 something that is possible in the long term. 57 00:03:29,440 --> 00:03:34,040 And then our second product, which we've demonstrated to work 58 00:03:34,120 --> 00:03:38,200 with monkeys is Blind side, which would enable someone who 59 00:03:38,200 --> 00:03:42,680 is completely blind or lost both eyes or completely lost their 60 00:03:42,680 --> 00:03:48,040 optic nerve to be able to see. So that's that's something that 61 00:03:48,040 --> 00:03:49,800 we hope to demonstrate in the future. 62 00:03:55,470 --> 00:03:59,160 So let's just give you a sense of what the device is a way to 63 00:03:59,160 --> 00:04:01,120 think about the neural link device is kind of like a a 64 00:04:01,240 --> 00:04:06,360 Fitbit or an Apple Watch with with tiny wires or electrodes. 65 00:04:06,360 --> 00:04:10,560 Those those tiny wires are implanted in the in the brain 66 00:04:10,600 --> 00:04:14,280 and they read and write electrical signals. 67 00:04:14,840 --> 00:04:19,079 So a lot of people think the brain is incredibly mysterious 68 00:04:19,640 --> 00:04:21,079 thing. It's it, it is mysterious in a 69 00:04:21,079 --> 00:04:24,040 lot of ways, but but it is actually, it does operate with 70 00:04:24,040 --> 00:04:26,640 like electrical signals. So if you can read and write 71 00:04:26,640 --> 00:04:29,200 those electrical signals, you can interface with the brain. 72 00:04:29,880 --> 00:04:37,240 And the devices is sized so that it is the same size as the as 73 00:04:37,240 --> 00:04:38,600 the piece of skull that is removed. 74 00:04:38,600 --> 00:04:43,360 So it's like a few centimeters diameter of skull that's 75 00:04:43,360 --> 00:04:45,360 removed. We replaced that with the device 76 00:04:46,280 --> 00:04:50,640 after implanting the tiny wires with a surgical robot and that 77 00:04:50,640 --> 00:04:54,320 enables read write capability to the neurons. 78 00:04:54,880 --> 00:04:56,960 Completely wirelessly. Yeah, yes, exactly. 79 00:04:57,640 --> 00:05:01,840 It's completely wirelessly. So like I could I could have a 80 00:05:01,840 --> 00:05:03,000 neural link right now, you wouldn't know. 81 00:05:04,560 --> 00:05:08,360 And it charges inductively. So you could just basically have 82 00:05:08,360 --> 00:05:13,400 an electromagnetic ad that you charge the device with. 83 00:05:13,640 --> 00:05:15,760 So yeah, it's like an Apple Watch. 84 00:05:15,760 --> 00:05:19,160 Exactly so. Except that it's actually a much 85 00:05:19,280 --> 00:05:23,040 harder technical challenge to solve given that there's limit 86 00:05:23,040 --> 00:05:26,560 as to how much heat the brain tissue whereas in for phones. 87 00:05:26,560 --> 00:05:28,160 And have a hot watch. Or whatever you really care how 88 00:05:28,440 --> 00:05:29,960 much if it's sitting on a table. Sure. 89 00:05:31,200 --> 00:05:33,360 Yeah, it's also it's got to go through scan and stuff as well 90 00:05:33,360 --> 00:05:36,800 in our case. So it is a tougher challenge to 91 00:05:36,800 --> 00:05:40,960 to charge and to have I bandwidth communications given 92 00:05:40,960 --> 00:05:43,480 that it's got to go through skin and hair and stuff. 93 00:05:43,640 --> 00:05:45,200 We have solved it. We have solved it, yeah. 94 00:05:48,360 --> 00:05:49,840 So yeah. Yeah. 95 00:05:49,840 --> 00:05:54,280 So our first step with the telepathy is basically to unlock 96 00:05:54,400 --> 00:05:59,440 digital independence for people with policies and to allow them 97 00:05:59,440 --> 00:06:02,400 to control the computer just with their mind without moving 98 00:06:02,400 --> 00:06:05,800 the body. And our goal is to provide them 99 00:06:05,800 --> 00:06:09,560 the same level of control, functionality and reliability 100 00:06:09,560 --> 00:06:12,720 that I have when I'm using a computer, even better than the 101 00:06:12,840 --> 00:06:15,400 level of control I have. And it's not a high bar for. 102 00:06:15,640 --> 00:06:19,280 You just to be clear, this guy, he's controlling this with his 103 00:06:19,280 --> 00:06:21,520 brain, so he's not like you can't see his hands in this 104 00:06:21,520 --> 00:06:23,560 video, but he's not using a mouse and keyboard. 105 00:06:23,800 --> 00:06:26,320 Just, you know, thinking about how to move the cursor and 106 00:06:26,320 --> 00:06:28,200 playing Civilization. No eye tracker. 107 00:06:28,200 --> 00:06:29,520 Right, there's no eye tracking from I. 108 00:06:29,600 --> 00:06:31,400 Mean he's lost. He's lost Junior. 109 00:06:31,440 --> 00:06:32,680 He's like, he's going to watch this on Twitter. 110 00:06:32,680 --> 00:06:34,000 Just thinking that's it. Just. 111 00:06:34,280 --> 00:06:35,280 Thinking like just a couple days. 112 00:06:35,280 --> 00:06:37,600 Like cursor move here. Yeah, yeah, yeah. 113 00:06:37,600 --> 00:06:39,280 This is like last night or two nights ago something. 114 00:06:39,360 --> 00:06:41,200 Yeah. I think I think the way he also 115 00:06:41,200 --> 00:06:43,960 described it is he's using the. Yeah. 116 00:06:44,480 --> 00:06:47,760 He has many more videos on his on the platform. 117 00:06:48,560 --> 00:06:51,960 Definitely check them out. Yeah, so he can, he's streaming 118 00:06:51,960 --> 00:06:54,440 that live and also can talk and like move his head without a 119 00:06:54,480 --> 00:06:57,280 problem multi time. Yeah, you guys like if you join 120 00:06:57,280 --> 00:06:59,400 this live stream, you can ask him questions, he'll he'll tell 121 00:06:59,400 --> 00:07:01,080 you all about what it's like to move. 122 00:07:02,160 --> 00:07:06,240 Also I think I haven't played civilization myself but I think 123 00:07:06,240 --> 00:07:08,120 this is actually not easy mode. This is expert this. 124 00:07:08,360 --> 00:07:11,120 Is emperor mode emperor mode? If you have played stiff emperor 125 00:07:11,120 --> 00:07:12,520 mode, it's like the highest difficulty level. 126 00:07:12,760 --> 00:07:15,280 Just the point is like this is a commonly demanding task while 127 00:07:15,280 --> 00:07:18,440 live streaming playing the hardest mode of team and he's 128 00:07:18,440 --> 00:07:22,040 able to do that while moving. Talking, engaging with you know 129 00:07:22,040 --> 00:07:25,440 the audience while playing. One of the other games he likes 130 00:07:25,440 --> 00:07:28,520 to play a lot is chess. I think it gets lost sometimes 131 00:07:28,520 --> 00:07:30,560 that he's actually playing speed chess. 132 00:07:30,560 --> 00:07:33,760 Against me, yeah. Which requires an incredibly 133 00:07:33,760 --> 00:07:38,360 high fidelity degree of control and and speed of control in 134 00:07:38,360 --> 00:07:45,000 order to be able to win. So also another cool stuff about 135 00:07:45,080 --> 00:07:49,040 about our device is that we can use it anywhere, anytime, also 136 00:07:49,040 --> 00:07:53,040 on a plane, during a flight, while creating really cool means 137 00:07:53,040 --> 00:07:57,840 of care. Also, our device unlocks things 138 00:07:57,840 --> 00:08:00,920 that previously were impossible for our participants. 139 00:08:01,360 --> 00:08:05,280 For example, we're able to connect him to his gaming 140 00:08:05,280 --> 00:08:08,680 consoles, which they Mario Kart with friends and family, and it 141 00:08:08,680 --> 00:08:12,960 was lovely to see them playing together after years that he 142 00:08:12,960 --> 00:08:14,640 couldn't do it since he's injured. 143 00:08:15,320 --> 00:08:18,040 Imagine if you're sitting umbrella over from the sky on a 144 00:08:18,080 --> 00:08:20,200 plane. You look over, he's making a cat 145 00:08:20,200 --> 00:08:23,520 meme. No hands, no movement. 146 00:08:23,560 --> 00:08:26,440 Yeah. Live in a real world. 147 00:08:26,520 --> 00:08:28,080 Yeah, it's strange, strange time. 148 00:08:29,880 --> 00:08:32,280 Yeah. And he loves using the device 149 00:08:32,280 --> 00:08:37,280 and using independently daily to watch videos, read, play games 150 00:08:38,640 --> 00:08:42,480 using the browser. And the key metrics that we care 151 00:08:42,480 --> 00:08:45,760 is to make sure our device is actually useful is to is 152 00:08:45,840 --> 00:08:48,600 basically the amount of hours we use the device daily and weekly. 153 00:08:49,000 --> 00:08:53,640 And we track it weekly since the since the surgery and on weeks 154 00:08:53,640 --> 00:08:57,600 that he's not too busy and not travelling, he can even reach 70 155 00:08:57,600 --> 00:09:00,400 hours of using the device a week. 156 00:09:00,600 --> 00:09:04,240 This is amazing and he would of course love to use it more, but 157 00:09:04,520 --> 00:09:09,120 need to run recent sessions, he needs to sleep sometimes and 158 00:09:09,120 --> 00:09:11,000 also of course to charge the device once in a while. 159 00:09:11,440 --> 00:09:13,200 Hopefully we'll improve that over time, I think. 160 00:09:14,040 --> 00:09:16,080 That maybe not obvious to people who are watching this. 161 00:09:16,080 --> 00:09:17,720 Like it's a normal MacBook he's controlling. 162 00:09:17,720 --> 00:09:20,560 This isn't like some limited efficient thing, but there's 163 00:09:20,560 --> 00:09:22,400 only a few options like he can just do anything that you can do 164 00:09:22,400 --> 00:09:24,800 on a MacBook Pro. Same one I have on my desk. 165 00:09:24,800 --> 00:09:25,840 Actually, it's the exact same one. 166 00:09:29,000 --> 00:09:33,360 And maybe another interesting point is that on the first day 167 00:09:33,360 --> 00:09:38,240 he used BCI brain control, he was able to break the previous 168 00:09:38,240 --> 00:09:42,480 world record for cursor control is by using the brain. 169 00:09:42,800 --> 00:09:47,600 And recently he even doubled it and was able to outperform about 170 00:09:47,600 --> 00:09:51,440 10% of our engineered neurolink. And you can be sure that we are 171 00:09:51,440 --> 00:09:53,040 very good in this game and very quick. 172 00:09:53,480 --> 00:09:56,280 And if you want to check out how well, how well you can do it, 173 00:09:56,280 --> 00:10:00,040 you can do it on our website. And it's very addictive games. 174 00:10:01,000 --> 00:10:02,920 Yeah, it's a very simple game. You just have to click on the 175 00:10:02,920 --> 00:10:07,040 square. But but it's it's it's it's 176 00:10:07,040 --> 00:10:11,240 actually, even though it sounds silly, it's it's quite a yeah, 177 00:10:11,360 --> 00:10:13,440 it can be quite a it sounds like it could be quite addictive. 178 00:10:13,440 --> 00:10:16,320 And it's especially if you get a low score and you think there's 179 00:10:16,360 --> 00:10:19,640 no way I got a. So I, I mean, anyone who wants 180 00:10:19,640 --> 00:10:23,040 to try this, I recommend going to theneurolink.com website and 181 00:10:23,040 --> 00:10:24,840 seeing, seeing if you can beat Nolan's record. 182 00:10:25,360 --> 00:10:27,520 And it's that you will find that's actually quite difficult 183 00:10:27,520 --> 00:10:31,320 to do so. And this is really with version 184 00:10:31,320 --> 00:10:34,280 one of the device and with only a small percentage of the 185 00:10:34,880 --> 00:10:36,680 electrodes that are that are working. 186 00:10:36,840 --> 00:10:40,320 So this is, this is really just the beginning, but even the 187 00:10:40,320 --> 00:10:43,760 beginning is twice as good as the world record. 188 00:10:43,880 --> 00:10:48,400 This is important to emphasize the, you know, the media has a 189 00:10:48,400 --> 00:10:53,000 habit of of saying that the glass is 10% empty, but but 190 00:10:53,000 --> 00:10:57,040 actually it's 90% full. So I think it's really quite an 191 00:10:57,040 --> 00:11:00,520 accomplishment of the neural link team to have achieved with 192 00:11:00,520 --> 00:11:05,480 first with the first patient, the first device twice the world 193 00:11:05,480 --> 00:11:09,080 record for the range computer bandwidth. 194 00:11:09,960 --> 00:11:16,080 That's really an astonishing, an amazingly great outcome and it's 195 00:11:16,080 --> 00:11:17,760 only going to be get better from here. 196 00:11:17,920 --> 00:11:21,320 So the potential is to ultimately get, I think to 197 00:11:21,640 --> 00:11:26,400 megabit level. So that's, that's part of the 198 00:11:26,400 --> 00:11:29,400 long term goal of, of improving the, the bandwidth of the brain 199 00:11:29,400 --> 00:11:33,160 computer interface. If you think about like how low 200 00:11:33,160 --> 00:11:37,000 the bandwidth normally is between a human and a device, 201 00:11:37,000 --> 00:11:38,640 it's the average bandwidth is extremely low. 202 00:11:38,640 --> 00:11:43,240 It's it's, I said less than one bit per second over the course 203 00:11:43,240 --> 00:11:45,960 of a day. If there are 86,400 seconds in a 204 00:11:45,960 --> 00:11:50,720 day, you're outputting less than that number of bits to any any 205 00:11:50,720 --> 00:11:53,800 given device except in perhaps very rare circumstances. 206 00:11:54,360 --> 00:12:05,080 So the this is actually quite important for for AI, you know, 207 00:12:05,080 --> 00:12:11,920 basically for for human AI, symbiosis is just being able to 208 00:12:11,920 --> 00:12:15,440 communicate at A at a speed. Yeah, I can follow. 209 00:12:19,600 --> 00:12:23,880 So yeah, just to emphasize again, he's performing at this 210 00:12:23,880 --> 00:12:28,080 extremely high level with about 15% of his channels functional. 211 00:12:30,080 --> 00:12:33,800 And so we want to mitigate any of the problems that led to that 212 00:12:33,800 --> 00:12:36,800 situation. So, you know, the brain is a 213 00:12:37,040 --> 00:12:39,920 fascinating organ. Share with you some of the 214 00:12:39,920 --> 00:12:42,800 secrets about the brain. During any typical brain 215 00:12:42,800 --> 00:12:46,880 surgery, a small amount of air is introduced into the skull. 216 00:12:47,920 --> 00:12:50,400 That's because neurosurgeons like to have as much room as 217 00:12:50,400 --> 00:12:54,880 possible around the brain. And so there's this little known 218 00:12:54,880 --> 00:13:01,720 control mechanism of allowing the CO2 concentration in the 219 00:13:01,720 --> 00:13:06,040 blood to rise a bit, which allows the brain to either 220 00:13:06,160 --> 00:13:09,760 expand or contract depending on where you target that CO2. 221 00:13:10,280 --> 00:13:13,400 But typically neurosurgeons will have the brain shrink by 222 00:13:13,440 --> 00:13:16,120 lowering CO2. What we're going to do in our 223 00:13:16,120 --> 00:13:20,600 future surgeries is keep the CO2 concentration actually quite 224 00:13:20,600 --> 00:13:23,760 normal, maybe even slightly elevated, and that'll allow the 225 00:13:23,760 --> 00:13:26,640 brain to stay its normal size and shape during surgery. 226 00:13:27,000 --> 00:13:29,520 That should eliminate this air pocket that we saw in the first 227 00:13:29,520 --> 00:13:32,560 participant. Now that air pocket we think may 228 00:13:32,560 --> 00:13:37,080 have contributed to eating up some of the thread slack is as 229 00:13:37,080 --> 00:13:40,320 the air bubble migrated to be under the implant, push the 230 00:13:40,320 --> 00:13:44,080 brain away from the implant. And so that's easy enough to 231 00:13:44,080 --> 00:13:48,960 fix. Another consideration that we 232 00:13:48,960 --> 00:13:53,920 want to focus on for our upcoming participants is that 233 00:13:54,400 --> 00:13:57,920 the brain, think of it like a really complex folded onion. 234 00:13:57,920 --> 00:14:02,840 It's layer upon layer of sheets of neurons all over the surface 235 00:14:02,840 --> 00:14:07,720 of the brain folded into this, you know, odd looking shape. 236 00:14:08,800 --> 00:14:12,560 The folds of the brain travel down deep into the brain and and 237 00:14:12,720 --> 00:14:15,920 along with it go those onion layers of neurons. 238 00:14:16,520 --> 00:14:20,520 And if we insert very close to one of the folds where there may 239 00:14:20,520 --> 00:14:24,120 be very useful information encoded in neurons, we may end 240 00:14:24,120 --> 00:14:28,400 up travelling with our threads parallel to some of the layers 241 00:14:28,400 --> 00:14:30,840 of neurons that we're most interested in, avoiding them 242 00:14:30,840 --> 00:14:33,520 entirely. To avoid that possibility, we're 243 00:14:33,520 --> 00:14:37,920 going to insert in our future participants more close to the 244 00:14:37,920 --> 00:14:43,200 middle of the apex of the folds, ensuring that we're crossing the 245 00:14:43,200 --> 00:14:45,560 layers of interest, layer five of the cortex. 246 00:14:45,840 --> 00:14:50,480 I also think that it's important to highlight here those tiny 247 00:14:50,480 --> 00:14:53,720 wires that Elon mentioned, They're fraction of a human 248 00:14:53,720 --> 00:14:55,080 hair. They're very flexible, 249 00:14:55,600 --> 00:14:58,720 intentionally so, because brain is constantly moving and you 250 00:14:58,720 --> 00:15:02,000 want the electrodes to be moving with the brain, causing less of 251 00:15:02,000 --> 00:15:07,040 the scarring. And it's actually impossible for 252 00:15:07,040 --> 00:15:10,440 a human neurosurgeon, however talented Matthew is to actually 253 00:15:10,440 --> 00:15:13,400 maneuver them by. So we have a surgical robot that 254 00:15:13,400 --> 00:15:16,320 we built that can actually precisely target them in any 3 255 00:15:16,320 --> 00:15:19,760 dimensional space, XY as well as Z with Micron level precision 256 00:15:20,040 --> 00:15:23,440 while avoiding vasculature so that you don't disrupt the the 257 00:15:23,800 --> 00:15:25,560 and and cause immune response from happening. 258 00:15:25,560 --> 00:15:28,680 So we we actually have the technology to be able to place 259 00:15:28,680 --> 00:15:31,120 them exactly where we want them in three. 260 00:15:31,680 --> 00:15:34,240 Yeah, it was truly amazing to see the surface of the brain 261 00:15:34,240 --> 00:15:37,680 after the robot had inserted all the electrodes on the 1st 262 00:15:37,680 --> 00:15:39,640 participant without a drop of blood. 263 00:15:39,640 --> 00:15:42,560 Insight. It's really quite an 264 00:15:42,560 --> 00:15:45,400 achievement. Yes, something that probably 265 00:15:45,480 --> 00:15:48,760 most people don't realize is that the the brain appears to be 266 00:15:48,760 --> 00:15:50,680 sort of somewhat undifferentiated. 267 00:15:50,680 --> 00:15:53,400 So if you look at the cortex, it looks like a whole bunch of 268 00:15:53,400 --> 00:15:58,080 folds that where, you know, maybe like it's, it's not 269 00:15:58,160 --> 00:16:00,720 obvious, just looking at a say, a picture of the brain that, 270 00:16:01,560 --> 00:16:03,360 that, that it's the brain is highly differentiated. 271 00:16:03,360 --> 00:16:08,160 That there's you pretty much know exactly where the part of 272 00:16:08,160 --> 00:16:11,760 the brain is that controls your right hand and your left hand 273 00:16:12,080 --> 00:16:15,160 and your leg. And like that kind of thing or 274 00:16:15,160 --> 00:16:19,520 vision, it's actually quite precisely located. 275 00:16:19,520 --> 00:16:23,640 It's not some people like might, might think, look at the brain 276 00:16:23,640 --> 00:16:24,840 like, oh, it could be, it could be anywhere. 277 00:16:24,840 --> 00:16:30,720 But actually we it's your brain is highly differentiated, even 278 00:16:30,720 --> 00:16:33,800 though it doesn't look it's yeah. 279 00:16:34,520 --> 00:16:37,080 Do you want to describe how we actually where like how we 280 00:16:37,080 --> 00:16:38,680 identify where to drill the? Yeah. 281 00:16:38,680 --> 00:16:43,360 So we can we can put a patient that is considering this implant 282 00:16:44,000 --> 00:16:47,960 into an fMRI, so a functional magnetic resonance imaging 283 00:16:47,960 --> 00:16:51,760 machine and ask them to imagine hand movements that, you know, 284 00:16:51,760 --> 00:16:53,920 because of the spinal cord injury don't happen. 285 00:16:54,320 --> 00:16:57,360 But just imagining those hand movements causes these areas of 286 00:16:57,360 --> 00:17:00,400 the brain to light up in the fMRI scanner. 287 00:17:00,760 --> 00:17:05,960 And so we have a pretty good idea based in, in fact, for each 288 00:17:05,960 --> 00:17:09,040 individual participant, which part of their brain is going to, 289 00:17:10,640 --> 00:17:13,599 you know, respond to imagined movements of the hand. 290 00:17:14,079 --> 00:17:17,240 And so we can map those imagined movements, much as we all do 291 00:17:17,839 --> 00:17:20,560 when moving a mouse to controlling a cursor on a 292 00:17:20,560 --> 00:17:22,400 screen, even without the use of a mouse. 293 00:17:25,000 --> 00:17:26,240 Yeah. But anyway, I think it's kind of 294 00:17:26,240 --> 00:17:29,400 an important point that like, it's not like you're the part of 295 00:17:29,400 --> 00:17:31,360 your brain that controls your hand might be anywhere in the 296 00:17:31,360 --> 00:17:33,120 cortex. It's this is not the case. 297 00:17:33,120 --> 00:17:35,320 It's going to be in a very specific region and it's going 298 00:17:35,320 --> 00:17:37,680 to be extremely common across people. 299 00:17:38,040 --> 00:17:43,480 Precision is key too. Yeah, the left-handed, 300 00:17:43,480 --> 00:17:45,120 right-handed thing in my mind too. 301 00:17:45,120 --> 00:17:48,080 Like if you're right-handed, you want the device on the left 302 00:17:48,080 --> 00:17:50,200 side, the actual lateral side to the hand. 303 00:17:50,320 --> 00:17:52,280 That's atomic. Yeah, the left side of your 304 00:17:52,280 --> 00:17:53,440 brain controls right side of your body. 305 00:17:53,440 --> 00:17:57,360 Everything's crossed. Another of the risk mitigations 306 00:17:57,360 --> 00:18:01,360 we're looking at in the future is that, you know that the 307 00:18:01,360 --> 00:18:04,320 implant has a certain size, the depth of the bottom of the 308 00:18:04,320 --> 00:18:08,240 implant actually thinner than the average human skull. 309 00:18:08,640 --> 00:18:13,680 And So what we want to be able to do is control the size of the 310 00:18:13,680 --> 00:18:18,200 gap under the implant. If the threads that travel from 311 00:18:18,200 --> 00:18:21,720 the implant into the brain as much slack as possible. 312 00:18:22,600 --> 00:18:26,480 We didn't do this in the first participant because we didn't 313 00:18:26,480 --> 00:18:29,240 want to, you know, manipulate any of their tissue that we 314 00:18:29,240 --> 00:18:32,400 didn't absolutely have to in upcoming implants. 315 00:18:32,800 --> 00:18:37,200 Our plan is to or sculpt the surface of the skull very 316 00:18:37,200 --> 00:18:41,560 intentionally to minimize the gap under the implant such that 317 00:18:41,560 --> 00:18:45,120 the bottom of the implant travels perfectly flush with the 318 00:18:45,200 --> 00:18:48,040 normal contour of the inner side of the skull. 319 00:18:48,560 --> 00:18:52,440 That will put the implant closer to the brain, eliminate some of 320 00:18:52,440 --> 00:18:56,080 the tension on the threads and we think it will reduce some of 321 00:18:56,080 --> 00:18:58,600 the tendency of threads to retract, right and. 322 00:18:59,040 --> 00:19:00,640 We actually built a tool to do right. 323 00:19:01,720 --> 00:19:04,360 Yeah, this is, this is actually, this is a very important detail. 324 00:19:05,760 --> 00:19:11,000 You really want the the inner contour of the skull to be flush 325 00:19:11,040 --> 00:19:14,720 so that the implant doesn't there's no the brain doesn't 326 00:19:14,720 --> 00:19:16,680 want to pucker up into the into the gap. 327 00:19:16,800 --> 00:19:22,840 That's really quite a big deal. So that like like minimize the 328 00:19:22,840 --> 00:19:27,480 air pocket and the implant being flush with the the inside 329 00:19:27,480 --> 00:19:31,320 contour of the skull is are two very important improvements. 330 00:19:31,600 --> 00:19:35,080 The additional benefit here is that you know you do see some 331 00:19:35,080 --> 00:19:38,720 amount of stick up, what we call stick up, so you minor bump in 332 00:19:38,720 --> 00:19:41,040 the head, but this actually eliminates it even further. 333 00:19:41,080 --> 00:19:43,960 Yeah, yeah. I mean, it's like really our 334 00:19:43,960 --> 00:19:48,120 goal is that, that if you run your hand over the top of the 335 00:19:48,120 --> 00:19:50,600 skull, you don't feel any, any bump, you don't feel any, any 336 00:19:50,600 --> 00:19:53,560 device. And that even if someone was 337 00:19:53,560 --> 00:19:55,040 bald, you wouldn't really even notice it. 338 00:19:55,920 --> 00:19:59,640 And and then from the the inner inner contour of the skull that 339 00:19:59,640 --> 00:20:02,480 the the brain or physical standpoint doesn't really notice 340 00:20:02,480 --> 00:20:05,960 that there's a divot in the skull because there's no divot. 341 00:20:14,240 --> 00:20:17,480 OK. Another aspect of the of the 342 00:20:17,480 --> 00:20:20,200 human brain that you know, obviously differs from any of 343 00:20:20,200 --> 00:20:23,360 the animals that we tested in is that the human brain is a lot 344 00:20:23,360 --> 00:20:26,080 bigger. And so you may not realize that 345 00:20:26,080 --> 00:20:32,040 that means that the human brain moves quite a bit more than any 346 00:20:32,040 --> 00:20:34,120 of these other smaller brained creatures. 347 00:20:34,720 --> 00:20:40,000 And so when we open the skull, we see the brain travel toward 348 00:20:40,480 --> 00:20:43,760 and away from the robot about 3mm in total as the heart beats 349 00:20:43,760 --> 00:20:46,240 and and the breathing takes place. 350 00:20:46,720 --> 00:20:50,680 And so that movement, you know it, it adds a small challenge 351 00:20:50,680 --> 00:20:54,960 for the robot in precisely choosing a depth to insert each 352 00:20:54,960 --> 00:20:57,080 thread. It's not an enormous challenge. 353 00:20:57,080 --> 00:21:00,320 And we've already upgraded the robot's capabilities to be able 354 00:21:00,320 --> 00:21:06,400 to even more precisely target depth in, in even a very rapidly 355 00:21:06,400 --> 00:21:09,560 moving brain with a high amplitude of movement. 356 00:21:13,040 --> 00:21:15,800 You may think the most obvious mitigation for threads that 357 00:21:15,800 --> 00:21:17,640 pulled out of the brain is to insert them deeper. 358 00:21:17,920 --> 00:21:22,400 We think so too. And so we're going to broaden 359 00:21:22,400 --> 00:21:25,360 the range of depths at which we insert threads. 360 00:21:25,920 --> 00:21:29,640 So, you know, for the very first participant, we had an enormous 361 00:21:29,640 --> 00:21:33,960 amount of data from our animal work and we had very highly 362 00:21:33,960 --> 00:21:39,200 optimized our insertion depth to maximize the crossing of the 363 00:21:39,200 --> 00:21:42,440 layers of interest in the cortex with the electrodes that we're 364 00:21:42,440 --> 00:21:46,680 recording from. Now that we know retraction is a 365 00:21:46,680 --> 00:21:51,320 possibility, we're going to insert at a variety of depths 366 00:21:51,320 --> 00:21:55,200 that even in several cases of differing amounts of retracting 367 00:21:55,200 --> 00:21:58,640 threads, we're going to have electrodes at the proper depth 368 00:21:59,160 --> 00:22:03,440 and with the deepest threads be able to track how much 369 00:22:03,440 --> 00:22:07,440 retraction has occurred across the surface of the brain from 370 00:22:07,480 --> 00:22:09,600 from each thread. And so we're going to, you know, 371 00:22:09,640 --> 00:22:13,480 both have more threads in the right layer and have better data 372 00:22:13,480 --> 00:22:15,160 on how much retraction has occurred. 373 00:22:15,920 --> 00:22:18,520 If you're a BCI nerd, you might know that being able to control 374 00:22:18,520 --> 00:22:23,240 individual Z depth per thread is not something that most neural 375 00:22:23,240 --> 00:22:25,600 interface devices offer. Most neural interface devices 376 00:22:25,600 --> 00:22:28,840 are kind of a static fixed rigid array that you push in and all 377 00:22:28,840 --> 00:22:30,240 the electrodes are on depth. Right. 378 00:22:30,680 --> 00:22:32,960 To be able to do this is actually pretty pretty novel. 379 00:22:32,960 --> 00:22:36,120 Part of the robot. Yeah, the historical approach is 380 00:22:36,120 --> 00:22:39,680 to actually pound in a sort of bed of nails with an air hammer 381 00:22:39,680 --> 00:22:41,800 into the. Brain, it looks crazy that that 382 00:22:41,800 --> 00:22:46,400 that is, Yeah, just with a with a pneumatic hammer. 383 00:22:47,000 --> 00:22:49,480 That's the that's the this is it sounds come somewhat barbaric. 384 00:22:49,480 --> 00:22:52,920 This is not what we do, but this is the what's been done before 385 00:22:52,920 --> 00:22:55,000 is. They're literally just hammering 386 00:22:55,000 --> 00:22:56,920 in what looks like a bed of nails with the brain. 387 00:22:57,080 --> 00:22:58,280 Which? Actually works. 388 00:22:58,280 --> 00:22:59,560 It's astonishing that it actually. 389 00:22:59,760 --> 00:23:02,640 Works, but I mean some people like manual like DBS probes. 390 00:23:02,640 --> 00:23:05,480 You're just sticking in by hand. Our research is just guiding 391 00:23:05,480 --> 00:23:07,760 them in. Those are several, several 392 00:23:07,760 --> 00:23:10,800 orders of magnitude more volume of brain tissue that you're 393 00:23:10,800 --> 00:23:12,360 destroying compared to what we're doing. 394 00:23:12,520 --> 00:23:14,960 But that deep brain simulation stuff does actually work, and it 395 00:23:14,960 --> 00:23:18,440 actually helps people a lot. Yeah, yeah, yeah. 396 00:23:18,440 --> 00:23:19,920 That's a great product. Yeah. 397 00:23:20,360 --> 00:23:23,320 But I mean, I think we'll we'll be able to do a much more 398 00:23:23,320 --> 00:23:24,760 finessed version of that down the road. 399 00:23:26,520 --> 00:23:29,960 So I mean, it's really difficult. 400 00:23:30,400 --> 00:23:32,600 Like the the neural link device is something that really 401 00:23:32,680 --> 00:23:36,320 absolutely minimizes damage to the brain, absolutely minimizes 402 00:23:36,320 --> 00:23:40,160 the load on the patient. And the goal is to allow someone 403 00:23:40,160 --> 00:23:41,480 to live a completely normal life. 404 00:23:42,440 --> 00:23:45,200 They work that you won't even notice that someone even has the 405 00:23:45,200 --> 00:23:49,240 device. So like I said, we're restoring 406 00:23:49,520 --> 00:23:51,720 the ability to control your computer and phone. 407 00:23:51,840 --> 00:23:56,360 That's how telepathy and then next device being able to allow 408 00:23:56,360 --> 00:23:58,120 people to see that cannot see before. 409 00:23:58,520 --> 00:24:00,920 And in fact, you could allow people to see kind of like Dora 410 00:24:00,920 --> 00:24:03,800 the Forge in Star Trek in any but whatever. 411 00:24:03,960 --> 00:24:06,160 Infrared. Yeah, infrared, ultraviolet, 412 00:24:07,560 --> 00:24:10,880 radar. So. 413 00:24:12,760 --> 00:24:16,160 So I think another way of saying it is that we want to give 414 00:24:16,160 --> 00:24:19,440 people superpowers. So it's not just that we're 415 00:24:19,440 --> 00:24:22,680 restoring your prior brain functionality, but that you 416 00:24:22,680 --> 00:24:25,240 actually have functionality far greater than a normal human. 417 00:24:26,640 --> 00:24:29,240 That's a super big deal. And I also think, you know, 418 00:24:29,440 --> 00:24:31,680 often times the questions that we get a lot is why do you have 419 00:24:31,680 --> 00:24:34,480 to actually go into the brain? What if you place it on the 420 00:24:34,520 --> 00:24:38,800 surface or outside the skull? Basically the Long story short, 421 00:24:38,800 --> 00:24:41,760 the physics of how it works, you really need to get the sensors, 422 00:24:41,760 --> 00:24:46,920 which are these facing in the brain next to the source which 423 00:24:47,200 --> 00:24:49,120 you're on as close to it as possible. 424 00:24:49,120 --> 00:24:52,240 Otherwise what you get is you get a population response and 425 00:24:52,240 --> 00:24:55,000 not be able to kind of do the level of controls that we 426 00:24:55,000 --> 00:24:59,360 believe of. Yeah, I mean, a good sort of 427 00:24:59,960 --> 00:25:02,120 analogy would be like if you're trying to understand what goes 428 00:25:02,120 --> 00:25:04,560 on in a factory, you kind of need to go into the factory. 429 00:25:04,560 --> 00:25:07,920 You can't just put a stethoscope on the wall and try to figure 430 00:25:07,920 --> 00:25:09,720 out what's going. Like anything on the outside of 431 00:25:09,720 --> 00:25:13,560 the trying to read things from the outside is like putting a 432 00:25:13,560 --> 00:25:15,480 stethoscope on the wall of a factory, trying to understand 433 00:25:15,480 --> 00:25:18,040 what's going in the factory. It's not going to be effective. 434 00:25:19,760 --> 00:25:22,160 You've got to be threads. You got to be in there. 435 00:25:23,600 --> 00:25:26,520 So. But I just want to be emphasized 436 00:25:26,520 --> 00:25:31,000 again, like the goal is to give people superpowers, not just to 437 00:25:31,000 --> 00:25:34,760 restore prior functionality. So it's very exciting. 438 00:25:35,760 --> 00:25:39,280 And I think that should give hope to a lot of people in the 439 00:25:39,280 --> 00:25:43,280 world that the future is going to be exciting and inspiring and 440 00:25:43,760 --> 00:25:46,800 the technology is going to give them superpowers. 441 00:25:46,800 --> 00:25:58,440 I mean, that's that's amazing. Yeah, I guess. 442 00:25:59,160 --> 00:26:01,600 These. Off yeah and could can you 443 00:26:01,600 --> 00:26:04,800 multitask but yeah in fact if you look at Nolan's streaming 444 00:26:04,800 --> 00:26:08,480 and you can just check out Nolan's streams on on the X 445 00:26:08,480 --> 00:26:12,360 platform he's multitasking all the time so he's playing video 446 00:26:12,360 --> 00:26:15,880 games while talking and. Listening to podcasts? 447 00:26:15,880 --> 00:26:18,280 Nice listening to podcasting. Yeah. 448 00:26:18,320 --> 00:26:21,600 Yeah, exactly. So it's really just like if you, 449 00:26:21,800 --> 00:26:24,880 you're using your hands and you, you can be, you know, playing a 450 00:26:24,880 --> 00:26:27,240 video game while talking so. I mean, don't take that word for 451 00:26:27,240 --> 00:26:27,960 it. Just go watch. 452 00:26:27,960 --> 00:26:29,920 I mean, yeah, yeah, he's out there on the Internet doing his 453 00:26:29,920 --> 00:26:31,040 thing. Yeah, yeah, exactly. 454 00:26:33,320 --> 00:26:35,640 So can he do keyboard shortcuts or is it just a mouse? 455 00:26:35,640 --> 00:26:36,920 Yeah. That that's actually what we're 456 00:26:36,920 --> 00:26:40,400 working on right now. Sure, so currently he's working 457 00:26:40,400 --> 00:26:42,920 with the mouse, but we are also exploring and decoding more 458 00:26:43,160 --> 00:26:45,240 dimensions from the neural activity. 459 00:26:45,680 --> 00:26:50,520 Multiple clicks. So to do shortcuts or just able 460 00:26:50,520 --> 00:26:56,240 to control more games, control games with an Xbox controller. 461 00:26:57,080 --> 00:26:59,520 But also in the future we expand, we plan to expand to 462 00:26:59,520 --> 00:27:03,480 decode text, not just the mouse control, but also allow our 463 00:27:03,480 --> 00:27:08,720 participant to type much faster. And yeah. 464 00:27:09,320 --> 00:27:11,080 Yeah, actually. So maybe going back to the 465 00:27:11,080 --> 00:27:13,160 discussion of thread retraction, you know, one of the very 466 00:27:13,160 --> 00:27:15,640 exciting parts to me about this story is that we're able to do 467 00:27:15,640 --> 00:27:18,600 so much with 15% of channels. You have more channels. 468 00:27:18,600 --> 00:27:21,160 What that actually offers you is not just faster mouse control, 469 00:27:21,480 --> 00:27:23,880 because in the motor cortex, neurons don't all represent the 470 00:27:23,880 --> 00:27:25,480 same thing. So if you're trying to 471 00:27:25,480 --> 00:27:28,480 understand, like, you know, what an individual finger is trying 472 00:27:28,480 --> 00:27:31,360 to do, you might or might not have an Electro next to it. 473 00:27:31,360 --> 00:27:32,960 And the more channels you have in the brain, the higher 474 00:27:32,960 --> 00:27:35,880 likelihood you have, you know, representation or decodability 475 00:27:35,880 --> 00:27:38,440 of all fingers on the hand. It's like you're trying to do 476 00:27:38,440 --> 00:27:40,080 something like output text at a fast rate. 477 00:27:40,080 --> 00:27:41,920 It's something that matters a lot for people who are 478 00:27:41,920 --> 00:27:44,680 completely locked in, who cannot speak at all, who are trying to, 479 00:27:44,680 --> 00:27:46,920 you know, just say I love you to them, to a loved one in their 480 00:27:46,920 --> 00:27:49,160 family or ask for a glass of water or a scratch or whatever. 481 00:27:49,160 --> 00:27:51,280 You know, being able to type at a faster rate, it's extremely 482 00:27:51,280 --> 00:27:52,960 important. And the more fingers you have 483 00:27:52,960 --> 00:27:55,360 access to, higher probability, you can do that efficiently. 484 00:27:56,360 --> 00:27:59,760 And so, yeah, you know, I'm super excited about how high the 485 00:27:59,760 --> 00:28:02,760 ceiling is we can that we can get to as we resolve this 486 00:28:02,760 --> 00:28:05,040 standard traction issue. Yeah, I mean we're like we're 487 00:28:05,040 --> 00:28:08,160 currently at approximately 1010 bits per second be great. 488 00:28:08,160 --> 00:28:13,480 But ultimately we want to get to megabit and I think say 489 00:28:13,640 --> 00:28:15,080 ultimately whole brain interface. 490 00:28:15,400 --> 00:28:19,280 I think you know, many years from now I think Gigabit level 491 00:28:19,280 --> 00:28:23,640 is possible. So that's that's pretty 492 00:28:23,640 --> 00:28:27,600 astonishing. Now you know with this is still 493 00:28:27,600 --> 00:28:30,680 version one of our device, as we mentioned, it's version one with 494 00:28:30,680 --> 00:28:36,480 only 15% of the threads working. The the current device has 64 495 00:28:36,480 --> 00:28:38,960 threads with 16 electrodes on each thread. 496 00:28:39,640 --> 00:28:44,760 Our next device has 128 threads with, with eight electrodes per 497 00:28:44,880 --> 00:28:47,880 per thread. Because as we get more confident 498 00:28:47,880 --> 00:28:52,560 about how, where exactly to place the the electrode the the 499 00:28:52,560 --> 00:28:55,200 thread, you need fewer electrodes per thread. 500 00:28:55,480 --> 00:28:57,840 So we can essentially with the current device without 501 00:28:57,840 --> 00:29:02,320 substantial changes potentially double the bandwidth if we 502 00:29:03,000 --> 00:29:06,720 accurate with the with the placement of of the threads and 503 00:29:06,720 --> 00:29:09,280 then our next generation device will have maybe even. 504 00:29:09,280 --> 00:29:13,920 More channels, Yeah, yeah. So yeah, so next device for 505 00:29:13,920 --> 00:29:18,480 every four, yeah, 3000 channels. So this will just keep getting 506 00:29:18,480 --> 00:29:21,960 better and better really moving up I think in orders of 507 00:29:21,960 --> 00:29:25,400 magnitude in factors of 10 basically in in what kind of 508 00:29:25,400 --> 00:29:29,720 bandwidth. So I think it won't be, it won't 509 00:29:29,720 --> 00:29:33,200 be all that long before someone with a neural link device can 510 00:29:33,200 --> 00:29:36,720 communicate faster than someone who is has a fully functional 511 00:29:36,920 --> 00:29:41,960 body. And yeah, so I think, you know, 512 00:29:41,960 --> 00:29:44,360 faster than the fastest speed typist or auctioneer. 513 00:29:45,120 --> 00:29:48,080 The E sports tournaments are. Going to be won't be able to 514 00:29:48,080 --> 00:29:52,000 speak faster than someone can communicate with a neural link 515 00:29:52,000 --> 00:29:54,960 telepathy device. It may be a very interesting 516 00:29:54,960 --> 00:29:58,320 part of this. Basically, we currently connect 517 00:29:58,600 --> 00:30:02,040 standard inputs to the computer through mouse and keyboards, but 518 00:30:02,040 --> 00:30:05,960 very soon as we will have a much part of bandwidth, we need to 519 00:30:05,960 --> 00:30:09,760 think about new ways to actually build the interface for the 520 00:30:09,840 --> 00:30:12,280 devices. This is something that we can't 521 00:30:13,000 --> 00:30:15,640 accept. Yeah, no, that's, that's a good 522 00:30:15,640 --> 00:30:20,160 point because the the current import devices are centered 523 00:30:20,160 --> 00:30:21,440 around human hands. Yeah. 524 00:30:21,720 --> 00:30:24,320 So it's like we've got these, you know, little meat sticks 525 00:30:24,320 --> 00:30:29,360 that we move and the this certain rate at which you can 526 00:30:29,360 --> 00:30:31,120 move your little move your fingers. 527 00:30:31,120 --> 00:30:33,880 And, and so we've got like the mouse and the keyboard and with 528 00:30:33,880 --> 00:30:36,240 a joystick control, you know, like Xbox controller or 529 00:30:36,240 --> 00:30:38,720 something like that. But you really don't need that. 530 00:30:39,200 --> 00:30:43,680 You can actually, you don't, you don't need, since you're no 531 00:30:43,680 --> 00:30:46,280 longer, if you're not trying to use your hands, you don't, you 532 00:30:46,280 --> 00:30:51,880 actually don't need those conventional control mechanisms. 533 00:30:53,320 --> 00:30:55,880 And so This is why like, ultimately I think you'll be 534 00:30:55,880 --> 00:31:01,200 able to do conceptual telepathy like where you can communicate 535 00:31:01,200 --> 00:31:05,800 entire concepts uncompressed to someone else with a neural link 536 00:31:06,040 --> 00:31:09,480 or to the computer. Even today we have some problems 537 00:31:09,480 --> 00:31:11,280 here where like, you know, if you don't feel the mouse 538 00:31:11,280 --> 00:31:13,280 clicking under your finger, how do you know it actually 539 00:31:13,280 --> 00:31:14,720 happened? Because you know, you're, you're 540 00:31:14,720 --> 00:31:16,440 seeing it on the screen, but you don't actually feel the mouse 541 00:31:16,440 --> 00:31:17,080 clip. You don't have the 542 00:31:17,080 --> 00:31:19,200 proprioceptive feedback of, you know, the keys under your 543 00:31:19,200 --> 00:31:20,600 fingertips or the track pad under your. 544 00:31:21,240 --> 00:31:24,040 There's all sorts of interesting UX challenges, how to actually 545 00:31:24,480 --> 00:31:27,320 give the user some sense of what their decoder's actually doing, 546 00:31:27,720 --> 00:31:29,000 or what the Neurlink's actually doing. 547 00:31:29,000 --> 00:31:35,320 I mean, they're trying to use. So wireless. 548 00:31:38,200 --> 00:31:41,000 Yeah, it's Bluetooth. Just a Bluetooth connection, 549 00:31:41,000 --> 00:31:44,440 just like how your normal Apple mouse or like Apple Magic 550 00:31:44,440 --> 00:31:45,880 Keyboard connects to your computer. 551 00:31:45,880 --> 00:31:49,480 Same exact thing. In fact in Yeah, we can 552 00:31:49,480 --> 00:31:52,040 basically have this exposed as an HID interface if we want. 553 00:31:52,040 --> 00:31:54,800 HID is just the name of the protocol for like sending fits 554 00:31:54,800 --> 00:31:58,680 my mouse into a computer. Yeah, I can plug into basically 555 00:31:58,680 --> 00:32:00,080 anything. Yeah. 556 00:32:00,240 --> 00:32:02,600 I mean, I think we, we chose that interface because it's 557 00:32:02,720 --> 00:32:05,360 ubiquitous. Yeah, basically any devices are, 558 00:32:05,360 --> 00:32:09,160 are have Bluetooth capabilities. Our our long term goal is to 559 00:32:09,160 --> 00:32:11,360 actually have our own protocol, you know, that is safe and 560 00:32:11,360 --> 00:32:13,280 secure. But for now, you know we've 561 00:32:13,280 --> 00:32:18,080 chosen it for interoperability. So the question is, can a neural 562 00:32:18,080 --> 00:32:20,560 link chip repair the paralysis in the long term? 563 00:32:21,720 --> 00:32:23,440 You know, we can't do that right now. 564 00:32:23,440 --> 00:32:27,360 We have done sort of preliminary work implanting a second neural 565 00:32:27,360 --> 00:32:34,320 link in the spinal cord and we can restore naturalistic looking 566 00:32:34,560 --> 00:32:37,440 and, and leg movements in animal models. 567 00:32:39,000 --> 00:32:42,040 But this isn't something that is, you know, don't, don't hold 568 00:32:42,040 --> 00:32:43,800 your breath waiting for it. It's going to be a while. 569 00:32:43,800 --> 00:32:48,080 We've got a lot of work to do. But yes, there's no reason in 570 00:32:48,080 --> 00:32:50,160 theory that we can't repair paralysis. 571 00:32:51,640 --> 00:32:56,960 Yeah. I mean essentially to, to, I 572 00:32:56,960 --> 00:33:00,000 mean, it there's, there's no, there's no physics barrier to 573 00:33:00,440 --> 00:33:03,560 fully solving paralysis. That is perhaps a way to say it 574 00:33:03,560 --> 00:33:09,440 that you've got signals coming from your motor cortex that if 575 00:33:09,440 --> 00:33:14,240 they are transferred past the point where the the nerves are 576 00:33:14,240 --> 00:33:17,480 damaged, essentially just it's basically a communications 577 00:33:17,480 --> 00:33:19,560 bridge. So you bridge the communications 578 00:33:19,560 --> 00:33:24,000 from the motor cortex past the the point in the neck or spine 579 00:33:24,000 --> 00:33:28,200 where the nose is damaged and you should like it is physic it 580 00:33:28,200 --> 00:33:31,320 is possible from a physics standpoint to restore full body 581 00:33:31,320 --> 00:33:34,520 functionality from a physics standpoint, it's a very hard 582 00:33:34,520 --> 00:33:38,720 technical problem, but it but it there is nothing that prevents 583 00:33:38,720 --> 00:33:39,960 it happening from a physics standpoint. 584 00:33:45,200 --> 00:33:49,560 So in terms of next phase of the rollout, well, we really want to 585 00:33:49,560 --> 00:33:52,800 make sure that we make as much progress as possible between 586 00:33:52,800 --> 00:33:55,840 each neural link patient. So this is we're only just 587 00:33:55,840 --> 00:34:00,400 moving now to our second neural link patient, but we we hope to 588 00:34:00,400 --> 00:34:04,680 have, you know, if things go well high single digits this 589 00:34:04,680 --> 00:34:09,239 year. And I don't know maybe if this 590 00:34:09,239 --> 00:34:12,880 is somewhat dependent on regulatory approval and how much 591 00:34:12,880 --> 00:34:16,920 technical progress we make, but within a few years, hopefully 592 00:34:16,920 --> 00:34:18,639 thousands. Yeah. 593 00:34:18,639 --> 00:34:20,960 And I think one thing that is important to highlight is that, 594 00:34:22,320 --> 00:34:24,960 you know, it's not that we've built only one device and one 595 00:34:24,960 --> 00:34:26,679 surgery. We've done hundreds of surgery. 596 00:34:26,679 --> 00:34:30,080 We've built thousands and thousands of devices even for 597 00:34:30,080 --> 00:34:33,560 just the the ability to unearth any sort of low frequency 598 00:34:33,719 --> 00:34:36,199 failure mode. So we have already been 599 00:34:36,199 --> 00:34:39,000 investing very heavily in infrastructure to be able to 600 00:34:39,000 --> 00:34:42,239 scale this thing on the device manufacturing side as well as on 601 00:34:42,239 --> 00:34:44,760 the surgery side of things. We want to be able to help as 602 00:34:44,760 --> 00:34:47,840 many people as quickly as well, yeah, go through obviously the 603 00:34:47,840 --> 00:34:51,000 appropriate hurdles that are regulatory challenges and 604 00:34:51,000 --> 00:34:56,040 proving out the device with. Yeah, and the, the, the device 605 00:34:56,040 --> 00:34:59,840 implantation really needs to become almost entirely if not 606 00:34:59,840 --> 00:35:03,800 entirely automatic in the same way that, say, LASIK eye surgery 607 00:35:03,800 --> 00:35:06,280 is done. You know, you don't have an 608 00:35:06,280 --> 00:35:09,400 ophthalmologist with a, a laser cutter by hand that that would 609 00:35:09,400 --> 00:35:13,240 be crazy. But the ophthalmologist overseas 610 00:35:13,240 --> 00:35:16,840 the, the LASIK machine and make sure that the settings are 611 00:35:16,840 --> 00:35:19,840 correct and then the machine does everything and restores 612 00:35:19,840 --> 00:35:22,400 your eyesight. It's really remarkable how how 613 00:35:22,400 --> 00:35:25,400 many people have had their eyesight restored with, with 614 00:35:25,400 --> 00:35:26,640 LASIK. And I think there's another one 615 00:35:26,640 --> 00:35:28,480 called Smile. It's they, they keep making it 616 00:35:28,480 --> 00:35:30,080 better. We need to have something 617 00:35:30,080 --> 00:35:33,240 similar for a Neuralink implantation so that you 618 00:35:33,240 --> 00:35:36,480 basically sit down and whatever the, the, whatever, whatever 619 00:35:36,520 --> 00:35:39,480 kind of upgrades or, you know, brain fixes are needed. 620 00:35:40,600 --> 00:35:43,600 That's that's reviewed by medical expert. 621 00:35:44,320 --> 00:35:47,960 Obviously we want to make sure that that is reviewed correctly, 622 00:35:47,960 --> 00:35:50,160 but but it really needs to be automatic. 623 00:35:50,240 --> 00:35:54,560 So you sit down and and within 10 minutes you have a Neuralink 624 00:35:54,560 --> 00:35:57,360 device installed very, very fast. 625 00:35:57,360 --> 00:36:00,520 I mean, it's very sort of cyberpunk, you know, Deus Ex if 626 00:36:00,520 --> 00:36:06,400 you played those games and. We'll Neuralink start to 627 00:36:06,400 --> 00:36:08,360 interface with other devices like wheelchair. 628 00:36:08,840 --> 00:36:10,840 It's great question. We currently focusing on 629 00:36:12,520 --> 00:36:17,160 controlling computers and unlock independence in the virtual 630 00:36:17,160 --> 00:36:19,440 world. Of course, our plan is as we 631 00:36:19,440 --> 00:36:21,640 mentioned earlier, robotic arm and a wheelchair. 632 00:36:21,640 --> 00:36:25,360 To unlock independence in the physical world is of course at 633 00:36:25,440 --> 00:36:27,760 the additional risk if you make your computer. 634 00:36:27,760 --> 00:36:31,960 There's some yeah to that, but we are working with the FDA to 635 00:36:32,080 --> 00:36:35,000 allow us to be quite some. Well, it seems like if if the 636 00:36:35,200 --> 00:36:37,280 wheelchair has a. An app. 637 00:36:37,880 --> 00:36:39,560 Well, the wheelchair just need to have some have an interface. 638 00:36:39,560 --> 00:36:42,600 It does, yeah. So wheelchair has a Bluetooth 639 00:36:42,600 --> 00:36:44,520 interface. You could just Bluetooth 640 00:36:44,520 --> 00:36:49,480 interface to the wheelchair and and that's probably something we 641 00:36:49,480 --> 00:36:51,600 should do. We're working on pretty soon, 642 00:36:51,600 --> 00:36:52,560 yeah. It's really a matter of 643 00:36:52,600 --> 00:36:54,080 paperwork. I'm showing that you can do it 644 00:36:54,080 --> 00:36:55,360 safely. You don't want to drive off. 645 00:36:55,400 --> 00:36:58,120 A Cliff, Well, I think we could. Well, we can limit the speed. 646 00:36:58,120 --> 00:36:58,800 Yeah. Yeah, exactly. 647 00:36:58,800 --> 00:37:03,480 So it doesn't go careening often to disaster, but you know, so 648 00:37:03,600 --> 00:37:08,400 just make it go slowly at first. But yeah, so being able to sort 649 00:37:08,400 --> 00:37:12,720 of it really the New York device just should work generally for 650 00:37:12,720 --> 00:37:17,080 anything that's got a Bluetooth. Interface, including potentially 651 00:37:17,080 --> 00:37:23,000 an Optimus. Yes, yeah, you yes, you could 652 00:37:23,000 --> 00:37:29,400 communicate with Optimus. Yep, absolutely Optimus. 653 00:37:29,400 --> 00:37:31,520 We'll also be able to talk to Optimus. 654 00:37:31,520 --> 00:37:37,200 But like, but you could just, yeah, instead of talking just, 655 00:37:37,200 --> 00:37:40,000 you could just beam it directly. Or if, if someone has lost the 656 00:37:40,000 --> 00:37:42,760 use of speech, then then they can still communicate to an 657 00:37:42,760 --> 00:37:45,480 Optimist. They, they can communicate 658 00:37:45,480 --> 00:37:48,080 telepathically to Optimists or by Bluetooth. 659 00:37:49,600 --> 00:37:54,520 And, and, and so even if someone has, you know, completely less 660 00:37:54,520 --> 00:37:57,760 the ability to speak, they could still control Optimist or their 661 00:37:57,760 --> 00:37:58,560 computer or phone. I. 662 00:37:59,040 --> 00:38:00,960 Mean also like if you have an optimist and you have a neuron, 663 00:38:01,000 --> 00:38:03,600 you can just directly map the brain signal to control of the 664 00:38:03,600 --> 00:38:05,920 physical arm of the robot. And that's a very meaningful 665 00:38:05,920 --> 00:38:07,520 thing. Like if you're, you know, folks 666 00:38:07,520 --> 00:38:09,520 that have spinal cord injury, one of the biggest requests is 667 00:38:09,520 --> 00:38:12,720 to be able to scratch yourself. It's something that quite 668 00:38:12,720 --> 00:38:15,320 annoying actually. And if you have a scratch on 669 00:38:15,320 --> 00:38:17,360 your face, you can't fall asleep until you scratch it. 670 00:38:18,120 --> 00:38:19,960 You know, it's very convenient to be able to move something 671 00:38:19,960 --> 00:38:22,120 physically towards you, to be able to scratch similar things 672 00:38:22,120 --> 00:38:23,520 like eating food. You know, if you need somebody 673 00:38:23,520 --> 00:38:26,880 to feed you very hard to have dinner with friends in a way 674 00:38:26,880 --> 00:38:30,280 that is, you know, sort of a normal social experience. 675 00:38:30,280 --> 00:38:33,440 And so if you can feed yourself, pick up a fork and actually eat 676 00:38:33,440 --> 00:38:36,320 a piece of chicken on your own, you know, that's a big deal. 677 00:38:37,320 --> 00:38:40,520 It prevents and saves a lot of interactions with caretakers and 678 00:38:40,520 --> 00:38:43,040 other people in your life that you rely on to take care of it, 679 00:38:43,400 --> 00:38:47,320 But it really increases you. And I think an exciting 680 00:38:47,320 --> 00:38:50,960 possibility long term also is to say if you take parts of the 681 00:38:50,960 --> 00:38:53,400 Optimus Optimus humanoid robot and you combine that with a 682 00:38:53,400 --> 00:38:56,560 neural link, let's say somebody has lost their arms or legs. 683 00:38:57,720 --> 00:39:00,920 Well, we, we could actually attach an Optimus arm or Optimus 684 00:39:00,920 --> 00:39:05,840 legs and do a neural link implant so that the, the motor 685 00:39:05,840 --> 00:39:09,200 commands from your brain that go would go to your biological 686 00:39:09,200 --> 00:39:10,800 arms. Now go to your robot arms or 687 00:39:10,800 --> 00:39:13,480 robot legs. And again, you you'd have 688 00:39:13,480 --> 00:39:15,440 basically cybernetic superpowers. 689 00:39:15,840 --> 00:39:18,280 Actually, so the latency from the neural link to your hand 690 00:39:18,440 --> 00:39:21,320 would probably be slightly faster than it is just to go to 691 00:39:21,320 --> 00:39:23,200 your physical hand. So you can imagine like if 692 00:39:23,200 --> 00:39:25,720 you're a piano player or AI don't know anything that 693 00:39:25,720 --> 00:39:28,360 requires extremely fast hand movements, that you could 694 00:39:28,360 --> 00:39:30,760 actually have a pretty imbalanced right hand robotic 695 00:39:30,760 --> 00:39:32,680 arm control versus left hand physical arm control. 696 00:39:33,160 --> 00:39:36,440 There's more. Yeah, like I said, it's just 697 00:39:36,440 --> 00:39:40,000 kind of a cyberpunk Deus Ex in the future where you have 698 00:39:40,000 --> 00:39:43,080 cybernetic upgrades that are actually better than your 699 00:39:43,240 --> 00:39:51,400 biological limbs. And it's, it's certainly the, 700 00:39:51,800 --> 00:39:55,520 we'll have a much, you know, as as particularly as we expand to 701 00:39:55,520 --> 00:40:01,280 a large number of of, of customers or patients for neural 702 00:40:01,280 --> 00:40:04,280 link, the understanding of the brain will improve dramatically 703 00:40:04,920 --> 00:40:09,880 because really there isn't a fine, very fine grained 704 00:40:09,880 --> 00:40:12,640 understanding of the brain today because the, it's just the 705 00:40:12,640 --> 00:40:16,080 sensors aren't good enough. You've got fMRI, which is pretty 706 00:40:16,080 --> 00:40:19,320 good, but it's still not as good as actually having high 707 00:40:19,320 --> 00:40:20,440 bandwidth electrodes in the brain. 708 00:40:21,160 --> 00:40:24,000 Yeah, I think this is under appreciated as a research tool 709 00:40:24,000 --> 00:40:27,600 to to move that whole effort forward of really knowing, you 710 00:40:27,600 --> 00:40:32,000 know, what the physical substance of human thought is. 711 00:40:32,200 --> 00:40:35,320 We don't know to the to the degree that we need to. 712 00:40:35,560 --> 00:40:39,480 So Neuralink is actually a very powerful research tool. 713 00:40:40,800 --> 00:40:46,200 Yeah, I mean, we, I think we can ultimately understand and and 714 00:40:46,200 --> 00:40:51,280 fix it's quite severe psychosis or like if somebody's got like 715 00:40:51,280 --> 00:40:56,280 the if somebody's got like a. Like a delusion that they have a 716 00:40:56,280 --> 00:40:56,960 chip in their brain. Yeah. 717 00:40:58,360 --> 00:41:00,240 I was wondering if you were going to mention that one. 718 00:41:00,640 --> 00:41:03,440 We just want to be clear. There's only one person with a 719 00:41:03,440 --> 00:41:07,160 new linked chip in their brain. So for people out there who 720 00:41:07,160 --> 00:41:09,720 think we've put a chip in their brain, we'd like to assure you 721 00:41:10,480 --> 00:41:13,800 or what it's worth, you probably won't believe us, but we did not 722 00:41:13,800 --> 00:41:20,760 put a chip in your brain. OK, So there's actually a 723 00:41:20,760 --> 00:41:23,080 remarkable number of people who think we have put a chip in 724 00:41:23,080 --> 00:41:26,760 their brain, but we have not. But in the future, if you would 725 00:41:26,760 --> 00:41:30,320 like us to put a chip in your brain, which will perhaps help 726 00:41:30,320 --> 00:41:32,840 with the issue of thinking that you have a chip in your brain, 727 00:41:33,600 --> 00:41:37,520 then we will be able to do so. So there are people that have 728 00:41:37,920 --> 00:41:40,560 severe schizophrenia. They've got basically things 729 00:41:40,560 --> 00:41:43,200 that their brain is malfunctioning in some way. 730 00:41:43,600 --> 00:41:49,240 And, and this is actually due to really like physical circuitry 731 00:41:49,240 --> 00:41:50,360 issues. You can think of the brain as 732 00:41:50,400 --> 00:41:54,960 like really it's a, it's a biological computer. 733 00:41:54,960 --> 00:41:59,480 And if some of the circuits are crossed, it's going to, you 734 00:41:59,480 --> 00:42:02,160 know, it's going to crash or it's going to have issues that 735 00:42:02,680 --> 00:42:06,280 was not work. But with a Neuralink device, we 736 00:42:06,280 --> 00:42:11,280 can fix those issues and, you know, give someone who I think 737 00:42:11,640 --> 00:42:16,200 has to say severe schizophrenia or psychosis of some kind, allow 738 00:42:16,200 --> 00:42:19,320 them to live a normal life. I think that is one of the 739 00:42:19,320 --> 00:42:31,440 likely things in the future. So, yeah, I mean, yeah, you can 740 00:42:31,440 --> 00:42:34,400 certainly imagine, like, I'm sure people have, like, parents, 741 00:42:34,400 --> 00:42:41,040 grandparents who've, you know, have memory that's not working 742 00:42:41,040 --> 00:42:44,720 as well as it used to be. Sometimes they forget who who 743 00:42:44,720 --> 00:42:46,960 their grandchildren are or what day it is. 744 00:42:46,960 --> 00:42:49,360 And this is something that on your linked device could help 745 00:42:49,360 --> 00:42:52,840 fix I. Mean that that's actually one of 746 00:42:52,840 --> 00:43:02,080 the personal reason in in in many way like forms of you're 747 00:43:02,080 --> 00:43:04,880 you're literally losing your and then part of your right then, 748 00:43:04,880 --> 00:43:09,000 which is a just a very, very. Yeah. 749 00:43:11,280 --> 00:43:13,040 And it's really just, it's a glitch in the biological 750 00:43:13,040 --> 00:43:18,240 computer that is a fixable glitch like, like it's a short 751 00:43:18,240 --> 00:43:22,640 circuit essentially. How does the device charge and 752 00:43:22,640 --> 00:43:24,440 how long does the charge last? Yeah. 753 00:43:24,440 --> 00:43:27,560 So the current version that Nolan has, it lasts for four to 754 00:43:27,560 --> 00:43:31,200 five hours on a single charge, and it takes about 45 minutes to 755 00:43:31,200 --> 00:43:32,920 charge. The thing we've learned from 756 00:43:32,920 --> 00:43:35,320 Nolan is that that's actually one of the main limiters for him 757 00:43:35,320 --> 00:43:37,680 using it more. It's actually pretty hard to use 758 00:43:37,680 --> 00:43:40,360 a product more than like 70 hours a week. 759 00:43:40,520 --> 00:43:42,960 But that's about what he has he used it for in some weeks. 760 00:43:43,200 --> 00:43:45,400 Yeah, 70 hours in a week, Yeah. I mean, just for context, like 761 00:43:45,400 --> 00:43:47,000 you sleep roughly 8 hours a night. 762 00:43:47,000 --> 00:43:48,880 So that's, you know, we're doing better than the bed. 763 00:43:49,120 --> 00:43:51,200 The bed is 56 hours a week if you use roughly. 764 00:43:51,800 --> 00:43:55,120 And so 70 hours a week if uses. I challenge you to think about 765 00:43:55,120 --> 00:43:56,760 products that you've actually used for that duration. 766 00:43:57,920 --> 00:43:59,640 But that's OK. Some of these points are worth 767 00:43:59,640 --> 00:44:03,240 like emphasizing again, like the that Nolan, our first neural 768 00:44:03,240 --> 00:44:08,120 link recipient, has used the neural link device for 70 hours 769 00:44:08,120 --> 00:44:11,120 in a week, which is incredible. You probably won't enjoy the 770 00:44:11,120 --> 00:44:15,840 time sharing his computer use publicly, but I mean, I assure 771 00:44:15,840 --> 00:44:19,400 use for productive things only. But actually, so one of the 772 00:44:19,400 --> 00:44:21,120 things we've learned is that in the next version of the device, 773 00:44:21,120 --> 00:44:23,520 we really need to like double or you know, increase that battery 774 00:44:23,520 --> 00:44:25,040 life. And so I think DJ, the next 775 00:44:25,040 --> 00:44:28,720 version is going to be double. Actually actually double without 776 00:44:28,720 --> 00:44:32,440 without increasing the charge. Correct in charging time to 777 00:44:32,440 --> 00:44:34,600 double the battery life, meaning you should get roughly 8 hours 778 00:44:34,600 --> 00:44:36,160 of use. And the goal is to actually get 779 00:44:36,200 --> 00:44:39,240 to all they use so you can just charge, you know, maybe in your 780 00:44:39,240 --> 00:44:41,760 sleep your. Sleeping pillow exactly as soon 781 00:44:41,760 --> 00:44:44,720 as you've got like 16 hours of usage, then you basically have 782 00:44:44,720 --> 00:44:47,000 24 hours of usage because it can charge while you're sleeping. 783 00:44:47,680 --> 00:44:49,240 One of the things that's important, I think, to call out 784 00:44:49,240 --> 00:44:51,800 here is if you're paralyzed, you can't, you know, put the charger 785 00:44:51,800 --> 00:44:53,600 over your head yourself. And so it's important to think 786 00:44:53,600 --> 00:44:56,880 about like it's not just a duration of better use, but also 787 00:44:56,880 --> 00:44:58,680 can you recharge it yourself independently. 788 00:44:59,200 --> 00:45:00,960 So we spend a lot of time thinking about how to make that 789 00:45:01,080 --> 00:45:03,480 feasible because then that means that you can, this is what no 790 00:45:03,480 --> 00:45:05,520 one does. You can use the device, charge 791 00:45:05,520 --> 00:45:08,400 it, use the device, charge it, use the device without needing 792 00:45:08,400 --> 00:45:11,120 anybody to come in and sort of help you with that, which is a 793 00:45:11,120 --> 00:45:13,640 big deal if you're trying to play save until 5:00 AM at night 794 00:45:13,640 --> 00:45:16,520 when your family's asleep. And the way in which he does 795 00:45:16,520 --> 00:45:19,000 that is that there is a charger coil that's a big or you know 796 00:45:19,320 --> 00:45:22,840 about this big. And we actually put it in the 797 00:45:23,240 --> 00:45:26,800 sleeve of a, of a hat, a little beanie or a beanie. 798 00:45:26,800 --> 00:45:29,880 And then he wears it and then says with the voice command and 799 00:45:29,880 --> 00:45:31,200 charge. Charger energized. 800 00:45:31,200 --> 00:45:40,410 That's the one he likes, yeah. How would writing work? 801 00:45:41,210 --> 00:45:48,650 So so yeah, the current device that Nolan has is reading. 802 00:45:49,370 --> 00:45:52,690 So it's trying to read his essentially like worst movement 803 00:45:52,970 --> 00:45:56,850 from from one one hand. That's also, you know, with 804 00:45:56,890 --> 00:46:00,050 pointing out like in the future, like I think we're pretty cool 805 00:46:00,050 --> 00:46:03,600 to you can build into a second implant that would allow the 806 00:46:03,640 --> 00:46:07,800 other hand to be used and also have higher, obviously higher 807 00:46:07,800 --> 00:46:09,960 active electrode count. So then you can play 2 808 00:46:10,040 --> 00:46:12,040 essentially play games 200 because that's normally how you 809 00:46:12,040 --> 00:46:16,080 play games. And but then with with writing, 810 00:46:16,320 --> 00:46:19,160 it's really just it's an electrical impulse instead of 811 00:46:19,160 --> 00:46:22,320 like reading electrical impulses from the neuron to you issue an 812 00:46:22,320 --> 00:46:25,360 electrical impulse, which is obviously critical for vision. 813 00:46:25,360 --> 00:46:30,080 So vision is, is writing, which is just training and electrical 814 00:46:30,080 --> 00:46:32,280 impulse in the vision part of the brain. 815 00:46:33,720 --> 00:46:37,200 And that like activates a pixel. So we, we actually do have this 816 00:46:37,200 --> 00:46:39,960 working in monkeys. So we had, we've had it working 817 00:46:39,960 --> 00:46:42,800 with monkeys for a while now where you can sort of flash a 818 00:46:42,800 --> 00:46:46,280 pixel and then you watch where the monkey, obviously the 819 00:46:46,280 --> 00:46:49,320 monkey's like, what monkey's a little surprised to see like, 820 00:46:49,320 --> 00:46:51,280 hey, there's a flash here and a flash here. 821 00:46:51,720 --> 00:46:54,080 But it's gets used to it after a while. 822 00:46:54,480 --> 00:46:57,120 But it just you can you can see that that the pixel is in the 823 00:46:57,120 --> 00:46:59,720 right location because the monkey's eyes will dark to that 824 00:46:59,720 --> 00:47:01,240 location. It's not on the screen, like 825 00:47:01,240 --> 00:47:02,240 there's no pixel on the. Screen. 826 00:47:02,240 --> 00:47:03,560 There's no pixel on the. Screen. 827 00:47:03,720 --> 00:47:06,240 Yeah, just like. You just verify that that the 828 00:47:06,560 --> 00:47:08,520 that you're triggering a pixel in the right part of the brain. 829 00:47:08,520 --> 00:47:13,320 So, you know, the initial resolution for vision will be 830 00:47:13,320 --> 00:47:16,080 relatively low, you know, sort of Atari graphics type of thing. 831 00:47:16,080 --> 00:47:19,560 But over time, it could potentially be better than 832 00:47:19,560 --> 00:47:22,400 normal vision. And then I guess in terms of 833 00:47:22,400 --> 00:47:27,040 some additional applications for where writing to the brain can 834 00:47:27,040 --> 00:47:32,040 be useful order applications as Bliss mentioned there is. 835 00:47:32,240 --> 00:47:33,880 Feedback. There's a proprioceptive 836 00:47:33,880 --> 00:47:35,440 feedback, there's a tactile feedback. 837 00:47:35,440 --> 00:47:37,360 Especially for a robot arm, Like if you're trying to grasp a call 838 00:47:37,560 --> 00:47:38,800 right, you need to know if you've got it. 839 00:47:39,080 --> 00:47:41,760 Yeah, 1 to 1 egg. That's an egg, yeah. 840 00:47:42,120 --> 00:47:45,800 It's a very much a delicate balance of not just initiating 841 00:47:45,800 --> 00:47:47,840 the movement but getting the feedback and controlling it 842 00:47:48,000 --> 00:47:50,040 accordingly. So there there is a some meta 843 00:47:50,040 --> 00:47:53,040 sensory cortex that's right adjacent to motor cortex that 844 00:47:53,600 --> 00:48:07,080 could could be beneficable. Motor movements, so any changes 845 00:48:07,080 --> 00:48:11,080 in neural growth after device is inserted, we don't see any any 846 00:48:11,400 --> 00:48:16,440 signs of neural damage. But I and I guess we, we have 847 00:48:16,440 --> 00:48:21,680 seen some rebound on some of the electrodes, right? 848 00:48:23,120 --> 00:48:24,560 Correct. And then also, I mean, I guess, 849 00:48:24,600 --> 00:48:27,080 I guess, you know, rain is very plastic. 850 00:48:27,680 --> 00:48:31,640 So it's not that plastic well. It does diminish quite a bit 851 00:48:31,640 --> 00:48:39,040 after each and 20. Throughout childhood, actually, 852 00:48:39,160 --> 00:48:44,160 when you get to about 25 rain really done cooking. 853 00:48:44,560 --> 00:48:51,080 Yeah. But there are, there is a little 854 00:48:51,080 --> 00:48:54,480 bit of damage done with each insertion, but it's a miniscule 855 00:48:54,480 --> 00:48:57,000 amount compared to anything else out there. 856 00:48:57,000 --> 00:49:03,320 And so it's an easy amount of damage recover from. 857 00:49:03,320 --> 00:49:08,720 And it's really only detectable on cutting pieces of the brain 858 00:49:09,000 --> 00:49:13,520 after after the animal's no longer alive and looking at them 859 00:49:13,520 --> 00:49:15,920 under a microscope, you can't really tell during life that 860 00:49:15,920 --> 00:49:19,160 there's been any brain. Good. 861 00:49:19,160 --> 00:49:21,680 Another way to interpret this question, have there any changes 862 00:49:21,680 --> 00:49:23,320 in neural growth after the device is inserted? 863 00:49:23,680 --> 00:49:25,920 One way to interpret that is like the user learning how to 864 00:49:25,920 --> 00:49:27,600 use the device. And I think on that side of 865 00:49:27,600 --> 00:49:29,200 things, there's been tremendous progress. 866 00:49:29,480 --> 00:49:33,120 He's put in hundreds of hours trying to figure out the best 867 00:49:33,120 --> 00:49:34,840 way to use this device because he really thinks that, you know, 868 00:49:34,840 --> 00:49:36,560 if he can figure this out, he can help share this knowledge. 869 00:49:38,280 --> 00:49:40,320 I mean, he's like on Friday night at 8:00 PM, you know, he's 870 00:49:40,320 --> 00:49:43,080 starting a session of like, you know, figuring out himself how 871 00:49:43,080 --> 00:49:45,640 to how to push his own performance to the next level. 872 00:49:46,200 --> 00:49:48,640 And that's really a unique learning process because there's 873 00:49:48,880 --> 00:49:50,640 not many people in the world that had the experience of 874 00:49:50,640 --> 00:49:53,200 moving something. And so there's a lot of nuance 875 00:49:53,200 --> 00:49:56,440 to like, OK, how exactly should I imagine or attempt to move my 876 00:49:56,440 --> 00:49:59,960 wrist to get the thing to. Yeah, he's really dialed that 877 00:49:59,960 --> 00:50:02,680 into a. Also, just the sheer number of 878 00:50:02,680 --> 00:50:06,280 hours that he's using even in the past six months, right? 879 00:50:06,320 --> 00:50:09,040 Yeah, in many ways. Like, I mean, he's using it in 880 00:50:09,040 --> 00:50:10,880 his travel and in his plane, right? 881 00:50:11,360 --> 00:50:13,440 Effectively, PCI has left the lab, yeah. 882 00:50:13,720 --> 00:50:16,200 Yeah, one of the questions is how close are we converting 883 00:50:16,200 --> 00:50:19,320 thoughts into text? I mean, right, Right now it's 884 00:50:19,320 --> 00:50:21,880 more about moving curse from the screen on on a virtual keyboard, 885 00:50:22,520 --> 00:50:27,680 but long term you should be able to really estimate entire words 886 00:50:27,760 --> 00:50:29,200 faster than anyone could possibly type. 887 00:50:30,320 --> 00:50:34,920 I'm able to type Hello World today, but we're still in the 888 00:50:34,920 --> 00:50:36,720 early days making that a posh experience. 889 00:50:36,720 --> 00:50:39,600 I. Mean the other things that we're 890 00:50:39,600 --> 00:50:40,800 looking at is sign language, right? 891 00:50:40,920 --> 00:50:45,000 At the end of the day, it is a movement of and into, right? 892 00:50:45,520 --> 00:50:49,240 Yeah, it's true. Was the brain trying to 893 00:50:49,240 --> 00:50:52,280 naturally push the threads out? I mean, this is sort of a 894 00:50:52,280 --> 00:50:54,320 universal feature of any implant in the body. 895 00:50:54,320 --> 00:50:58,280 The body tries to reject it. And the goal of the surgeons and 896 00:50:58,280 --> 00:51:00,600 the technology team is to fight that. 897 00:51:00,920 --> 00:51:04,640 And so with artificial hips and with, you know, screws in the 898 00:51:04,640 --> 00:51:08,040 spine, we've done a really good job of finding biocompatible 899 00:51:08,040 --> 00:51:13,320 materials and techniques to fix those implants in the body. 900 00:51:13,760 --> 00:51:16,680 I mean, past a certain age, it's getting hard to find someone 901 00:51:16,680 --> 00:51:18,760 without some kind of implant, you know, knee. 902 00:51:19,200 --> 00:51:21,560 Hip. Some kind of screws in their 903 00:51:21,560 --> 00:51:25,360 spine. And so we've got this problem 904 00:51:25,920 --> 00:51:29,680 pretty well solved. So to answer your question, yes, 905 00:51:29,680 --> 00:51:33,760 the body is trying to get rid of any implant, but we can ensure 906 00:51:33,760 --> 00:51:37,560 that basically can't. It's also worth highlighting 907 00:51:37,560 --> 00:51:42,240 that the threads have not actually moved in the past five 908 00:51:42,240 --> 00:51:44,160 months. There's, there's some still 909 00:51:44,160 --> 00:51:46,680 minor movements in terms of like some maybe maybe getting pushed 910 00:51:46,720 --> 00:51:49,400 in a little bit, pushed out a little bit, but it's, it's more 911 00:51:49,400 --> 00:51:51,640 or less very stable and been stable for. 912 00:51:52,280 --> 00:51:56,240 And the reason for that is, you know, once you, once you do a 913 00:51:56,240 --> 00:51:59,680 brain surgery, it takes some time for the tissues to come in 914 00:51:59,680 --> 00:52:02,960 and then, and then you know, our tissue or the neo membrane to 915 00:52:02,960 --> 00:52:04,920 actually come in and then anchor the threats in place. 916 00:52:05,160 --> 00:52:07,720 And once that happens, everything has been stable and 917 00:52:08,160 --> 00:52:10,760 seen much movement. That's where the world record 918 00:52:10,840 --> 00:52:13,520 performance starts to come. In Yeah, that was a couple weeks 919 00:52:13,520 --> 00:52:18,560 ago. Yeah, the threats, like it is 920 00:52:18,560 --> 00:52:20,520 important that the threats be extremely tiny. 921 00:52:20,520 --> 00:52:23,080 If they're extremely tiny, then the brain does not. 922 00:52:23,480 --> 00:52:26,640 The smaller they are, the less likely the brain is to react to 923 00:52:26,640 --> 00:52:28,600 to them. So that's why you want the 924 00:52:28,600 --> 00:52:31,760 threats to be extremely tiny and also to minimize any damage to 925 00:52:31,760 --> 00:52:37,760 neurons, so. On that note, we do plan to 926 00:52:37,760 --> 00:52:41,360 actually share some of the, you know, the tissue response in 927 00:52:41,360 --> 00:52:44,440 detail and some of the the later upcoming updates. 928 00:52:44,920 --> 00:52:49,840 Yeah, it is quite a challenging, it's challenging on many fronts 929 00:52:49,840 --> 00:52:52,760 to do something like this because you're, you're trying to 930 00:52:52,880 --> 00:52:56,400 read and read and write electrical signals, but you need 931 00:52:56,400 --> 00:53:01,280 to have the, the threads themselves need to be like 932 00:53:01,600 --> 00:53:06,880 electrically isolated and and not subject to corrosion in the 933 00:53:06,880 --> 00:53:09,320 body. So like the, you know, just 934 00:53:09,800 --> 00:53:13,280 metal by itself is somewhat subject to corrosion or, or 935 00:53:13,280 --> 00:53:17,440 being attacked. So it's it's, it's like in terms 936 00:53:17,440 --> 00:53:19,760 of the various coatings and things to actually make this 937 00:53:19,760 --> 00:53:23,080 electrode work while not actually eroding its performance 938 00:53:23,080 --> 00:53:26,120 over time is, is very difficult. Human bodies are very, very 939 00:53:26,120 --> 00:53:27,640 harsh environment, very harsh environment. 940 00:53:27,640 --> 00:53:31,520 It's a it's a bag of salt water with bad sensors that's elevated 941 00:53:31,520 --> 00:53:33,080 temperature that is well regulated. 942 00:53:33,560 --> 00:53:35,480 I mean, I'm sure people have experienced dropping their 943 00:53:35,480 --> 00:53:38,200 electronic devices in a sea water and in an instant. 944 00:53:38,920 --> 00:53:43,000 Yeah, yeah. So we'll sort of wrap wrap this 945 00:53:43,000 --> 00:53:48,960 up soon if there's like a few few last questions, I guess. 946 00:53:49,080 --> 00:53:51,800 So a good question. So what about upgrades? 947 00:53:52,640 --> 00:53:55,920 So yeah, we do think it's going to be important to be able to 948 00:53:55,920 --> 00:53:59,560 upgrade the device over time, just like you wouldn't want like 949 00:53:59,560 --> 00:54:01,320 an iPhone 1 stuck in your brain forever. 950 00:54:02,680 --> 00:54:05,760 You know, if you've got an iPhone 15, you probably want the 951 00:54:05,760 --> 00:54:11,400 iPhone 15, not the iPhone 1. So I think people over time will 952 00:54:13,000 --> 00:54:14,520 be able to upgrade their neural link. 953 00:54:14,520 --> 00:54:17,640 So we'll take the neural link device out and put a new one in. 954 00:54:19,080 --> 00:54:24,920 And we have done this with some of our animals and they're 955 00:54:24,920 --> 00:54:28,840 actually in one case we did with we upgraded device three times 956 00:54:28,840 --> 00:54:32,720 and so with a pig. We did with a monkey as well. 957 00:54:32,720 --> 00:54:36,280 Monkey is able to do PCI. Yeah, and he's doing fine. 958 00:54:36,920 --> 00:54:39,240 Pager has the latest implant. Actually, he hit his, I think 959 00:54:39,240 --> 00:54:41,680 his record with the last, yeah. With the with an upgrade. 960 00:54:42,800 --> 00:54:44,040 No, it still beat him though Will. 961 00:54:44,480 --> 00:54:45,800 It still beat him. Yes, this is true. 962 00:54:46,000 --> 00:54:47,880 Humans are top of the species leaderboard right now. 963 00:54:48,160 --> 00:54:49,440 Pager's like what? Like 8 or something. 964 00:54:49,520 --> 00:54:53,480 Pager's like 8.5 BPS, OK, but it's a very high score, yeah. 965 00:54:53,640 --> 00:54:55,080 I'm not trying to put Pager down. 966 00:54:55,080 --> 00:54:57,080 And also to train a monkey to do that, It's like a whole 967 00:54:57,080 --> 00:54:58,920 challenge on it's own. We have like the best animal 968 00:54:58,920 --> 00:55:01,600 care team world. Yeah, I just do 1/2 size. 969 00:55:01,600 --> 00:55:05,280 We, we, we do our absolute best to take care of the animals. 970 00:55:06,240 --> 00:55:10,960 And when we had like a USDA inspector come through, she said 971 00:55:10,960 --> 00:55:15,520 that this was the the nicest animal facility she's ever seen 972 00:55:15,520 --> 00:55:16,920 in her entire life. So. 973 00:55:17,040 --> 00:55:22,240 Breakfast on an app like. The the the monkey orders room 974 00:55:22,240 --> 00:55:24,160 service. Yeah. 975 00:55:24,160 --> 00:55:27,240 We're, we're, we're monkey room service, which is rare. 976 00:55:28,320 --> 00:55:31,000 Like we're the only ones who offer monkey room service. 977 00:55:31,680 --> 00:55:35,200 So we really do everything we can to maximize the welfare of 978 00:55:35,200 --> 00:55:38,520 the animals. So all right, with that, thank 979 00:55:38,520 --> 00:55:41,280 you everyone for tuning in. Hope you found this interesting.