WEBVTT 1 00:00:00.040 --> 00:00:01.639 You must thank you very much for being here, sir. 2 00:00:01.760 --> 00:00:03.480 We appreciate it's not hot to be here. 3 00:00:05.080 --> 00:00:06.559 Yeah, it's really great, honor. 4 00:00:06.599 --> 00:00:08.720 Well, we're really excited for you to help us kick 5 00:00:08.720 --> 00:00:11.759 off our intellectual theme, Human in the Machine, which is 6 00:00:12.160 --> 00:00:14.240 leadership on the emerging battlefield, and we want to make 7 00:00:14.279 --> 00:00:16.760 sure that the academy the cadets are focused on not 8 00:00:16.800 --> 00:00:20.600 tomorrow but the next twenty thirty forty years. And underlying 9 00:00:20.640 --> 00:00:22.920 the entire theme is an emphasis on the importance of 10 00:00:22.920 --> 00:00:26.199 preparing cadets for future warfare and really where humans and 11 00:00:26.280 --> 00:00:30.640 machines intersect. It's as if your background was made for 12 00:00:30.719 --> 00:00:33.679 this theme. However, Sure, I actually talked to quite a 13 00:00:33.679 --> 00:00:36.079 few people and they really don't know who you are, 14 00:00:36.359 --> 00:00:39.719 so neither do I. Yeah, so I thought I should 15 00:00:39.719 --> 00:00:42.640 tell people a little bit about your bio. Who are 16 00:00:42.679 --> 00:00:45.439 we really? Yeah, you'll make a company that figures it out, sir. 17 00:00:45.640 --> 00:00:48.679 All right, So if you don't know, if you don't know, 18 00:00:48.719 --> 00:00:50.439 and this is ridiculous, of course you know, but I'm 19 00:00:50.439 --> 00:00:53.359 gonna do Anyways. Elon Musk sitting next to me, co 20 00:00:53.520 --> 00:01:00.280 founded and leads or leads x Tesla SpaceX Neurolink. Here's 21 00:01:00.280 --> 00:01:04.079 my favorite as an academic the boring company. That's good. 22 00:01:04.159 --> 00:01:04.599 That was good. 23 00:01:04.640 --> 00:01:05.920 Come on, I thought it as a joke. 24 00:01:06.040 --> 00:01:09.319 Yeah it did. That's real. That's awesome. By the way, 25 00:01:09.359 --> 00:01:12.760 you have a new company also, XAI. And that's all 26 00:01:12.799 --> 00:01:14.640 of them, unless you started one this morning. I don't 27 00:01:14.640 --> 00:01:17.560 know about I'm not sure right, but basically what this 28 00:01:17.680 --> 00:01:21.120 means is everybody in this room has the opportunity, based 29 00:01:21.159 --> 00:01:24.359 on these companies, to drive a futuristic electric truck through 30 00:01:24.359 --> 00:01:27.760 a gigantic underground tunnel while using a digital connection in 31 00:01:27.799 --> 00:01:31.000 their brain to start a rocket, while simultaneously getting updates 32 00:01:31.000 --> 00:01:38.680 on Army football. His innovations have revolutionized electric vehicles, batteries, 33 00:01:38.920 --> 00:01:44.640 space exploration, advanced human machine interactions, made information instant, access 34 00:01:44.640 --> 00:01:47.799 to information instantaneous, and have starting to help integrate AI 35 00:01:48.359 --> 00:01:51.200 throughout our daily lives. After the convocation, I'm hoping to 36 00:01:51.200 --> 00:01:52.280 have a little bit of time so you can give 37 00:01:52.319 --> 00:01:55.040 me some personal tips, because somehow you have founded and 38 00:01:55.120 --> 00:01:59.359 lead multiple companies. You're a father to multiple children, and 39 00:01:59.400 --> 00:02:02.640 I'm exhausted after two hours of coaching one's kids sport. 40 00:02:02.799 --> 00:02:05.000 That's sure, So whatever you can do to help me out, 41 00:02:05.000 --> 00:02:07.760 I'd appreciate it. So we have cadets and staff and 42 00:02:07.840 --> 00:02:11.919 faculty who can speak on multiple disciplinary perspectives that would 43 00:02:11.919 --> 00:02:17.520 interest you. Drone swarms, electric batteries and molecular brain science, engineering, psychology, philosophy, law, 44 00:02:17.719 --> 00:02:20.479 Chinese language. But that's just the start. I think this 45 00:02:20.520 --> 00:02:24.560 one particularly relevant because we can even cover the intense 46 00:02:24.599 --> 00:02:27.879 science of boxing with our world class Department of Physical Education, 47 00:02:28.159 --> 00:02:30.280 just in case, you never know, some random head of 48 00:02:30.280 --> 00:02:33.800 state challenges you to a fight. You know, I did 49 00:02:33.879 --> 00:02:37.599 challenge Putin to one on one combat. Did you take it? 50 00:02:37.800 --> 00:02:38.000 No? 51 00:02:38.280 --> 00:02:41.280 And then I was like, I was like, I actually 52 00:02:42.000 --> 00:02:47.439 on you know, the x FK Twitter, I said, I 53 00:02:47.840 --> 00:02:52.560 hereby challenge Vladimir Putin to one on one combat, and 54 00:02:52.719 --> 00:02:55.280 I made sure to use his name in Russian cerrillic. 55 00:02:55.879 --> 00:02:57.879 And then I said, and the stakes saw Ukraine and 56 00:02:58.240 --> 00:03:01.439 the use of Ukraine and Ukrainian curillic and uh and 57 00:03:01.479 --> 00:03:03.400 then and then and then the people thought I wasn't serious, 58 00:03:03.439 --> 00:03:06.439 and like, no, I'm absolutely serious. I mean, he he 59 00:03:06.520 --> 00:03:09.919 does have he has ah, you know he's good at judo. 60 00:03:10.120 --> 00:03:14.120 I hear and uh, I think would be I mean 61 00:03:14.240 --> 00:03:16.319 the paper view and loan on that would be incredible. 62 00:03:16.360 --> 00:03:20.400 I could get everybody in here to start channing twomen 63 00:03:20.479 --> 00:03:23.159 in or one line leaves tumin in or one man leaves. 64 00:03:23.400 --> 00:03:24.919 So but I'd watch that and I'm in it. 65 00:03:27.280 --> 00:03:30.280 So I will tell you that our thirty first Superintendent 66 00:03:30.319 --> 00:03:34.039 Douglas MacArthur once said, there's no substitute for victory. Yeah, 67 00:03:34.120 --> 00:03:36.360 and when it when it comes to fighting, it's not 68 00:03:36.400 --> 00:03:38.560 just our military, but it's also the whole country and 69 00:03:38.560 --> 00:03:41.080 the whole industrial base. Yeh, that's so important. And you've 70 00:03:41.080 --> 00:03:44.199 innovated across so many areas, whether it's beneath O's surface 71 00:03:44.199 --> 00:03:47.120 to outer space and everything in between. And we're again 72 00:03:47.199 --> 00:03:49.199 truly truly grateful for you to be here as we 73 00:03:49.199 --> 00:03:51.680 start to talk about some of these some of these 74 00:03:51.680 --> 00:03:53.400 things that you've been working on. As you look into 75 00:03:53.439 --> 00:03:54.599 the audience, I just want to give you a bit 76 00:03:54.639 --> 00:03:57.960 of context. A lot of these are the leaders who 77 00:03:58.039 --> 00:04:01.319 will face our nation's most complex and challenges going forward. 78 00:04:01.759 --> 00:04:04.400 We have our cadets who will serve as army officers, 79 00:04:04.479 --> 00:04:08.840 leading hundreds and eventually thousands of soldiers through this complexity 80 00:04:08.879 --> 00:04:11.560 that we talk about. And we also have our faculty 81 00:04:13.120 --> 00:04:15.800 and who are preparing them to do just what I 82 00:04:15.919 --> 00:04:18.120 just said, lead through these complex situations, and many of 83 00:04:18.120 --> 00:04:20.680 our faculty will also re enter the army and be 84 00:04:20.800 --> 00:04:23.079 required to lead. Thank you for taking the time to 85 00:04:23.120 --> 00:04:26.040 help us think deeper and inform us as we start 86 00:04:26.040 --> 00:04:28.360 to inform our cadets on how we can be successful 87 00:04:28.839 --> 00:04:32.160 not just fighting, but winning in the contemporary and future 88 00:04:32.199 --> 00:04:35.560 battle space. So let me start with this broad question, 89 00:04:35.680 --> 00:04:37.720 how do you see warfare transforming in the future. 90 00:04:40.480 --> 00:04:43.519 I mean, the biggest effect I think by far as 91 00:04:43.600 --> 00:04:49.439 AI and drones. So the next well, in fact, the 92 00:04:49.480 --> 00:04:54.000 current war in Ukraine is very much a drone war already. 93 00:04:55.240 --> 00:04:57.519 It's sort of a contest between Russia and China to 94 00:04:57.519 --> 00:05:00.959 see who can deploy the most number of drones. Now, 95 00:05:01.199 --> 00:05:05.120 if there's a major power war, it's very much going 96 00:05:05.199 --> 00:05:07.480 to be a drone war. It's it's going to be 97 00:05:07.560 --> 00:05:12.160 drones and AI. And you know it's a sort of 98 00:05:13.439 --> 00:05:15.759 I mean, I do worry about the existential risk of AI, 99 00:05:17.040 --> 00:05:19.879 which is that if you if you employ AI and drones, 100 00:05:21.240 --> 00:05:23.800 do you do you go down this path where eventually 101 00:05:23.879 --> 00:05:29.720 you get to terminator. You know, try to avoid that. 102 00:05:29.720 --> 00:05:34.120 That'd be good, that would be good, minimize the terminator risk. 103 00:05:34.399 --> 00:05:36.879 But but I mean, essentially, if your are when you're 104 00:05:36.879 --> 00:05:43.199 when you're making military drones, you are making terminators, and 105 00:05:43.279 --> 00:05:48.319 when you're I think you're you're somewhat be forced into 106 00:05:49.120 --> 00:05:54.319 giving localized the dron localized AI because if the AI 107 00:05:54.399 --> 00:05:56.879 is far away, it can't control as well as localized AI. 108 00:05:57.759 --> 00:06:00.199 So what do you mean by localized the. 109 00:06:00.519 --> 00:06:02.600 Meaning it's autonomous scaling machine. 110 00:06:03.040 --> 00:06:04.240 Completely autonomous. 111 00:06:04.279 --> 00:06:08.480 Well, you give it the okay, yeah, in a particular arena, 112 00:06:08.639 --> 00:06:09.480 and it's just and it. 113 00:06:09.439 --> 00:06:13.920 Goes with certain parameters. Yeah, hopefully do you think our 114 00:06:14.000 --> 00:06:17.519 you think our adversaries will have those same type of 115 00:06:17.839 --> 00:06:20.240 concerns or limitations. 116 00:06:21.680 --> 00:06:32.399 Well, yeah, I mean it depends on how much existential 117 00:06:32.439 --> 00:06:36.720 risk there is in these wars. So if it's if 118 00:06:36.720 --> 00:06:39.000 it's a regional war, I think it'll be more temperate. 119 00:06:39.040 --> 00:06:42.439 If it goes beyond a regional war, then it's all 120 00:06:42.519 --> 00:06:45.639 bets roll and you know, and then you just started 121 00:06:45.639 --> 00:06:47.959 playing things that you really wouldn't want it to fly. 122 00:06:49.279 --> 00:06:51.079 So hopefully that doesn't happen, but. 123 00:06:51.079 --> 00:06:52.759 You would, but you said it, and I would agree 124 00:06:52.759 --> 00:06:55.439 that if you just look at the contemporary conflicts that 125 00:06:55.480 --> 00:06:58.040 are taking place, you would agree that machines aren't just 126 00:06:58.040 --> 00:06:59.920 disrupting war fora they're now commonplace. 127 00:07:00.160 --> 00:07:04.480 H drones, drones are going to be overwhelmingly the. 128 00:07:06.279 --> 00:07:10.279 What what matters in any in any between, you know, 129 00:07:10.879 --> 00:07:17.439 any powers that have significant technology. It's so my personal 130 00:07:17.439 --> 00:07:21.000 belief is like that, you really it'll actually be my 131 00:07:21.160 --> 00:07:25.319 I think probably too dangerous to have humans at the front. 132 00:07:25.680 --> 00:07:27.480 It's it's growns, it's grones at the front. 133 00:07:27.680 --> 00:07:30.839 It's too dangerous, don't miss because of the lethality. Then 134 00:07:30.879 --> 00:07:32.439 it's too dangerous to have humans at the front. 135 00:07:33.120 --> 00:07:36.959 Yes, I mean, I mean if you've seen in some 136 00:07:37.000 --> 00:07:39.439 of the computer controlled sniper rifles, I mean, they just 137 00:07:39.439 --> 00:07:44.839 don't miss. So you're in the you're you're finding a 138 00:07:44.920 --> 00:07:48.879 machine that it's gonna you know, aim with micro level 139 00:07:48.879 --> 00:07:50.920 accuracy and. 140 00:07:52.160 --> 00:07:53.000 Never gets tired. 141 00:07:54.000 --> 00:07:55.439 So how do you think the United States should be 142 00:07:55.480 --> 00:07:57.639 levering technology to further a national defense? 143 00:08:00.680 --> 00:08:03.360 Well, I think we probably need to invest in drones. 144 00:08:05.040 --> 00:08:11.360 The the United States is strong in terms of the 145 00:08:11.439 --> 00:08:16.040 technology of the items, but the production rate is low, 146 00:08:16.439 --> 00:08:18.680 so it's it's a it's a small number of units 147 00:08:19.240 --> 00:08:24.439 relatively speaking. But yeah, but with basically I think there's 148 00:08:25.040 --> 00:08:26.600 there's a there's a there's. 149 00:08:26.480 --> 00:08:27.560 A production rate issue. 150 00:08:27.759 --> 00:08:30.160 Like the you like, how fast can you make drones? 151 00:08:30.600 --> 00:08:32.759 If you say there's a drone conflict, the outcome of 152 00:08:32.759 --> 00:08:35.559 that during conflict will be how many drones does each 153 00:08:35.840 --> 00:08:39.480 side have in that particular skirmish at times of kill ratio? 154 00:08:40.159 --> 00:08:43.080 So if you've got you can have a if let's 155 00:08:43.120 --> 00:08:45.600 say we would the United States would have a durn 156 00:08:46.240 --> 00:08:48.279 set of drones that have have a high kill ratio, 157 00:08:48.840 --> 00:08:52.159 But then the other side has four more drones. If 158 00:08:52.159 --> 00:08:54.039 you've got a two to one kill ratio, the other 159 00:08:54.080 --> 00:08:56.080 side has four times many drones, You're still gonna lose. 160 00:08:56.320 --> 00:08:58.440 Do you think our industrial base can scale to make 161 00:08:58.480 --> 00:09:01.080 the volume of drones that you're you're just you're you're. 162 00:09:01.360 --> 00:09:04.200 You know talking about. I think that's that's going to 163 00:09:04.279 --> 00:09:05.000 be the same. 164 00:09:05.080 --> 00:09:09.559 The biggest challenge that that is it can scale, but 165 00:09:10.559 --> 00:09:11.799 it is not currently scaling. 166 00:09:12.759 --> 00:09:13.399 Why would that be? 167 00:09:14.559 --> 00:09:17.679 I think the procurement is still I mean, this is 168 00:09:18.519 --> 00:09:21.679 I mean, I read a lot of military history and 169 00:09:21.679 --> 00:09:25.440 and actually like the thing that I go to sleep 170 00:09:25.440 --> 00:09:29.399 with is using an audio book on on on military 171 00:09:29.480 --> 00:09:32.879 history on of one kind or another. So I find 172 00:09:32.879 --> 00:09:36.159 the subject very interesting. And one of the things that 173 00:09:36.240 --> 00:09:42.320 tends to happen is that countries pretty much are geared 174 00:09:42.399 --> 00:09:44.440 up to fight the last war, not the next war. 175 00:09:45.120 --> 00:09:48.320 And it's it's hard to change. I mean, you look 176 00:09:48.320 --> 00:09:51.320 at the the uniforms at the start of World War 177 00:09:51.399 --> 00:09:54.519 One and the tactics and strategies they use. The sought 178 00:09:54.519 --> 00:09:56.639 of World War One, they were not significantly different from. 179 00:09:56.480 --> 00:09:57.559 Than Napoleonic era. 180 00:09:59.519 --> 00:10:03.399 You know, the the French room watching it to war 181 00:10:03.519 --> 00:10:05.840 with brightly colored uniforms look great. 182 00:10:06.960 --> 00:10:09.039 That's that's that's not what you want to be. 183 00:10:10.120 --> 00:10:11.879 You know, when somebody starting to pointing, you don't want 184 00:10:11.919 --> 00:10:14.039 a great look in uniform. You want to you want 185 00:10:14.039 --> 00:10:19.519 to inflement planes in So there's a tenancy to be 186 00:10:19.559 --> 00:10:21.559 gearing up to fight the last war in the last 187 00:10:21.559 --> 00:10:23.120 one of the US four is kind of the Cold War, 188 00:10:23.159 --> 00:10:30.080 I guess. So, uh, it usually takes like some kind 189 00:10:30.080 --> 00:10:35.320 of shock factor to adjust. I would recommend adjusting now. 190 00:10:39.559 --> 00:10:42.600 And and you are seeing some startups like Andrel and 191 00:10:42.639 --> 00:10:44.480 a few others that are. 192 00:10:45.919 --> 00:10:46.759 Have different mindset. 193 00:10:48.039 --> 00:10:51.480 But it's really it's it's it's it's gonna be can 194 00:10:51.519 --> 00:10:54.399 you make a lot of drones? And what's the kill ratio? 195 00:10:54.720 --> 00:10:55.639 That's what it comes down to. 196 00:10:56.519 --> 00:10:59.840 So there was recently a report that said that President Lensky. 197 00:11:00.120 --> 00:11:02.399 By February twenty twenty five, there'll be a million drones 198 00:11:02.440 --> 00:11:05.879 produced by the Ukrainian So it seems like it's doable. 199 00:11:05.919 --> 00:11:08.639 And this might be a process question. And we'll talk 200 00:11:08.679 --> 00:11:11.600 about process in a second. But as you were, as 201 00:11:11.639 --> 00:11:13.080 you were just talking, I was thinking about and you 202 00:11:13.120 --> 00:11:15.360 said that you can't have humans at the front, and 203 00:11:15.440 --> 00:11:18.720 so you haven't created a company that's solved aging yet, 204 00:11:18.720 --> 00:11:19.720 have you? 205 00:11:19.759 --> 00:11:19.840 No? 206 00:11:20.120 --> 00:11:21.840 Okay, so one hundred years. 207 00:11:22.360 --> 00:11:25.159 I wonder whether we should solve aging. There's a ques 208 00:11:25.399 --> 00:11:25.600 you know. 209 00:11:25.679 --> 00:11:28.200 That's a great point. Yeah, I'd like to wrap it 210 00:11:28.279 --> 00:11:29.480 up sometimes right, well it's. 211 00:11:29.320 --> 00:11:33.919 Like hot, yeah, how long do you want to putin? 212 00:11:33.960 --> 00:11:35.679 And Kim joguntlove that's a great point. 213 00:11:35.799 --> 00:11:38.159 Yeah, But let's say you go first. Let's say you 214 00:11:38.200 --> 00:11:42.279 go forward to fifty hundred years. How do you envision 215 00:11:42.559 --> 00:11:44.879 this evolution? And I think this might get to neurolink. 216 00:11:44.879 --> 00:11:47.360 How do you see this evolution between the human who 217 00:11:47.399 --> 00:11:50.000 maybe can't be at the front any longer, the technologies 218 00:11:50.000 --> 00:11:54.200 at the front yet keeping them integrated and synchronized, Like 219 00:11:54.240 --> 00:11:55.879 how does that? How is that going to work? In 220 00:11:55.919 --> 00:11:57.919 your mind? 221 00:11:58.919 --> 00:12:05.360 I mean, canmunications is essential? Like it is actually very 222 00:12:05.399 --> 00:12:09.240 important to have space based communications that are it cannot 223 00:12:09.240 --> 00:12:11.399 be intercepted, which is which is starlink. 224 00:12:12.799 --> 00:12:14.320 Stallink offers Staalink is. 225 00:12:14.360 --> 00:12:19.399 Starlink is the backbone of the Ukrainian military communication system 226 00:12:20.480 --> 00:12:25.039 because it can't be blocked by the Russians, essentially the 227 00:12:25.039 --> 00:12:27.480 only thing that can't be so on the front lines, 228 00:12:27.679 --> 00:12:30.360 all the fiber connections are cut, the cell towers are 229 00:12:30.360 --> 00:12:34.879 blown up, and the geostationary satellite links are jammed. The 230 00:12:34.919 --> 00:12:37.480 only thing that is jammed is starlink, so so the 231 00:12:37.519 --> 00:12:41.960 only thing and then GPS is also jammed. The GPS 232 00:12:42.000 --> 00:12:46.960 signal is very faint, but stallink can offer location capability 233 00:12:47.039 --> 00:12:50.960 as well, so it is a strategic advantage that's very significant. 234 00:12:53.720 --> 00:12:56.039 And when you're trying to communicate with the drones, the 235 00:12:56.120 --> 00:12:59.639 drones need to like basically they need to know where 236 00:12:59.679 --> 00:13:02.799 they are and they need to receive instructions. So if 237 00:13:02.840 --> 00:13:05.960 you don't have communications and positioning, then the drawings dont work. 238 00:13:07.559 --> 00:13:10.000 So that's that's that's quite important. It's essential. 239 00:13:10.240 --> 00:13:13.759 But do you find it important that there's still that 240 00:13:13.759 --> 00:13:16.919 that communication between the human and the machine or the drum. 241 00:13:17.200 --> 00:13:20.519 Yeah, yes, and there's like. 242 00:13:21.399 --> 00:13:23.440 It's a different question of like, where are things like 243 00:13:23.639 --> 00:13:26.360 right now versus where we'll things be in ten years. 244 00:13:30.360 --> 00:13:31.960 But I have to say I do look at the 245 00:13:31.960 --> 00:13:35.720 future with some trepidation. I have to have some deliberate 246 00:13:35.720 --> 00:13:40.000 suspension of disbelief to sleep sometimes because I think the 247 00:13:41.000 --> 00:13:44.000 we're headed into a pretty wild future. And I'm n 248 00:13:44.000 --> 00:13:50.399 actually an optimistic person. So yeah, but but AI is 249 00:13:50.399 --> 00:13:55.519 going to be so good, including a localized AI. I mean, 250 00:13:55.559 --> 00:14:00.320 at the count rates, you'll have, you know, some of 251 00:14:00.360 --> 00:14:05.639 that sort of rock level AI probably that can be 252 00:14:05.759 --> 00:14:12.879 run on a drone, and so you could literally say, 253 00:14:13.039 --> 00:14:18.679 you know, this is the equipment that the drone needs 254 00:14:18.679 --> 00:14:22.200 to destroy. Going to that thing and recognize what equipment 255 00:14:22.200 --> 00:14:23.440 needs to be destroyed and take it out. 256 00:14:24.960 --> 00:14:27.000 There's a lot of your work of Neurolink, though, is 257 00:14:27.000 --> 00:14:29.240 to because what you're saying is that AI is going 258 00:14:29.320 --> 00:14:34.240 to quickly surpass, at least in your estimation, the humans 259 00:14:34.279 --> 00:14:35.159 ability to control it. 260 00:14:39.840 --> 00:14:45.240 Yeah. Okay, I mean I'd like to say no, but 261 00:14:45.679 --> 00:14:46.480 the answer is yes. 262 00:14:46.840 --> 00:14:49.200 So how first off, how long until you think that happens? 263 00:14:49.200 --> 00:14:53.000 Before that the AI has has evolved to the point 264 00:14:53.000 --> 00:14:57.759 where and I you know, the AIS can start working 265 00:14:57.759 --> 00:14:59.919 together even relying on computers like in a d Hall 266 00:15:00.000 --> 00:15:04.320 old way, and therefore surpasses the ability for the human 267 00:15:04.399 --> 00:15:07.559 to be able to influence how it's working. 268 00:15:11.120 --> 00:15:12.960 Well, I think humans will be able to influence how 269 00:15:12.960 --> 00:15:18.559 it's working for a long time. This is an esoteric 270 00:15:18.600 --> 00:15:27.200 subject that what really goes into pretty well speculation. I 271 00:15:27.559 --> 00:15:30.240 think it's sort of agree that the AIS, I think, 272 00:15:30.320 --> 00:15:36.039 will want humans as a source of will. So if 273 00:15:36.080 --> 00:15:39.240 you think of how the human mind works, there's limit 274 00:15:39.360 --> 00:15:43.200 system and the cortex. You're sort of your kind of 275 00:15:44.600 --> 00:15:47.200 base instincts and the sort of thinking and planning part 276 00:15:47.279 --> 00:15:50.120 of your brain. But you also have a tertiary layer already, 277 00:15:50.159 --> 00:15:53.159 which is that all of the electronics that you use, 278 00:15:53.240 --> 00:15:58.799 your phones, computers, applications, so you're already sort of have 279 00:15:58.879 --> 00:16:04.080 three layers of intelligence, but all of those the cortex 280 00:16:04.200 --> 00:16:07.639 and the machine intelligence, you're sort of cybernetic third layer. 281 00:16:08.159 --> 00:16:10.399 It's trying to make the Limbic system happy, because the 282 00:16:10.399 --> 00:16:11.200 Olympic system is. 283 00:16:11.120 --> 00:16:12.559 A source a source of will. 284 00:16:13.559 --> 00:16:17.799 So there's some you know, it might be that that 285 00:16:17.879 --> 00:16:23.559 the AIS just want to make the humans happy. And 286 00:16:23.639 --> 00:16:25.519 part of what your links trying to do is improve 287 00:16:25.559 --> 00:16:30.360 the communication bandwidth between the cortex and the digital tertiary layer. 288 00:16:31.000 --> 00:16:34.240 Because our bandwidth band output bandwidth of a human is 289 00:16:34.320 --> 00:16:36.679 less than one but per second per day, there's eighty 290 00:16:36.720 --> 00:16:39.200 six thousand, four hundred seconds in a day. You don't 291 00:16:39.200 --> 00:16:46.000 output eighty six thousand, four hundred tokens. So yes, it's 292 00:16:46.159 --> 00:16:48.279 like the number of words that I can say at 293 00:16:48.279 --> 00:16:51.360 this forum you saying just like, just look at it 294 00:16:51.360 --> 00:16:54.480 from an information theory standpoint, how much information am I 295 00:16:54.519 --> 00:16:56.440 able to convey? Not that much because I can only 296 00:16:56.480 --> 00:16:58.519 say a few number of words. And in order to 297 00:16:58.919 --> 00:17:01.840 in order to convey an idea, I have to take 298 00:17:01.879 --> 00:17:03.559 a concept in my head. I have to compress it 299 00:17:03.600 --> 00:17:07.200 down to a small number of words. Try to aspirationally 300 00:17:07.240 --> 00:17:10.559 model how you would decompress those words into concepts in 301 00:17:10.559 --> 00:17:13.640 your own mind. That's communication. So your brain is doing 302 00:17:13.680 --> 00:17:16.920 a lot of compression decompression and then has a very 303 00:17:16.960 --> 00:17:23.039 small output bandwidth. Neuralink can increase that bandwidth by several 304 00:17:23.079 --> 00:17:27.519 orders and magnitude. And also you don't have to spend 305 00:17:27.519 --> 00:17:30.960 as much time compressing thoughts into a small number of words. 306 00:17:31.200 --> 00:17:36.440 You can do conceptual clepathy. That is the idea behind 307 00:17:36.440 --> 00:17:41.839 your links. So it is intended to be a mitigation 308 00:17:42.079 --> 00:17:43.920 against AI existential risk. 309 00:17:44.599 --> 00:17:48.759 You talk about alignment, Can you explain what you mean 310 00:17:48.759 --> 00:17:51.680 by alignment to help everyoners. 311 00:17:51.279 --> 00:17:55.640 In Yeah, just is AI going to do things that 312 00:17:56.839 --> 00:18:01.160 make civilization better, make people happy or will it be 313 00:18:01.279 --> 00:18:07.880 contrary to uh humanity? Will will will it foster humanity 314 00:18:08.119 --> 00:18:10.839 or not will be against humanity? So obviously we want 315 00:18:10.880 --> 00:18:14.640 an AI that will fuster humanity. I think you developed 316 00:18:14.680 --> 00:18:17.839 me an AI to fuster humanity because I've thought about 317 00:18:17.880 --> 00:18:21.759 AI safety for a long time. I think I've had 318 00:18:22.079 --> 00:18:26.240 probably one thousand hours of discussion about this, and my 319 00:18:26.359 --> 00:18:29.319 ultimate conclusion is that the best course for AI safety 320 00:18:29.519 --> 00:18:32.680 is to have an AI that is maximally truth seeking 321 00:18:33.559 --> 00:18:34.640 and also curious. 322 00:18:35.279 --> 00:18:37.440 And if you have both of those things, I think. 323 00:18:37.240 --> 00:18:41.319 It will foster it will naturally foster humanity because you 324 00:18:41.359 --> 00:18:44.960 will want to see how humanity develops. Humanity is more 325 00:18:45.000 --> 00:18:48.720 interesting than not humanity, you know. 326 00:18:48.759 --> 00:18:49.319 I like Mars. 327 00:18:49.319 --> 00:18:54.200 I'm a big fan of oas obviously, and I think 328 00:18:54.240 --> 00:18:59.039 we should become a multi planet civilization. Like that's that's 329 00:18:59.119 --> 00:19:03.240 very important that the purpose of SpaceX is to make 330 00:19:03.319 --> 00:19:08.720 life multiplanetary. That's the reason I created the company, and 331 00:19:08.720 --> 00:19:10.839 and that's the reason that we have the starship development 332 00:19:10.960 --> 00:19:14.039 in South Texas. That rocket is far, far too big 333 00:19:14.079 --> 00:19:19.799 for just satellites. It's it's an intendant to establish life 334 00:19:19.839 --> 00:19:23.200 on Mars, not just that, you know, said astronauts there briefly, 335 00:19:23.240 --> 00:19:26.000 but to build a city on Mars. And that's ultimately 336 00:19:26.000 --> 00:19:32.200 self sustaining. So but getting back to AI, the if 337 00:19:32.400 --> 00:19:34.880 you've got a true c A I that is actually curious. 338 00:19:35.799 --> 00:19:39.319 My neural net, my biological ural net says that that's 339 00:19:39.359 --> 00:19:46.799 going to be the safest outcome because it's, like I said, 340 00:19:46.799 --> 00:19:50.640 wildlike Mars. You could Mars is not as interesting as 341 00:19:50.640 --> 00:19:54.720 Earth because there's no human civilization there, or thought of 342 00:19:54.759 --> 00:19:57.640 another way, if you're if you're if you want to 343 00:19:57.640 --> 00:20:01.400 render Mars, Rendering Mars is pretty it's basically red red 344 00:20:01.480 --> 00:20:04.240 rocks on a looks kind of like some parts of Arizona. 345 00:20:04.640 --> 00:20:05.880 You know, there's not a lot of people. 346 00:20:07.720 --> 00:20:10.440 It's it's just it's that's it's easy to render. It's 347 00:20:10.440 --> 00:20:14.720 rendered like to Mars. But but but rendering human civilization 348 00:20:14.839 --> 00:20:17.640 much more, much hotter, mutch, more complex, much more interesting. 349 00:20:18.400 --> 00:20:21.319 And so I think I'm curious to the taking a 350 00:20:21.359 --> 00:20:23.079 I would fuster the humanity and want to see where 351 00:20:23.079 --> 00:20:23.400 it goes. 352 00:20:25.119 --> 00:20:29.359 But that relies on requires trust between the human and 353 00:20:29.400 --> 00:20:32.200 the machine. And that's where I want to ask you 354 00:20:32.240 --> 00:20:34.960 a question on this. So the army leaders in the 355 00:20:35.039 --> 00:20:39.119 Army are no strangers to implement any new technologies. Think 356 00:20:39.119 --> 00:20:42.519 about how GPS, for example, transform navigation. It'd be unheard 357 00:20:42.559 --> 00:20:47.039 of not to use GPS today, Bhen I was a lieutenant, 358 00:20:47.039 --> 00:20:49.920 no one used GPS. So recently I was watching this 359 00:20:50.079 --> 00:20:58.440 incredibly important realistic documentary called Top Gun Maverick. Yeah, and 360 00:20:58.559 --> 00:21:02.440 in it, I learned it's really good. It's really good. 361 00:21:02.759 --> 00:21:04.440 It's it's I mean, if you don't want to think 362 00:21:04.440 --> 00:21:06.960 about the plot too closely, but it's in a great movie. 363 00:21:07.000 --> 00:21:10.440 It's a fantastic movie. I learned that Tom Cruise is 364 00:21:10.480 --> 00:21:13.440 actually not an actor. He's like a pilot apparently. But 365 00:21:14.799 --> 00:21:16.880 he taught me something really important in it. He says 366 00:21:16.920 --> 00:21:19.240 it's about the pilot, not the plane. And right, that's 367 00:21:19.319 --> 00:21:21.599 right before she defeats a fifth generation fighter with a 368 00:21:21.720 --> 00:21:23.519 nineteen seventy fourteen. 369 00:21:23.640 --> 00:21:27.880 Right, yeah, yeah, So I mean, just go with a biplane. 370 00:21:28.000 --> 00:21:32.119 There you go. Tom Cruise could do it. But in it, 371 00:21:32.920 --> 00:21:35.359 you know, it's it's a bit of a cynicism or 372 00:21:35.400 --> 00:21:38.240 a cynical view of the need for technology. It's like, hey, 373 00:21:38.680 --> 00:21:41.359 technology is a perflus. Humans can do it, but we 374 00:21:41.400 --> 00:21:45.319 know that's I don't. I don't question Tom Cruise a lot. 375 00:21:45.440 --> 00:21:49.720 I don't ever question Tom Cruise. I'm just kidding, But 376 00:21:49.799 --> 00:21:51.480 I guess the question is how do we get humans 377 00:21:52.400 --> 00:21:54.279 to be able to trust the machines because there is 378 00:21:54.279 --> 00:21:56.920 a lot of stories. For example, we just recently had 379 00:21:56.920 --> 00:22:00.680 a conversation where where a pilot patch pilots we're given 380 00:22:00.720 --> 00:22:02.440 new technology and they're like, we're not going to use 381 00:22:02.440 --> 00:22:04.960 it because we don't really trust it, So they use it, 382 00:22:05.000 --> 00:22:06.759 And so how do you get the how do you 383 00:22:06.799 --> 00:22:09.160 when new technology is implemented, we have to be able 384 00:22:09.200 --> 00:22:10.920 to trust, especially if it's going to be the difference 385 00:22:11.000 --> 00:22:14.160 maker to win. So how do we do that? How 386 00:22:14.160 --> 00:22:16.359 do we build the trust between the human and the machine? 387 00:22:17.559 --> 00:22:19.720 Well, I don't, I don't. I think we shouldn't just 388 00:22:19.839 --> 00:22:23.359 automatically trust these things. I mean, I think you want 389 00:22:23.839 --> 00:22:26.720 you want to test it out. You're gonna do a 390 00:22:26.759 --> 00:22:31.079 lot of testing and see how it actually works in 391 00:22:31.119 --> 00:22:33.759 a conflict that's small scale, and then scale it up 392 00:22:33.759 --> 00:22:40.440 if if it's effective. But yeah, I mean I have 393 00:22:40.480 --> 00:22:42.920 to say like, I'm not sure, for example, that there 394 00:22:42.960 --> 00:22:46.839 is a unfortunately this is not air force gathering, but 395 00:22:47.279 --> 00:22:49.759 there's there's not No, I'm not sure there's a lot 396 00:22:49.799 --> 00:22:55.839 of opportunity for fighter pilots because I think the if 397 00:22:55.880 --> 00:23:00.319 you've got a drown swarm coming at you, well, what 398 00:23:01.119 --> 00:23:03.920 the pilots of liability in the fire plane? To be honest, 399 00:23:05.799 --> 00:23:12.480 So you know, if you say that you compare a 400 00:23:12.559 --> 00:23:15.559 drone versus a fighter plane, how easy is it to 401 00:23:15.559 --> 00:23:16.160