1 00:00:00,000 --> 00:00:02,877 But, you've been saying recently that, you know, one of the things that's 2 00:00:02,960 --> 00:00:05,037 inevitable is some kind of an AI 3 00:00:05,120 --> 00:00:07,877 disaster, and we're hoping that it's 4 00:00:07,960 --> 00:00:10,037 like 100 people that die, 5 00:00:10,120 --> 00:00:15,037 and not 1,000 or 10,000 or a million or or even 100 million. 6 00:00:15,120 --> 00:00:18,717 Uh but, you know, suppose that it does play out that there's a catastrophe 7 00:00:18,800 --> 00:00:23,157 that's a wake-up call. What do you want to do with that wake-up 8 00:00:23,240 --> 00:00:26,397 call? Um the one I'm particularly worried about is biological, and it goes 9 00:00:26,480 --> 00:00:30,397 something like this. You take some existing pathogen, and using biological 10 00:00:30,480 --> 00:00:34,317 techniques, which I won't discuss, you can modify it enough that it cannot be 11 00:00:34,400 --> 00:00:37,197 detected, but it's still quite dangerous. That's an example of a 12 00:00:37,280 --> 00:00:42,117 threat. Guys, Eric Schmidt just gave a pretty shocking interview about AI. In 13 00:00:42,200 --> 00:00:45,757 this interview, he talked about where AI is headed right now, and how dangerous 14 00:00:45,840 --> 00:00:49,637 it's it's going to get in the next 2 years. He warned that if governments of 15 00:00:49,720 --> 00:00:52,077 the US and China don't put any 16 00:00:52,160 --> 00:00:54,517 safeguards around AI, it's going to take 17 00:00:54,600 --> 00:00:59,277 over literally everything. Eric also talked about AI disasters, which could 18 00:00:59,360 --> 00:01:03,597 destroy our world if our governments don't regulate this industry right now. 19 00:01:03,680 --> 00:01:07,197 Guys, the audio of this interview is a little bit choppy in the first 2 20 00:01:07,280 --> 00:01:11,277 minutes, but I guarantee that this interview is a real mind-opener. Let's 21 00:01:11,360 --> 00:01:14,077 watch the interview, and I'll explain everything as we go. 22 00:01:14,160 --> 00:01:17,197 >> So, are we going to win the AI race, and you know, what are the scenarios where 23 00:01:17,280 --> 00:01:21,957 we win and lose? It damn looks like we will. And uh let 24 00:01:22,040 --> 00:01:25,917 me define the way And so, I think the the separate that's 25 00:01:26,000 --> 00:01:28,837 so consensus, which is I call it, is 26 00:01:28,920 --> 00:01:33,037 what people in separate that's so believe, which charitably was you. 27 00:01:33,120 --> 00:01:34,037 Uh 28 00:01:34,120 --> 00:01:37,637 Kids, that I you only just be built from 29 00:01:37,720 --> 00:01:42,317 current magnetic computing to various forms of recursive 30 00:01:42,400 --> 00:01:43,837 self-improvement 31 00:01:43,920 --> 00:01:47,397 to eventual AGI and superintelligence. 32 00:01:47,480 --> 00:01:49,597 Um in order to do that, it requires that 33 00:01:49,680 --> 00:01:52,357 you use the enormous amount of hardware. 34 00:01:52,440 --> 00:01:56,357 Uh Google TPUs, the biggest chips, and so when everybody in the audience knows 35 00:01:56,440 --> 00:01:59,677 us that it sure looks like the hardware 36 00:01:59,760 --> 00:02:02,317 restrictions that the Trump and Biden 37 00:02:02,400 --> 00:02:05,037 administrations have put on China are 38 00:02:05,120 --> 00:02:09,477 going to prevent them from competing at that space. I was recently in Shanghai 39 00:02:09,560 --> 00:02:12,957 for a few days. I have good relationships with the Chinese. 40 00:02:13,040 --> 00:02:16,477 And my conclusion is they're they're they're fighting a different game. 41 00:02:16,560 --> 00:02:18,797 They're regular about AI in every 42 00:02:18,880 --> 00:02:21,877 product, every service, everything, but 43 00:02:21,960 --> 00:02:26,317 in a more classical way. Whereas America's been a secret for AGI. 44 00:02:26,400 --> 00:02:29,917 And I was quite worried that we would end up in a superintelligence phrase 45 00:02:30,000 --> 00:02:35,037 where you would end up with such enormous gains that one side would have 46 00:02:35,120 --> 00:02:39,397 to actually pre- would have to actually attack the other one. 47 00:02:39,480 --> 00:02:43,717 Uh and you know, there's a preemptive attack. It looks like that fear from me 48 00:02:43,800 --> 00:02:48,037 was not well grounded in fact. I think we're going to be okay for a few years. 49 00:02:48,120 --> 00:02:52,677 Really? What what about robotics? We seem to be pretty far behind in that front. 50 00:02:52,760 --> 00:02:54,757 Yeah, if if if it's okay for me to be 51 00:02:54,840 --> 00:02:58,237 completely blunt, um the Chinese are 52 00:02:58,320 --> 00:03:00,277 doing the same thing in robotics that 53 00:03:00,360 --> 00:03:02,917 they have done in electronic vehicles. 54 00:03:03,000 --> 00:03:05,397 Um in case you're confused while the 55 00:03:05,480 --> 00:03:08,277 current government in the US gets rid of 56 00:03:08,360 --> 00:03:11,437 solar and wind subsidies and promos, 57 00:03:11,520 --> 00:03:14,277 China put in 172 58 00:03:14,360 --> 00:03:16,437 gigawatts of solar last year. I think 59 00:03:16,520 --> 00:03:19,597 it's the number. It's remarkable. 60 00:03:19,680 --> 00:03:22,957 So, the the Chinese race around 61 00:03:23,040 --> 00:03:27,477 solar and EVs, electronic vehicles of one kind or another, looks like they've 62 00:03:27,560 --> 00:03:32,597 won. They're using all of those technologies, in particular new ways of 63 00:03:32,680 --> 00:03:36,677 building stepper motors and other very inexpensive, very powerful, you know, 64 00:03:36,760 --> 00:03:39,157 physical thing. Good example is Unitree 65 00:03:39,240 --> 00:03:42,957 just launched the R1 uh for $6,000 66 00:03:43,040 --> 00:03:44,997 uh available in December. I'm I've 67 00:03:45,080 --> 00:03:47,557 ordered one. We'll see how good it is. 68 00:03:47,640 --> 00:03:50,437 But the arrival of humanoid robots um is 69 00:03:50,520 --> 00:03:54,837 likely to be dominated by China. I'm not suggesting there won't be areas where 70 00:03:54,920 --> 00:03:58,957 the US will play. The US will still have spaces of very 71 00:03:59,040 --> 00:04:03,317 high-end, very sophisticated stuff. Our software is so much better than the 72 00:04:03,400 --> 00:04:06,917 Chinese software, but at the hardware level, I think you should assume that 73 00:04:07,000 --> 00:04:09,877 the world be awash in inexpensive 74 00:04:09,960 --> 00:04:12,317 Chinese robots. In the same sense that 75 00:04:12,400 --> 00:04:15,157 they'll be awash in inexpensive Chinese 76 00:04:15,240 --> 00:04:19,677 electric vehicles. Who knew? And so this is going to be my last question, 77 00:04:19,760 --> 00:04:23,317 but let me bump it to the front since you just beautifully segued into it. 78 00:04:23,400 --> 00:04:28,277 In the race for AGI, but then also in parallel robotics, 79 00:04:28,360 --> 00:04:31,677 China's going to be miles ahead in electricity production. 80 00:04:31,760 --> 00:04:34,397 And in the very short term, we're all going to be chip constrained, TPU 81 00:04:34,480 --> 00:04:38,917 constrained, but if you look 3-4 years out, they know that the fabs are running 82 00:04:39,000 --> 00:04:41,837 at full throttle, the chips are coming out by the millions, 83 00:04:41,920 --> 00:04:44,797 then you're suddenly electricity constrained. 84 00:04:44,880 --> 00:04:47,917 Is there a vulnerability for America there? 85 00:04:48,000 --> 00:04:50,517 It's a huge issue. So, again, let's use 86 00:04:50,600 --> 00:04:53,157 China versus the US as a metaphor. 87 00:04:53,240 --> 00:04:57,117 What are China's strengths? And I'm not praising China. I'm just be trying to 88 00:04:57,200 --> 00:05:01,477 report it. They have solved their electric power problem. 89 00:05:01,560 --> 00:05:05,077 They also have full control over social media, so they don't have the kind of 90 00:05:05,160 --> 00:05:07,597 problems that people here complain about. 91 00:05:07,680 --> 00:05:10,517 They have enormously talented software 92 00:05:10,600 --> 00:05:14,637 people, and they don't have enough hardware. 93 00:05:14,720 --> 00:05:16,357 I think that's roughly where they are. 94 00:05:16,440 --> 00:05:19,197 They're also incredible capitalists. 95 00:05:19,280 --> 00:05:24,117 They they call it Chinese Chinese socialism with 96 00:05:24,200 --> 00:05:28,837 socialism with Chinese characteristics, but trust me, it's pure 97 00:05:28,920 --> 00:05:32,157 raw capitalism. Let's call it what it is. 98 00:05:32,240 --> 00:05:36,357 In the US, we have the many benefits everybody understands. We do not have 99 00:05:36,440 --> 00:05:40,717 enough electricity, and our hard and our our at least in our consumer stuff, the 100 00:05:40,800 --> 00:05:42,917 Chinese are likely to beat us. 101 00:05:43,000 --> 00:05:44,717 Our hardware architectures are 102 00:05:44,800 --> 00:05:48,197 fantastic, and I'm including the Amazon 103 00:05:48,280 --> 00:05:50,437 chip, obviously the Nvidia chips, TPU 104 00:05:50,520 --> 00:05:55,437 V5, which I'm happy to say I was part of TPU V version 1. 105 00:05:55,520 --> 00:05:59,037 All of this stuff is incredible. In the following part, Eric talks about the 106 00:05:59,120 --> 00:06:03,237 lack of energy production in the United States. He says that the US is going to 107 00:06:03,320 --> 00:06:07,237 fall behind China if we don't generate enough electricity to pursue our AI 108 00:06:07,320 --> 00:06:09,517 goals. Guys, in my opinion, the way 109 00:06:09,600 --> 00:06:14,317 China is moving right now, I feel like we've already lost the AI race. I mean, 110 00:06:14,400 --> 00:06:18,677 the amount of electricity we produce is nowhere close to China. China has been 111 00:06:18,760 --> 00:06:22,437 upgrading its grids massively just to meet the energy demands of its data 112 00:06:22,520 --> 00:06:24,757 centers. So, I mean, the real question 113 00:06:24,840 --> 00:06:27,557 is, can the US build power plants fast 114 00:06:27,640 --> 00:06:32,277 enough to compete in the AI race? The answer, in my opinion, is no. So, if you 115 00:06:32,360 --> 00:06:34,397 think about it, and I testified in 116 00:06:34,480 --> 00:06:39,357 Congress a month or two ago on this, we looked at the amount of 117 00:06:39,440 --> 00:06:40,997 electricity required in the United 118 00:06:41,080 --> 00:06:45,757 States to power the expected demand 119 00:06:45,840 --> 00:06:48,517 of data centers, and we needed, in our 120 00:06:48,600 --> 00:06:51,837 calculation, by 2030, 121 00:06:51,920 --> 00:06:55,197 92 gigawatts to be built. And for 122 00:06:55,280 --> 00:07:00,357 reference, a big nuclear power plant is somewhere between 1 and 1 and 1/2 123 00:07:00,440 --> 00:07:03,357 gigawatts. To give you a sense of how many nuclear 124 00:07:03,440 --> 00:07:07,157 power plants are getting started in America, effectively zero. 125 00:07:07,240 --> 00:07:10,997 So, we had hoped, and I've hoped in my lobbying 126 00:07:11,080 --> 00:07:15,757 and test testifying that the government would fast track 127 00:07:15,840 --> 00:07:17,557 availability of all kinds of 128 00:07:17,640 --> 00:07:20,597 electricity. And indeed, they have 129 00:07:20,680 --> 00:07:25,077 promoted oil and gas, but they've also hobbled 130 00:07:25,160 --> 00:07:27,237 solar and wind to a terrible degree, 131 00:07:27,320 --> 00:07:32,237 which is an error. The country needs more energy. 132 00:07:32,320 --> 00:07:35,677 And if we don't get more energy, we're not going to be able to fully exploit 133 00:07:35,760 --> 00:07:39,517 the lead we have in AI and AGI. It's very clear. 134 00:07:39,600 --> 00:07:42,837 And by the way, the obvious next question is, what do we do? 135 00:07:42,920 --> 00:07:46,877 Well, there are scenarios. For example, the president went to Saudi and the UAE 136 00:07:46,960 --> 00:07:49,597 and did huge deals for multiple gigawatts. 137 00:07:49,680 --> 00:07:53,637 And so we might find ourselves in a situation where our training for our 138 00:07:53,720 --> 00:07:57,877 most important thing, the thing which are the the essence of America, which is 139 00:07:57,960 --> 00:08:02,677 American intelligence, is actually being developed in kingdoms. And that may be 140 00:08:02,760 --> 00:08:06,117 the only fallback we have. That is really weird, and that begs a very 141 00:08:06,200 --> 00:08:10,117 difficult question. Do you mind if I ask you a a tough one? 142 00:08:10,200 --> 00:08:15,237 Go ahead. Uh I've been, you know, uh kind of uh inspired by you for most of 143 00:08:15,320 --> 00:08:20,477 my adult life. And uh I I really have have been 144 00:08:20,560 --> 00:08:23,677 watching your uh podcast where you're talking about, look, 145 00:08:23,760 --> 00:08:26,597 when you you remember you when you guest lectured Erik Brynjolfsson's class over 146 00:08:26,680 --> 00:08:28,957 at Stanford, actually you said when you were running Google you felt like you 147 00:08:29,040 --> 00:08:32,477 made decisions three times faster than any company on the planet. But then when 148 00:08:32,560 --> 00:08:36,117 you got into the federal government, you felt like the decision making was 1/3 as 149 00:08:36,200 --> 00:08:39,877 fast as even a slow company. So from your point of view, it's like, you know, 150 00:08:39,960 --> 00:08:42,357 a tiny fraction of the pace that we need to move. 151 00:08:42,440 --> 00:08:44,877 >> Okay, next, Erik talks about some pretty 152 00:08:44,960 --> 00:08:48,797 serious stuff. He says we're going to see a lot of AI disasters if the 153 00:08:48,880 --> 00:08:52,797 governments around the world don't regulate AI companies. He warned that 154 00:08:52,880 --> 00:08:55,437 the pace at which AI is advancing, it 155 00:08:55,520 --> 00:08:58,437 has the potential to topple democracies. 156 00:08:58,520 --> 00:09:02,557 Guys, in my opinion, I believe that every major government around the world 157 00:09:02,640 --> 00:09:05,237 knows that AI is both a risk and a 158 00:09:05,320 --> 00:09:09,717 competitive advantage. If one country slows down, the other won't. That's why 159 00:09:09,800 --> 00:09:13,877 major players like the US and China are not putting any guardrails around their 160 00:09:13,960 --> 00:09:16,557 AI companies. They want that AI edge. 161 00:09:16,640 --> 00:09:19,477 >> But you've been saying recently that, you know, one of the things that's 162 00:09:19,560 --> 00:09:21,677 inevitable is some kind of an AI 163 00:09:21,760 --> 00:09:24,477 disaster, and we're hoping that it's 164 00:09:24,560 --> 00:09:26,677 like 100 people that die, 165 00:09:26,760 --> 00:09:31,637 and not 1,000 or 10,000 or a million or or even 100 million. 166 00:09:31,720 --> 00:09:35,317 Uh but, you know, suppose that it does play out that there's a catastrophe 167 00:09:35,400 --> 00:09:39,757 that's a wake-up call, what do you want to do with that wake-up 168 00:09:39,840 --> 00:09:44,397 call? What what's the next move for Eric Schmidt after the wake-up call? 169 00:09:44,480 --> 00:09:48,797 Well, so so the background here is that uh Dr. Kissinger, Henry, and I spent an 170 00:09:48,880 --> 00:09:52,197 awful lot of time talking about the period in the 1950s where he was a key 171 00:09:52,280 --> 00:09:54,757 component of all of these things. 172 00:09:54,840 --> 00:10:00,037 And the So so what he did was he he used 173 00:10:00,120 --> 00:10:02,837 the fact that we had used the nuclear bomb 174 00:10:02,920 --> 00:10:07,037 to negotiate over about a 15-year period 175 00:10:07,120 --> 00:10:08,837 a set of treaties that restricted 176 00:10:08,920 --> 00:10:14,557 nuclear proliferation. Those treaties when they were um 177 00:10:14,640 --> 00:10:17,997 uh when they were negotiated have allowed 178 00:10:18,080 --> 00:10:22,797 us to be alive today. So these were centrally important. Without but without 179 00:10:22,880 --> 00:10:26,557 controlling the spread of enriched uranium and the other secrets, we would 180 00:10:26,640 --> 00:10:28,517 all be toast, literally, because of 181 00:10:28,600 --> 00:10:31,157 crazy people and so forth. Is there an 182 00:10:31,240 --> 00:10:34,237 analogous set of things that we can do? 183 00:10:34,320 --> 00:10:38,197 The problem here is uh I guess the good news is we're not in a war, we haven't 184 00:10:38,280 --> 00:10:41,877 had a nuclear bomb, we don't have that thing to discuss. So we can talk about 185 00:10:41,960 --> 00:10:45,877 it, but governments tend to act reactively. So let me talk There are 186 00:10:45,960 --> 00:10:49,917 three obvious threats right now, which I think are fairly well understood. The 187 00:10:50,000 --> 00:10:53,557 first is misinformation and the software that we're all 188 00:10:53,640 --> 00:10:56,197 collectively giving people 189 00:10:56,280 --> 00:11:00,677 um allows for all sorts of misinformation, fake videos, fake news, 190 00:11:00,760 --> 00:11:02,997 what have you. We all understand this. 191 00:11:03,080 --> 00:11:04,917 It's all open source, that's done. 192 00:11:05,000 --> 00:11:08,477 That's a threat to democracies and maybe to dictatorships, certainly to 193 00:11:08,560 --> 00:11:11,517 democracies. The second one is cyber. 194 00:11:11,600 --> 00:11:15,117 And I think one way to understand cyber is that if you can write code, you can 195 00:11:15,200 --> 00:11:17,757 also write cyber cyber attacks. It's the 196 00:11:17,840 --> 00:11:19,757 same logic. And you have these 197 00:11:19,840 --> 00:11:22,837 incredible gains in software. It's 198 00:11:22,920 --> 00:11:27,077 frightening how good these soft Remember, my career is as a programmer. 199 00:11:27,160 --> 00:11:28,517 These things program better than I ever 200 00:11:28,600 --> 00:11:31,797 did. It's like shocking, right? And then 201 00:11:31,880 --> 00:11:34,117 the third one is bio. And I think most 202 00:11:34,200 --> 00:11:36,997 people believe that one of those three 203 00:11:37,080 --> 00:11:40,117 will create some kind of mini crisis 204 00:11:40,200 --> 00:11:44,637 that will then cause the governments to say, "Hang on. Let's have a conversation 205 00:11:44,720 --> 00:11:49,117 about how to really deal with the downsize." The upsides are incredible. 206 00:11:49,200 --> 00:11:52,517 Right? And I want America to win, and I want us to run as fast as we can, and 207 00:11:52,600 --> 00:11:56,357 we're indeed doing that with the Trump administration, which is great. 208 00:11:56,440 --> 00:12:00,757 Right? But we have to be aware that these possi- these things are possible. 209 00:12:00,840 --> 00:12:03,517 Um the one I'm particularly worried about is biological, and it goes 210 00:12:03,600 --> 00:12:07,477 something like this. You take some existing pathogen, and using biological 211 00:12:07,560 --> 00:12:11,437 techniques, which I won't discuss, you can modify it enough that it cannot be 212 00:12:11,520 --> 00:12:14,317 detected, but it's still quite dangerous. That's an example of a 213 00:12:14,400 --> 00:12:16,157 threat. There are many others. 214 00:12:16,240 --> 00:12:20,637 Yeah, I that's also the easiest, I think, which is scary. 215 00:12:20,720 --> 00:12:24,517 Um of the of the immediate, you know, CBRN threats, threats, so cyber, 216 00:12:24,600 --> 00:12:26,837 biological, radiological, nuclear. 217 00:12:26,920 --> 00:12:29,957 Uh biological is the one you can kind of do in a basement with three people, but 218 00:12:30,040 --> 00:12:33,197 it's also the one that's hardest to contain. So, 219 00:12:33,280 --> 00:12:35,037 anyway, uh I hope the theory is right. 220 00:12:35,120 --> 00:12:37,557 So, so then uh how do you deal with 221 00:12:37,640 --> 00:12:42,637 proliferation? You know, the Biden approach was, "Okay, let's contain AGI 222 00:12:42,720 --> 00:12:46,437 to these five companies, you know, Google being one of them, and then let's 223 00:12:46,520 --> 00:12:51,237 say any model with over 1E26 training flops has to register with the federal 224 00:12:51,320 --> 00:12:54,717 government, and then we'll keep it all contained." 225 00:12:54,800 --> 00:12:57,277 Uh so, that all got scrapped immediately, you know, after the 226 00:12:57,360 --> 00:13:01,637 election, and got replaced by the new David Sacks document, which the David 227 00:13:01,720 --> 00:13:06,077 Sacks document is much more about how do we move as quickly as possible and and 228 00:13:06,160 --> 00:13:09,357 win the race. Um but it doesn't really address 229 00:13:09,440 --> 00:13:11,357 proliferation, and obviously in America, 230 00:13:11,440 --> 00:13:13,957 we want startups uh and researchers to 231 00:13:14,040 --> 00:13:19,037 have incredible access to technology and compute. On the other hand, you know, 232 00:13:19,120 --> 00:13:23,517 the three people in a basement making a biological weapon is is a real scary 233 00:13:23,600 --> 00:13:25,837 thing. So, how do you balance those? 234 00:13:25,920 --> 00:13:29,557 Look, it turns out if you read the David Sacks President Trump announcement, 235 00:13:29,640 --> 00:13:31,277 they're very clear that they want to 236 00:13:31,360 --> 00:13:35,117 continue to study the security aspects 237 00:13:35,200 --> 00:13:37,237 of AI, especially in a geopolitical way. 238 00:13:37,320 --> 00:13:40,357 And that's code for China versus the US. 239 00:13:40,440 --> 00:13:44,557 And so they continue in their proposal to fund things involving nation-state 240 00:13:44,640 --> 00:13:46,797 attacks and so forth and so on. And I 241 00:13:46,880 --> 00:13:49,317 fully support that kind of stuff. 242 00:13:49,400 --> 00:13:54,037 Um there's probably a difference on the misinformation social media stuff 243 00:13:54,120 --> 00:13:57,837 between the two. But for example, the Biden rule was 244 00:13:57,920 --> 00:14:01,997 simply that you had to report if you're doing a a training run 10 to the 26 or 245 00:14:02,080 --> 00:14:05,677 greater. I was part of the group that made that number up. Uh we made it up 246 00:14:05,760 --> 00:14:07,557 cuz we had no better number. Uh I'm not 247 00:14:07,640 --> 00:14:09,597 suggesting it's the right number. 248 00:14:09,680 --> 00:14:14,077 And the the conclusion you come to is 249 00:14:14,160 --> 00:14:18,837 it's probably the case that that we know 250 00:14:18,920 --> 00:14:22,557 our government, I don't know this, but I'm guessing, knows where the training 251 00:14:22,640 --> 00:14:25,157 runs are going on in in China and 252 00:14:25,240 --> 00:14:27,037 because of espionage. And it's probably 253 00:14:27,120 --> 00:14:30,477 the case that the Chinese have espionage 254 00:14:30,560 --> 00:14:32,277 on the US knowing where our training 255 00:14:32,360 --> 00:14:34,757 runs are. So I'm not sure the nature of 256 00:14:34,840 --> 00:14:37,757 the training run runs is a secret. And 257 00:14:37,840 --> 00:14:39,877 frankly, everybody knows where the data 258 00:14:39,960 --> 00:14:42,597 centers are cuz they're immense. So 259 00:14:42,680 --> 00:14:47,477 uh we will see. Another attack on the 10 26 is that the training is getting more 260 00:14:47,560 --> 00:14:50,637 um more cost-efficient. If you look at the 261 00:14:50,720 --> 00:14:54,917 moving they move from something called FP16 to FP8, 262 00:14:55,000 --> 00:14:58,037 that means eight precision floating point. People are now moving to four 263 00:14:58,120 --> 00:14:59,357 four bit floating point, which is 264 00:14:59,440 --> 00:15:02,117 bizarre. It turns out these trainings 265 00:15:02,200 --> 00:15:07,517 algorithms seem to be quite tolerant for floating point in in precision, which is 266 00:15:07,600 --> 00:15:10,917 a shock to me. So again, we're getting more efficient 267 00:15:11,000 --> 00:15:15,277 in training. This creates more of a proliferation problem. Okay, next Eric 268 00:15:15,360 --> 00:15:20,117 talks about the threat from open-source AI. Guys, I don't agree with Eric here. 269 00:15:20,200 --> 00:15:24,037 I believe these tech CEOs are going after open-source models because they 270 00:15:24,120 --> 00:15:25,917 just want to protect their investments. 271 00:15:26,000 --> 00:15:29,957 Their goal here is to criminalize open-source AI. I don't think 272 00:15:30,040 --> 00:15:33,797 open-source is the real threat. The real threat is that one of the big 273 00:15:33,880 --> 00:15:38,037 corporations will eventually train a super-intelligent AI with the wrong 274 00:15:38,120 --> 00:15:42,237 terminal goal. If I can If I can say in general the proliferation issue, 275 00:15:42,320 --> 00:15:45,837 I'm not worried about big companies and big countries 276 00:15:45,920 --> 00:15:49,797 because I can count them. You know, there'll be 10 huge data 277 00:15:49,880 --> 00:15:52,677 centers, 10 huge training runs around the world. 278 00:15:52,760 --> 00:15:55,877 I'm much more worried about the open-source groups 279 00:15:55,960 --> 00:15:59,077 which can operate in the shadows. They don't have to solve every problem. They 280 00:15:59,160 --> 00:16:02,997 just have to do one thing well. They can patch together the open-source which is 281 00:16:03,080 --> 00:16:05,237 generally available and is good enough. 282 00:16:05,320 --> 00:16:07,357 If you look at the quality of uh deep 283 00:16:07,440 --> 00:16:11,637 seek R R2 R1 and now R2 coming, sure 284 00:16:11,720 --> 00:16:16,677 looks like it's in 80% or 90% of these top closed models. The closed models 285 00:16:16,760 --> 00:16:20,717 people, which obviously includes Google, get very defensive over this because 286 00:16:20,800 --> 00:16:23,917 they say, "Look, those are kind of synthetic. They have diffused the 287 00:16:24,000 --> 00:16:27,677 models. They've trained our best information." They're not All of that is 288 00:16:27,760 --> 00:16:32,077 true. They're correct. But they are nevertheless useful for specific things 289 00:16:32,160 --> 00:16:34,197 and they could be used in a 290 00:16:34,280 --> 00:16:39,237 proliferation issue to do various forms of cyber and biological attacks. Well, 291 00:16:39,320 --> 00:16:42,437 yeah, the current rule of thumb is that distillation and transfer learning is 292 00:16:42,520 --> 00:16:46,717 about 1% of the cost of the original training but gets you to the same 293 00:16:46,800 --> 00:16:50,277 destination. So, if you take, you know, FP4 which you mentioned, you know, you 294 00:16:50,360 --> 00:16:55,357 get an 8x performance boost over FP32. So, you get an 8x on the on the, 295 00:16:55,440 --> 00:16:58,197 you know, the the weights and then you've got another 296 00:16:58,280 --> 00:17:02,357 100x on the transfer learning and it's really hard to draw an analogy to 297 00:17:02,440 --> 00:17:06,957 nuclear or chemical weapons in the past cuz you couldn't kind of shrink them a 298 00:17:07,040 --> 00:17:09,237 thousand x under the covers and get the 299 00:17:09,320 --> 00:17:12,077 same result. But AI is really weird that 300 00:17:12,160 --> 00:17:14,317 way. It's It's very compressible, very 301 00:17:14,400 --> 00:17:16,317 fluid. 302 00:17:16,400 --> 00:17:19,197 Uh well, again, we we we will 303 00:17:19,280 --> 00:17:21,717 It does not look like we have a very 304 00:17:21,800 --> 00:17:24,517 good solution for distillation. 305 00:17:24,600 --> 00:17:27,917 It looks like opponent of a company 306 00:17:28,000 --> 00:17:32,837 can mask their queries to look like normal sets of users and then distill 307 00:17:32,920 --> 00:17:36,397 the models. One of my friends has thought about this 308 00:17:36,480 --> 00:17:40,437 a lot and he thinks that the eventual state in the United 309 00:17:40,520 --> 00:17:43,317 States is that the biggest models will 310 00:17:43,400 --> 00:17:45,317 never be released. 311 00:17:45,400 --> 00:17:47,557 And that the companies will distill 312 00:17:47,640 --> 00:17:50,717 their own models down for that reason. 313 00:17:50,800 --> 00:17:53,757 That's his opinion. We'll see if that's true. 314 00:17:53,840 --> 00:17:58,477 And this produces a bizarre outcome where the biggest models in the United 315 00:17:58,560 --> 00:18:03,237 States are closed source and the biggest models in China are open 316 00:18:03,320 --> 00:18:05,997 source. And the geopolitical issue there, of 317 00:18:06,080 --> 00:18:08,597 course, is that open source is free and 318 00:18:08,680 --> 00:18:10,877 the closed source models are not free. 319 00:18:10,960 --> 00:18:14,597 And so the vast majority of governments and countries who don't have the kind of 320 00:18:14,680 --> 00:18:18,397 money that the West does and so forth will end up standardizing on Chinese 321 00:18:18,480 --> 00:18:22,197 models, not because they're better, but because they're free. And so we'll see 322 00:18:22,280 --> 00:18:25,317 if that's true or not, but I do worry about that. All right, guys. That's it 323 00:18:25,400 --> 00:18:30,320 for today's video. I'll see you guys again next week with another one.