WEBVTT 1 00:00:00.840 --> 00:00:03.439 You know, what I've said is that we really have 2 00:00:03.680 --> 00:00:08.400 here rule of the bureaucracy as opposed to rule of 3 00:00:08.439 --> 00:00:09.599 the people democracy. 4 00:00:10.119 --> 00:00:15.160 So in order we want to restore the rule of 5 00:00:15.160 --> 00:00:15.519 the people. 6 00:00:15.519 --> 00:00:17.879 And so what that means is reducing the size of 7 00:00:17.920 --> 00:00:26.519 the federal government, uh, basically reducing regulation. You know, there's 8 00:00:26.519 --> 00:00:29.199 there's a there's a tremendous amount of of overregulation that's 9 00:00:29.199 --> 00:00:32.640 happened over time. And this is this isn't an inevitable 10 00:00:33.479 --> 00:00:36.560 consequence of a long period of prosperity, is. 11 00:00:36.520 --> 00:00:40.880 That you're going to get more and more rules and regulations. 12 00:00:40.920 --> 00:00:44.840 More laws accumulate over time, and the normal forcing function 13 00:00:45.079 --> 00:00:48.119 for getting rid of rules and regulations is war. So 14 00:00:48.439 --> 00:00:50.759 it needs to be some kind of existential war where 15 00:00:50.920 --> 00:00:57.600 you have to do a reset in order to avoid 16 00:00:57.640 --> 00:01:00.359 being defeated in a war. This is literally the throughout 17 00:01:00.359 --> 00:01:03.520 history has been the main forcing function for clearing out 18 00:01:03.640 --> 00:01:07.200 an accumulation of laws and regulation. In the absence of that, 19 00:01:08.280 --> 00:01:11.239 every year you get more laws and regulations, until eventually 20 00:01:11.239 --> 00:01:14.599 everything is illegal and nothing is permitted. And that's sort 21 00:01:14.599 --> 00:01:21.879 of the situation we have these days. So the aspiration 22 00:01:21.959 --> 00:01:27.879 here there is a reduction in regulation and reduction in 23 00:01:27.920 --> 00:01:33.040 government spending such that the economy is able to grow faster. 24 00:01:34.719 --> 00:01:37.280 Maybe the economy can grow at four or five percent 25 00:01:37.359 --> 00:01:41.599 potentially in terms of real useful goods and services output, 26 00:01:42.239 --> 00:01:48.000 and then government spending can be reproduced by about three 27 00:01:48.079 --> 00:01:50.239 or four percent of the economy, about maybe a trillion 28 00:01:50.280 --> 00:01:53.000 dollars or more. And the net effect of that would 29 00:01:53.040 --> 00:01:58.120 be no inflation from twenty five to twenty twenty six. 30 00:01:58.959 --> 00:02:00.799 So quite remarkable. 31 00:02:01.079 --> 00:02:03.480 And also, if the US government is buying less debt, 32 00:02:03.480 --> 00:02:05.920 which I think will be the case, if it's if 33 00:02:05.920 --> 00:02:07.879 the depths it drops from two trillion to one trillion, 34 00:02:08.919 --> 00:02:10.759 then there will be one trillion. 35 00:02:10.439 --> 00:02:12.719 Less debt the governor's spy, which will drop. Of course, 36 00:02:12.840 --> 00:02:13.680 interest rates. 37 00:02:13.479 --> 00:02:21.680 To drop significantly, and that means people's mortgage payments, car payments, 38 00:02:21.879 --> 00:02:25.599 credit card payments, student loans, whatever debt they have, will 39 00:02:25.960 --> 00:02:28.400 their debt payments will be less. So I think this 40 00:02:28.479 --> 00:02:33.360 is something that will benefit the average American. I think 41 00:02:33.560 --> 00:02:35.879 some of the things we're doing also will be helpful 42 00:02:35.960 --> 00:02:40.400 to hopefully helpful to other countries because with the new administration, 43 00:02:40.879 --> 00:02:46.360 there's less interested in interfering with the affairs of other countries. 44 00:02:47.560 --> 00:02:52.319 You know, I think a lot of the times the 45 00:02:52.439 --> 00:02:53.639 United States has been kind. 46 00:02:53.479 --> 00:02:59.639 Of pushy in international affairs, which may resonate with a 47 00:02:59.680 --> 00:03:04.479 number of the audience. And I think we should in 48 00:03:04.599 --> 00:03:08.560 general leave other countries to their own business and basically 49 00:03:09.120 --> 00:03:15.000 America's mind its own businesses, you know, rather than push 50 00:03:15.080 --> 00:03:16.879 forward regime change all over the place. 51 00:03:18.719 --> 00:03:23.800 So yeah, it was probably good by the countries too. 52 00:03:24.520 --> 00:03:26.319 So instead of waiting for the war to happened, you 53 00:03:26.360 --> 00:03:28.520 went to war against the bureaucracy in the government. 54 00:03:30.039 --> 00:03:35.800 Yes, we're well, we're essentially just we're with you know, 55 00:03:37.400 --> 00:03:42.960 the support and direction of the President Trump. We are 56 00:03:43.360 --> 00:03:48.560 reducing the size of the bureaucracy, getting rid of excess 57 00:03:48.599 --> 00:03:54.879 regulatory regulations. And there's also so many agencies and regulatory 58 00:03:54.879 --> 00:03:57.719 authorities that they're actually stepped on each other's feet. It's 59 00:03:57.919 --> 00:04:01.000 it's kind of like having a sports game where there 60 00:04:01.000 --> 00:04:03.879 are too many referees on the field, like more referees. 61 00:04:03.520 --> 00:04:04.639 Than players at times. 62 00:04:05.080 --> 00:04:07.520 Now, that would be a silly game, you know, if 63 00:04:07.520 --> 00:04:10.319 the players can't pass the ball without hitting a referee, 64 00:04:10.360 --> 00:04:13.240 you know. But it's kind of getting to that point 65 00:04:13.240 --> 00:04:17.079 in the US. So there's roughly four hundred and fifty 66 00:04:17.639 --> 00:04:22.519 federal agencies of one kind or another. That's that's more agencies. 67 00:04:22.600 --> 00:04:25.079 That's almost that's almost an average of two agencies per 68 00:04:25.160 --> 00:04:29.160 year since the formation of the United States. So, I mean, 69 00:04:29.199 --> 00:04:31.720 how many agencies do you really need to run a country? 70 00:04:32.120 --> 00:04:35.160 I'm ninety nine, not four hundred and fifty, that's for sure. 71 00:04:36.279 --> 00:04:40.759 So, and and how do you guarantee that all the 72 00:04:41.079 --> 00:04:44.399 incredible achievements that you aim to have in terms of savings, 73 00:04:44.439 --> 00:04:47.839 in terms of you know, impacting the lives of the 74 00:04:47.879 --> 00:04:50.560 American people, I'm not going to be reversed in four years. 75 00:04:50.720 --> 00:04:54.439 Typically the cycle gets reversed every four years. You know, 76 00:04:54.519 --> 00:04:56.759 do you think it's going to be so impacted that 77 00:04:56.839 --> 00:04:59.600 it won't be reversed? Is there any ways that you can, 78 00:04:59.759 --> 00:05:01.920 you know, ensure that the progress is going to be continuous? 79 00:05:04.040 --> 00:05:05.519 Well, I think I think we need we do need 80 00:05:05.560 --> 00:05:12.639 to delete entire agencies, as opposed to leave part of 81 00:05:12.680 --> 00:05:14.519 them behind, because if you leave part of them behind, 82 00:05:15.439 --> 00:05:18.360 it's easy. It's kind of like if they're leaving a weed. 83 00:05:18.399 --> 00:05:21.399 If you don't remove the roots of the weed, then 84 00:05:21.639 --> 00:05:23.319 it's easy for the weed to grow back. But if 85 00:05:23.319 --> 00:05:26.360 you remove the roots of the weed, it doesn't stop weeds. 86 00:05:26.160 --> 00:05:27.920 From ever growing. Back, but it makes it harder. 87 00:05:28.680 --> 00:05:32.720 So so we have to really delete entire agencies, many 88 00:05:32.759 --> 00:05:37.199 of them. And that's not to say there won't be 89 00:05:37.240 --> 00:05:41.519 an increase over time of bureaucracy in some new administration, 90 00:05:41.639 --> 00:05:44.959 but it will, it'll be from a much lower baseline. 91 00:05:46.480 --> 00:05:49.639 So it's it's a step in the right direction. I think. Well, 92 00:05:50.040 --> 00:05:51.480 the overwatching goal here is like. 93 00:05:51.600 --> 00:05:55.519 It's to lay the validation for prosperity that will last 94 00:05:55.560 --> 00:06:02.199 many decades, you know, maybe centuries and yeah, but forever. 95 00:06:02.319 --> 00:06:07.120 Nothing's forever. But I think we can strengthen the foundations 96 00:06:07.160 --> 00:06:08.959 of the United States substantially. 97 00:06:10.680 --> 00:06:13.560 And what lessons can other governments learn from us? You 98 00:06:13.600 --> 00:06:17.279 say tech support on your shirt? Is that only technology 99 00:06:17.360 --> 00:06:21.800 or is there other things? How do you approach efficiency? 100 00:06:22.040 --> 00:06:28.519 Well, a shocking percentage of the problem or maybe not 101 00:06:28.560 --> 00:06:32.600 talking for those who know it, but a big percentage 102 00:06:32.680 --> 00:06:36.360 of the problem is improving the technology that the government 103 00:06:36.439 --> 00:06:40.720 runs on. So the US government runs on a collection 104 00:06:40.759 --> 00:06:43.519 of thousands of computers, so many of them antiquated running 105 00:06:43.560 --> 00:06:45.920 very old software that then the computers don't talk to 106 00:06:45.959 --> 00:06:50.720 each other. And so that's why tech support is kind 107 00:06:50.720 --> 00:06:54.160 of a real thing. In order to make the got 108 00:06:54.240 --> 00:06:57.399 more efficient. You have to improve the technology. You may 109 00:06:57.399 --> 00:07:02.040 have read about the example recently with when President Trump 110 00:07:02.079 --> 00:07:07.560 was signing one of the Doge Executive Orders of the 111 00:07:08.600 --> 00:07:13.160 difficulty of US covert worker is retiring, Like the retire 112 00:07:13.279 --> 00:07:17.600 maximum retirement rate is ten thousand a month, and the 113 00:07:17.639 --> 00:07:22.240 reason for that is because the retirement is entirely payerwork. 114 00:07:22.439 --> 00:07:26.199 Right now, it's manually calculated paperwork that's put in an 115 00:07:26.319 --> 00:07:29.240 envelope and then taken down a mind shaft and stored 116 00:07:29.279 --> 00:07:33.519 in a mine. And then the you know, one of 117 00:07:33.560 --> 00:07:36.959 the things that affects the rate at which better wor 118 00:07:36.959 --> 00:07:39.800 workers can retire is the speed of the elevator in 119 00:07:39.879 --> 00:07:43.439 a mine in Pennsylvania, which is bizarre because. 120 00:07:43.199 --> 00:07:45.000 It's not it should be digital, you know. 121 00:07:46.079 --> 00:07:47.920 So then when we said, well, why isn't it digital, 122 00:07:47.959 --> 00:07:51.040 it's they said, well, we have had a digital digitization 123 00:07:51.160 --> 00:07:55.480 program going since twenty fourteen, so that's been eleven years. 124 00:07:56.240 --> 00:07:59.240 So then we asked, well, so what how much progress 125 00:07:59.240 --> 00:08:00.439 have you made? They said B. 126 00:08:01.000 --> 00:08:03.879 I mean you're giving yourself a grade of B. No, 127 00:08:04.079 --> 00:08:08.480 we're on the letter B. So like hmm, okay, we're 128 00:08:08.480 --> 00:08:11.639 going to need to really provide some tech support here, 129 00:08:13.800 --> 00:08:15.439 like otherwise literally people can't. 130 00:08:15.240 --> 00:08:18.839 Even retire, like even if they want to. It's pretty bad. 131 00:08:20.240 --> 00:08:20.439 You know. 132 00:08:20.480 --> 00:08:23.199 There's this is a there's a lot of software systems 133 00:08:23.240 --> 00:08:26.399 that need to be updated and fixed, in some cases deleted. 134 00:08:27.800 --> 00:08:30.079 A lot of things that should really should be automated. 135 00:08:30.120 --> 00:08:34.759 I mean, in terms of the number of say you're 136 00:08:34.840 --> 00:08:37.679 citizens that are operating the mine, it's about a thousand 137 00:08:37.679 --> 00:08:41.279 people are working on this mine where the papers are stored. 138 00:08:41.639 --> 00:08:44.519 But they should they should ideally be working on something else. 139 00:08:44.799 --> 00:08:47.720 That is, they should be working on producing goods and 140 00:08:47.759 --> 00:08:48.679 services that are. 141 00:08:48.559 --> 00:08:51.960 Of much higher value to the public. You know. 142 00:08:53.279 --> 00:08:55.879 So, I mean, really, even even if somebody just grew 143 00:08:55.960 --> 00:08:58.639 tomatoes in their garment and so them at the farmer's market, 144 00:08:58.679 --> 00:09:03.039 that would be more useful than carrying minila envelopes down 145 00:09:03.039 --> 00:09:06.039 in mind shraft, you know, safe to say. 146 00:09:07.320 --> 00:09:09.559 So a lot of stuff is it's like that. 147 00:09:09.399 --> 00:09:12.840 You know, it's not it's not like any one thing 148 00:09:12.919 --> 00:09:16.120 is particularly difficult, but there are like ten thousand things 149 00:09:16.320 --> 00:09:18.200 that you need to be improved. 150 00:09:19.159 --> 00:09:23.399 So it's efficiency through innovation rather than efficiency through housterity 151 00:09:23.519 --> 00:09:26.320 and cost cutting specifically, so you're trying to do both 152 00:09:26.320 --> 00:09:29.360 at the same time. Maybe focus more on tech support 153 00:09:29.720 --> 00:09:30.759 than cutting costs. 154 00:09:32.279 --> 00:09:37.080 Well, by improving the technology, the costs do reduce. So 155 00:09:37.480 --> 00:09:39.240 you know, it's very expensive to have a thousand people 156 00:09:39.320 --> 00:09:44.480 uperate online with during paid for retirement, whereas that really 157 00:09:44.600 --> 00:09:48.159 should just be digitized and be a computer. That's where 158 00:09:48.159 --> 00:09:50.120 the information is stored in the cloud and it's very 159 00:09:50.159 --> 00:09:56.360 straightforward and low costs. So what automation you know, will 160 00:09:56.399 --> 00:10:01.240 help there a lot. And and then but like a 161 00:10:01.240 --> 00:10:04.120 lot of things just really shouldn't exist, you know, they're 162 00:10:04.320 --> 00:10:05.200 kind of vestigial. 163 00:10:08.080 --> 00:10:11.559 You know, We've a lot of attention has been on 164 00:10:11.559 --> 00:10:11.960 the sort. 165 00:10:11.840 --> 00:10:14.320 Of USAID for example, you know when we looked a 166 00:10:14.360 --> 00:10:16.559 lot of those programs, but like. 167 00:10:18.120 --> 00:10:19.600 We should like look, why why. 168 00:10:19.559 --> 00:10:22.159 Does this actually exist? Is there really a need for it? 169 00:10:23.720 --> 00:10:26.799 You know, there's like national damage for democracy. But I'm like, okay, 170 00:10:26.840 --> 00:10:29.720 well how much democracy have they achieved lately? 171 00:10:29.799 --> 00:10:31.200 You know, I don't know, not much. 172 00:10:32.440 --> 00:10:35.480 So you know, the picture they have on the website 173 00:10:35.519 --> 00:10:38.120 is a picture of Reagan and get that's been a while. 174 00:10:38.399 --> 00:10:42.919 You know, that was like the eighties, so like obviously 175 00:10:43.279 --> 00:10:46.240 not opposed to democracy or you know, there's all these 176 00:10:46.279 --> 00:10:50.440 things that get funded, but we're like, well, why why 177 00:10:50.559 --> 00:10:52.639 is this new tax paramoney. I don't think it doesn't 178 00:10:52.639 --> 00:10:57.960 seem like it does. So you know, there's a lot 179 00:10:57.960 --> 00:11:00.960 of sort of pushing DEI worldwide that you know, this 180 00:11:01.200 --> 00:11:04.720 obviously the Trump administration doesn't agree with, and we want 181 00:11:04.759 --> 00:11:10.240 to terminate that stuff, which we are and you know, 182 00:11:10.559 --> 00:11:14.840 make sure the schools focus on improving basic education of kids, 183 00:11:15.759 --> 00:11:17.159 as President Trump said. 184 00:11:16.960 --> 00:11:19.360 I think yesterday, maybe today. 185 00:11:20.480 --> 00:11:23.600 The United States is currently ranked fortieth out of forty 186 00:11:23.799 --> 00:11:28.120 in the OECD for education, which is pretty bad. You know, 187 00:11:28.360 --> 00:11:31.200 you can't you I mean, let's but but not. But 188 00:11:31.240 --> 00:11:33.879 in terms of spending, the United States is spending a 189 00:11:33.879 --> 00:11:39.120 tremendous amount for students but achieving very weak results. So 190 00:11:40.159 --> 00:11:41.879 you know, that's just a case where okay, we need 191 00:11:41.879 --> 00:11:45.399 to spend less money and get better results. It's it's like, 192 00:11:45.440 --> 00:11:47.399 I mean, a lot of it, a lot of you 193 00:11:47.399 --> 00:11:48.759 can think of it's sort of like it's like in 194 00:11:48.799 --> 00:11:53.600 a way, it's like a big company, like a big corporation, 195 00:11:53.639 --> 00:11:58.919 American Incorporated. And you know, just like with Twitter, there 196 00:11:58.960 --> 00:12:00.879 was a lot of stuff that was being done that 197 00:12:01.039 --> 00:12:05.080 was unnecessary. You know, we case of Twitter, we reduce 198 00:12:05.159 --> 00:12:08.000 the staff by eighty percent, but at the same time, 199 00:12:08.080 --> 00:12:12.840 improved the functionality and capabilities of the site dramatically and 200 00:12:12.879 --> 00:12:16.120 accomplish more in a year than they previously accomplished in 201 00:12:16.120 --> 00:12:22.080 five years. So it's like a corporate turnaround, but at 202 00:12:22.120 --> 00:12:28.200 a much larger scale. And you know, we are giving 203 00:12:28.320 --> 00:12:34.360 generous you know, exit packages, like you know, if people retire, 204 00:12:34.399 --> 00:12:38.120 they get paid all the way through September. They can 205 00:12:38.159 --> 00:12:40.200 go on vacation, they can get a second job, they 206 00:12:40.200 --> 00:12:43.679 can do whatever they want. And we can't actually pay 207 00:12:43.679 --> 00:12:47.559 them any more than through September because that's the congressional 208 00:12:47.559 --> 00:12:50.879 appropriation is only through the end of the government financial 209 00:12:50.960 --> 00:12:54.639 year which ends in September. So you know, so I 210 00:12:54.679 --> 00:12:57.639 think there'll be like some disruption, but at the end 211 00:12:57.639 --> 00:13:00.600 of the day, we'll have people move from, like I said, 212 00:13:00.600 --> 00:13:03.720 from load of negative productivity roles to the in the 213 00:13:03.759 --> 00:13:06.639 government sector to higher productivity roles in the private sector. 214 00:13:08.320 --> 00:13:11.759 Can we pivot to our special intelligence And I'm sure 215 00:13:11.759 --> 00:13:15.240 you've been seeing what Deep Sea has done and all 216 00:13:15.360 --> 00:13:19.879 the claimed achievements that they've had. I know that we've 217 00:13:19.879 --> 00:13:22.679 been speaking for a while about GROCK three and that 218 00:13:22.799 --> 00:13:25.320 GROCK three is going to be through the slot and 219 00:13:25.360 --> 00:13:27.559 the ice piece. When are we going to see that 220 00:13:27.720 --> 00:13:30.200 and what capabilities can we expect from GROCK three. 221 00:13:32.919 --> 00:13:37.159 Well, I mean GROCK three, GROG three has very very 222 00:13:37.279 --> 00:13:42.200 very powerful reasoning capabilities. So in the test that we've 223 00:13:42.399 --> 00:13:46.519 done thus far, Grock three is are performing anything that's 224 00:13:46.559 --> 00:13:48.159 been released that we're aware of. 225 00:13:49.679 --> 00:13:52.440 So that's that's a that's a good sign. 226 00:13:54.600 --> 00:13:58.399 Yeah, it's In fact, at times, I think Rock three 227 00:13:58.840 --> 00:14:01.519 is kind of scary smart. You're like, wow, this thing 228 00:14:01.559 --> 00:14:05.879 is smart. It's kind of scary. Growthree is scary. So 229 00:14:05.919 --> 00:14:09.759 it's like, wow, this thing is you know, it comes 230 00:14:09.840 --> 00:14:13.919 up with solutions that you didn't even think were like, 231 00:14:14.000 --> 00:14:20.879 you wouldn't even anticipate, you know, not obvious solutions. So 232 00:14:21.000 --> 00:14:23.879 GROCK three was trained with the most amount of compute 233 00:14:24.000 --> 00:14:28.840 and I think very efficiently trained. Also, notably, GROC three 234 00:14:28.960 --> 00:14:35.159 was trained on a lot of synthetic data, so and 235 00:14:34.960 --> 00:14:36.960 and and then it goes back and forth through the 236 00:14:37.039 --> 00:14:40.919 data and tries to achieve logical consistency. 237 00:14:41.200 --> 00:14:44.080 So so what if it's got data. 238 00:14:43.840 --> 00:14:49.000 That is wrong, it'll actually reflect upon that and remove 239 00:14:49.039 --> 00:14:51.000 the data that is that is wrong. It does not 240 00:14:51.200 --> 00:14:55.240 concord with reality. So it's base reasoning is very good. 241 00:14:55.720 --> 00:14:57.639 In fact, the. 242 00:15:00.000 --> 00:15:03.559 Even without fine tuning Grock three. The base model is 243 00:15:03.600 --> 00:15:08.360 better than Grock two, so with so we're really in 244 00:15:08.360 --> 00:15:13.000 the final stages of polishing Grock three. Probably it gets 245 00:15:13.000 --> 00:15:18.000 released in about a week or two, so pretty pretty soon. 246 00:15:19.919 --> 00:15:21.799 I don't want to be hasty in the release because 247 00:15:22.240 --> 00:15:28.480 a lot of the the final polish is necessary for 248 00:15:28.639 --> 00:15:33.600 a great user experience. So in some ways, you can 249 00:15:33.600 --> 00:15:35.480 think of it like a house. You know that last 250 00:15:35.519 --> 00:15:38.480 five percent where you do the finish, the drywall and 251 00:15:38.720 --> 00:15:41.679 do the painting and the trimming. Even though it's not 252 00:15:41.759 --> 00:15:46.840 much work, it transforms the house. So that I just 253 00:15:46.879 --> 00:15:49.120 want to make sure that that last five percent is 254 00:15:49.159 --> 00:15:53.840 done really well. And it's a week maybe two weeks, 255 00:15:54.480 --> 00:15:58.279 and I think it'll be very good, and I think 256 00:15:58.279 --> 00:16:01.159 this might be we think be better than anything else, 257 00:16:02.519 --> 00:16:05.240 and then maybe this might be the last time that 258 00:16:06.279 --> 00:16:07.799 any AI is better than a brag. 259 00:16:09.960 --> 00:16:12.919 Looking forward to it, everyone's I think excited the bottle. 260 00:16:13.960 --> 00:16:16.600 So I just want to touch upon a topic that 261 00:16:16.759 --> 00:16:19.480 was quoted in the media. You offered I think they 262 00:16:19.480 --> 00:16:22.240 said a group that was led by you offered ninety 263 00:16:22.240 --> 00:16:25.200 seven billion for acquiring open AI. 264 00:16:26.039 --> 00:16:29.519 I little round numbers. 265 00:16:29.799 --> 00:16:34.480 So I was personally involved in the meeting that you 266 00:16:34.559 --> 00:16:38.480 and Sam hosted in twenty seventeen in La if you remember, 267 00:16:38.559 --> 00:16:40.159 and you know, at that point of time you were 268 00:16:40.200 --> 00:16:43.120 the single largest shareholder, but think you contributed fifty million 269 00:16:43.159 --> 00:16:44.799 to the company. So it must hurt. 270 00:16:45.000 --> 00:16:47.200 I don't have any shares, actually I have no shares 271 00:16:47.240 --> 00:16:47.879 in open Air. 272 00:16:48.240 --> 00:16:50.559 But at that time it was a nonprofit, right, and 273 00:16:50.879 --> 00:16:52.879 it must hurt that, you know, you need to be 274 00:16:52.919 --> 00:16:55.519 a ninety seven billion for something that you bead fifty 275 00:16:55.519 --> 00:16:56.919 million dollars for in the past. 276 00:16:58.039 --> 00:17:02.279 Yeah, but I have I have a fate Love's irony. 277 00:17:03.879 --> 00:17:06.960 So I have a specific question here. Can you actually 278 00:17:07.039 --> 00:17:10.480 build a company like open ai and take it to 279 00:17:10.519 --> 00:17:12.799 the scale that you want to take as a nonprofit? 280 00:17:12.839 --> 00:17:15.720 Is it possible that you build a company that requires 281 00:17:15.799 --> 00:17:18.559 billions of dollars in comput capabilities to build these models 282 00:17:18.880 --> 00:17:21.559 while being a nonprofit or was it wish for thinking 283 00:17:21.599 --> 00:17:24.000 in the beginning and then you know, you guys parted 284 00:17:24.039 --> 00:17:25.279 ways because it couldn't work. 285 00:17:27.160 --> 00:17:31.440 Well, I mean, I think the evidence is there in that. 286 00:17:31.480 --> 00:17:35.119 Open ai has gotten this far while having at least 287 00:17:35.160 --> 00:17:38.240 a sort of dual profit nonprofit role. 288 00:17:38.640 --> 00:17:41.400 What they're trying to do now is completely delete the 289 00:17:41.440 --> 00:17:50.279 nonprofit and that seems really going too far. 290 00:17:50.920 --> 00:17:55.599 You know, I provided all of the funding for opening 291 00:17:55.759 --> 00:17:58.200 in the beginning, for the first almost fifty million dollars 292 00:17:58.720 --> 00:18:02.279 for nothing as a nonprofit, and it was meant to 293 00:18:02.279 --> 00:18:05.839 be open source. And so you know, I think this 294 00:18:05.880 --> 00:18:09.759 is analogous to like if you pay it, but if 295 00:18:09.759 --> 00:18:13.920 you find a nonprofit to preserve the Amazon rainforest. But 296 00:18:14.039 --> 00:18:16.240 then they but instead they turned into a lumber companies 297 00:18:16.319 --> 00:18:18.000 and chop down the trees and sell them for wood. 298 00:18:18.400 --> 00:18:20.440 You were like, wait a second, that's the exact opposite 299 00:18:20.480 --> 00:18:23.000 of what I paid, what I donated the money for. 300 00:18:24.680 --> 00:18:28.480 So open Air is meant to be open source nonprofit 301 00:18:28.599 --> 00:18:32.680 and now it is close. They changed the name closed 302 00:18:32.680 --> 00:18:37.240 for maximum profit AI it closed for voracious profit. I 303 00:18:37.279 --> 00:18:40.119 mean they're like, whoa are they after money? 304 00:18:40.119 --> 00:18:44.400 Next level? So why does this change need to occur? 305 00:18:47.000 --> 00:18:50.640 Yeah, I am. You said that in two weeks is 306 00:18:50.680 --> 00:18:54.000 going to be the most powerful model yet. I know 307 00:18:54.079 --> 00:18:59.079 that's you've been at the forefront of many technologies. Where 308 00:18:59.079 --> 00:19:01.640 do you think the biggest economic returns of these models 309 00:19:01.640 --> 00:19:04.119 are going to come from? Because currently we're spending billions 310 00:19:04.160 --> 00:19:05.759 and I think you mentioned this before. It's like the 311 00:19:05.799 --> 00:19:08.920 gambler syndrome. We're going on spending billions and hoping to 312 00:19:08.920 --> 00:19:10.359 put out a profit. At the end of the day, 313 00:19:10.720 --> 00:19:12.799 where do you think the biggest impact in terms of 314 00:19:12.839 --> 00:19:13.480 returns are. 315 00:19:13.319 --> 00:19:13.680 Going to be. 316 00:19:16.000 --> 00:19:19.880 Well, I think once you once you have humanoid robots 317 00:19:20.920 --> 00:19:25.319 and deep intelligence, you can basically you basically have why 318 00:19:25.400 --> 00:19:31.279 is the infinite products and services available? So with Tesla 319 00:19:32.079 --> 00:19:36.799 building the most advanced humanoid robot you know, then then 320 00:19:36.839 --> 00:19:40.240 those human humanoid robots can be directed by deep intelligence 321 00:19:41.440 --> 00:19:44.599 at the data center level. You can say you can 322 00:19:45.000 --> 00:19:47.920 you can produce any product, produce provide any service. 323 00:19:49.640 --> 00:19:51.519 There's really no limit to the economy. 324 00:19:51.640 --> 00:19:57.119 At that point. You can make anything provided links to that. 325 00:19:58.359 --> 00:20:01.039 So I'm not sure at that. 326 00:20:01.000 --> 00:20:04.319 Point will money even be meaningful. I don't know, it 327 00:20:04.400 --> 00:20:07.160 might not be. You know, the if. 328 00:20:08.599 --> 00:20:13.000 If the because you know, the economic output is productivity 329 00:20:13.039 --> 00:20:15.279 per capita at times the capital times. 330 00:20:15.000 --> 00:20:17.519 How many you know people do you have? If and 331 00:20:17.559 --> 00:20:18.599 if if in the. 332 00:20:18.519 --> 00:20:20.599 Form of human rightd robots, you have no meaningful limit 333 00:20:20.640 --> 00:20:24.160 on the number of of robots and the robots can 334 00:20:24.920 --> 00:20:30.559 basically do anything, then you'll have a sort of a 335 00:20:30.640 --> 00:20:34.559 universal high income situation. Anyone will be able to have 336 00:20:35.720 --> 00:20:39.880 as many products and services as that they want, with 337 00:20:39.960 --> 00:20:42.440 the exception of things that say have artificial scarcity, like 338 00:20:42.480 --> 00:20:45.200 a particular piece of art or something like that, but 339 00:20:45.559 --> 00:20:49.160 for any goods and services that will be available to everyone. 340 00:20:50.400 --> 00:20:53.440 So you've been it's going to be a very different world, 341 00:20:53.599 --> 00:20:53.839 you know. 342 00:20:54.079 --> 00:20:57.480 In fact, I recommend that people read maybe the End 343 00:20:57.519 --> 00:21:03.279 Banks the culture books for a frame of reference. So, 344 00:21:03.920 --> 00:21:09.200 because money is really like a database or information system 345 00:21:09.359 --> 00:21:14.400 for resource allocation. But if you don't have a scarcity 346 00:21:14.400 --> 00:21:18.279 of resources, it's not clear what purpose money has. 347 00:21:19.920 --> 00:21:24.319 Have you watched the movie Idiocracy? Yes, how do you 348 00:21:24.319 --> 00:21:26.599 guarantee that we don't end up in that world? If 349 00:21:26.640 --> 00:21:29.559 we don't need money, if I can think for us 350 00:21:29.599 --> 00:21:32.599 and do all these tasks, if as people, you know, 351 00:21:32.640 --> 00:21:35.519 we're dependent on something else to run the society and 352 00:21:35.519 --> 00:21:37.599 everything around it, how do you not end up in 353 00:21:37.599 --> 00:21:39.240 that world in the long term? 354 00:21:39.319 --> 00:21:43.839 I mean, well, I think Idiocracy was basically saying that 355 00:21:43.880 --> 00:21:46.359 if only if smart people don't reproduce, but only donbe 356 00:21:46.359 --> 00:21:49.720 people do, then everyone's going to be done. I mean, 357 00:21:49.720 --> 00:21:53.720 that's basically what I'm saying. That's the opening sequence of Idiocracy. 358 00:21:53.759 --> 00:21:58.079 The first ten minutes are amazing, and I hear people 359 00:21:58.119 --> 00:22:02.119 an ironically say the statements that are set in the 360 00:22:02.160 --> 00:22:06.759 opening sequence of Idiocracy, where you know, they don't have 361 00:22:06.799 --> 00:22:09.680 that they're too busy with their careers to have kids, 362 00:22:10.200 --> 00:22:15.000 and they keep postponing having kids for their careers until 363 00:22:15.319 --> 00:22:16.880 they're too old to have kids, and then they don't 364 00:22:16.920 --> 00:22:21.839 have kids. And that's I've heard those many people be 365 00:22:22.000 --> 00:22:28.039 like that. So yeah, I mean, I don't know. I 366 00:22:28.039 --> 00:22:32.640 think we might be headed to a bimodal human intelligence 367 00:22:32.640 --> 00:22:37.799 distribution where there's a small number of it's it's kind 368 00:22:37.799 --> 00:22:41.640 of maybe like more like Brave New World, Eldest Lexley. 369 00:22:42.480 --> 00:22:46.079 We've got sort of a sort of a small group 370 00:22:46.119 --> 00:22:49.599 of very smart humans. But then maybe the average intelligence 371 00:22:50.000 --> 00:22:51.160 grow slower over. 372 00:22:51.000 --> 00:22:52.559 Time, potentially. 373 00:22:55.880 --> 00:22:59.319 Because we have a sort of mating you know, in 374 00:22:59.440 --> 00:23:02.680 the last a few decades that or several decades that 375 00:23:03.000 --> 00:23:07.640 did not exist before. So but human intelligence I think 376 00:23:07.680 --> 00:23:13.079 will be dwarfed by machinery intelligence. I'm not sure how 377 00:23:13.079 --> 00:23:16.000 to feel about that, except that it is it'll be 378 00:23:16.000 --> 00:23:21.680 inevitable that at some point human intelligence will be a 379 00:23:21.799 --> 00:23:24.319 very small fraction. 380 00:23:24.160 --> 00:23:25.440 Of total intelligence. 381 00:23:27.079 --> 00:23:30.359 Digital intelligence will be more than nineteen nine percent of 382 00:23:30.599 --> 00:23:32.440 all intelligence in the future. 383 00:23:34.279 --> 00:23:37.680 So hopefully the hopefully the computers are nice to us. 384 00:23:38.119 --> 00:23:41.680 But I think wish for thinking, wish for thinking. 385 00:23:42.400 --> 00:23:42.759 I hope. 386 00:23:42.759 --> 00:23:48.480 So I think it matters like how we bring up AI, 387 00:23:48.559 --> 00:23:51.440 because you can think of AI like a super genius child, 388 00:23:51.880 --> 00:23:53.799 and but it still matters even if you have a 389 00:23:53.839 --> 00:23:57.079 super genius child, like what sort of values do you 390 00:23:57.359 --> 00:23:59.519 instill in that child? What do you say that teach that? 391 00:23:59.759 --> 00:24:04.720 How do you you you know, how do you as 392 00:24:04.720 --> 00:24:07.400 a child child's growing up? What values do you teach 393 00:24:07.440 --> 00:24:11.279 the child? And something that I think is extremely important 394 00:24:11.440 --> 00:24:18.319 is to be maximally truth seeking. And I think that's 395 00:24:18.400 --> 00:24:21.839 that's like that, that's that's what's this, what's the most 396 00:24:21.880 --> 00:24:23.519 important thing for AI safety. 397 00:24:23.960 --> 00:24:25.720 I think it's to be maximally truth seeking. 398 00:24:26.119 --> 00:24:30.279 And I think also curiosity is important, and I think 399 00:24:30.319 --> 00:24:34.799 it was curious and truth seeking, it will I think 400 00:24:34.839 --> 00:24:41.200 it will foster humanity and because it would be curious 401 00:24:41.200 --> 00:24:46.240