1 00:00:00,760 --> 00:00:04,517 Hello again. If I keep doing this, I understand that 2 00:00:04,600 --> 00:00:09,837 there's vacancy at 60 minutes. Maybe I can If this investing thing doesn't work 3 00:00:09,920 --> 00:00:14,317 out, maybe I can do that. Another full disclosure, 4 00:00:14,400 --> 00:00:16,677 Sam is a friend. I'm also on the board 5 00:00:16,760 --> 00:00:19,437 of OpenAI, so 6 00:00:19,520 --> 00:00:23,797 I promise you I will not only ask him softball questions. I will channel my 7 00:00:23,880 --> 00:00:27,157 inner Brett Bear and and uh 8 00:00:27,240 --> 00:00:28,957 and Slink Gooper. 9 00:00:29,040 --> 00:00:31,557 So, Sam, as I said, 10 00:00:31,640 --> 00:00:35,397 CEO of OpenAI, one of the founders. 11 00:00:35,480 --> 00:00:39,197 He previously served as the president of Y Combinator, 12 00:00:39,280 --> 00:00:43,997 where he helped scale some of the world's most influential startup accelerators, 13 00:00:44,080 --> 00:00:48,037 and advise companies that will shape entire industries. 14 00:00:48,120 --> 00:00:49,877 At OpenAI, Sam has obviously been at the 15 00:00:49,960 --> 00:00:51,797 forefront of advancing artificial 16 00:00:51,880 --> 00:00:55,157 intelligence to the benefit of humanity, 17 00:00:55,240 --> 00:01:00,077 helping to bring cutting-edge AI tools to the hands of millions of people. How 18 00:01:00,160 --> 00:01:01,917 many people? Eight, 900 million people? 19 00:01:02,000 --> 00:01:05,757 >> More, yeah. >> and businesses around the world. 20 00:01:05,840 --> 00:01:10,437 So, Sam, let me start with what I think is the question that is on everybody's 21 00:01:10,520 --> 00:01:14,517 mind, okay? Which is 22 00:01:14,600 --> 00:01:16,917 where are we today 23 00:01:17,000 --> 00:01:19,317 in the world of AI? 24 00:01:19,400 --> 00:01:21,997 >> I I think at some point in the last few 25 00:01:22,080 --> 00:01:24,477 months, 26 00:01:24,560 --> 00:01:27,757 we really have crossed a threshold into 27 00:01:27,840 --> 00:01:30,477 major economic utility of these models. 28 00:01:30,560 --> 00:01:33,077 It may have happened a little bit earlier, but there was such an overhang 29 00:01:33,160 --> 00:01:37,117 before we figured out how to use these, and we had to not only continue to make 30 00:01:37,200 --> 00:01:40,997 the models get smarter, but figure out sort of the plumbing to make them easy 31 00:01:41,080 --> 00:01:43,557 to use. Where we're we're now in a world 32 00:01:43,640 --> 00:01:48,877 where the models are astounding people with the 33 00:01:48,960 --> 00:01:52,677 work they can do. Uh and I think this has been most 34 00:01:52,760 --> 00:01:54,797 noticeable in coding, yep, but it's also 35 00:01:54,880 --> 00:01:56,877 happening in science. It's happening in 36 00:01:56,960 --> 00:02:00,717 many fields of knowledge work. Uh 37 00:02:00,800 --> 00:02:04,037 at sort of with this orienting speed where people are saying like, "Man, 38 00:02:04,120 --> 00:02:09,237 these things that I thought were still years away are happening now and I have 39 00:02:09,320 --> 00:02:12,557 my job shifted from doing 40 00:02:12,640 --> 00:02:16,797 uh you know, direct technical work or uh you know, legal work to managing a 41 00:02:16,880 --> 00:02:19,077 team of agents doing this work." 42 00:02:19,160 --> 00:02:22,757 This is going to go much further. Uh I think we are at a very steep part of the 43 00:02:22,840 --> 00:02:23,797 curve 44 00:02:23,880 --> 00:02:27,557 and right now maybe you can trust say a 45 00:02:27,640 --> 00:02:32,077 AI software engineer to do a multi-hour task. Very soon it'll be a multi-day 46 00:02:32,160 --> 00:02:34,397 task and then a multi-week task. 47 00:02:34,480 --> 00:02:37,317 And not long after that, I think the 48 00:02:37,400 --> 00:02:39,637 paradigm will shift again and it'll feel 49 00:02:39,720 --> 00:02:44,517 like these AI systems are just connected to your life, to your company, whatever 50 00:02:44,600 --> 00:02:48,237 proactively thinking, working all the time. Uh 51 00:02:48,320 --> 00:02:52,157 and having full context on whatever they need to know and just sort of doing 52 00:02:52,240 --> 00:02:54,397 stuff like you would trust a senior employee to do. 53 00:02:54,480 --> 00:02:59,197 >> Mhm. And and you think that um companies have 54 00:02:59,280 --> 00:03:01,357 real have a real understanding of how 55 00:03:01,440 --> 00:03:03,637 these systems can help them and rein- 56 00:03:03,720 --> 00:03:07,157 reimagine how they do their businesses? 57 00:03:07,240 --> 00:03:11,157 >> Some do, some don't. Uh certainly the new generation of startups thinks 58 00:03:11,240 --> 00:03:14,277 differently than any generation of startups before. 59 00:03:14,360 --> 00:03:17,877 Uh it used to be that when we would talk to 60 00:03:17,960 --> 00:03:21,917 startups, they would talk about how many employees they needed. 61 00:03:22,000 --> 00:03:25,997 Uh now they generally don't want to hire a lot. They think that'll slow them down 62 00:03:26,080 --> 00:03:28,197 and they're all focused on how much compute they can get. 63 00:03:28,280 --> 00:03:31,117 >> Mhm. >> You know, can I reserve this much capacity? Can I do a cloud deal for 64 00:03:31,200 --> 00:03:33,877 that? Uh can I get this many tokens? 65 00:03:33,960 --> 00:03:38,317 And I think that is that that is a mental shift that bigger 66 00:03:38,400 --> 00:03:40,677 companies are going through more slowly 67 00:03:40,760 --> 00:03:44,277 but some are starting to do that. Uh one 68 00:03:44,360 --> 00:03:49,597 place I think you can see this happening is engineering orgs and product orgs 69 00:03:49,680 --> 00:03:54,037 talking about they're doubling, tripling what they're planning to ship this year. 70 00:03:54,120 --> 00:03:56,077 >> Mhm. And that has not happened before. 71 00:03:56,160 --> 00:03:57,117 >> Yeah. 72 00:03:57,200 --> 00:04:00,597 And you've been quite vocal in saying 73 00:04:00,680 --> 00:04:05,037 that um artificial general intelligence will 74 00:04:05,120 --> 00:04:06,997 come sooner rather than later. 75 00:04:07,080 --> 00:04:10,677 Want to share your views on how close we are and how soon it'll 76 00:04:10,760 --> 00:04:12,317 come? 77 00:04:12,400 --> 00:04:15,237 >> At at this point, I think the definition 78 00:04:15,320 --> 00:04:19,477 of AGI really matters. Some people would say we already got there. Some people 79 00:04:19,560 --> 00:04:22,517 say it's very close. Some people say we're kind of you know, it's maybe still 80 00:04:22,600 --> 00:04:28,517 a year away, but in any case, that word has ceased to have much meaning. 81 00:04:28,600 --> 00:04:31,917 >> Mhm. >> Uh there are maybe two thresholds that we 82 00:04:32,000 --> 00:04:33,597 could talk about that are interesting. 83 00:04:33,680 --> 00:04:35,677 >> Okay. >> Number one, when is there going to be 84 00:04:35,760 --> 00:04:40,077 more of the world's cognitive capacity inside of data centers than outside of 85 00:04:40,160 --> 00:04:45,677 them? And that to me feels like maybe it could happen. 86 00:04:45,760 --> 00:04:48,317 Huge error bars. I could be totally wrong, but maybe that could happen by 87 00:04:48,400 --> 00:04:51,237 like late 2028. >> Mhm. 88 00:04:51,320 --> 00:04:55,437 >> And that's an extraordinary shift in the world. 89 00:04:55,520 --> 00:04:56,677 Um 90 00:04:56,760 --> 00:05:00,477 the other one is when can 91 00:05:00,560 --> 00:05:03,157 a CEO of major company, a president of a 92 00:05:03,240 --> 00:05:05,437 major country, a Nobel Prize winning 93 00:05:05,520 --> 00:05:07,637 scientist, when can they not do their 94 00:05:07,720 --> 00:05:10,957 job without making heavy use of AI? This 95 00:05:11,040 --> 00:05:14,677 doesn't mean that there will be an AI CEO or an AI president, uh 96 00:05:14,760 --> 00:05:18,317 but it does mean that the the role of let's 97 00:05:18,400 --> 00:05:23,277 say a human CEO, when I think about my job, it's really quite different. You you 98 00:05:23,360 --> 00:05:25,277 still do need a person to 99 00:05:25,360 --> 00:05:27,277 stand behind decisions and kind of 100 00:05:27,360 --> 00:05:30,677 exercise human judgment and 101 00:05:30,760 --> 00:05:33,037 all of the understanding 102 00:05:33,120 --> 00:05:35,597 that we expect out of someone running an 103 00:05:35,680 --> 00:05:39,437 an important organization to do, but 104 00:05:39,520 --> 00:05:43,037 the actual parts of my role that I will 105 00:05:43,120 --> 00:05:47,477 increasingly have to rely on an AI to do because no human can. You know, no human 106 00:05:47,560 --> 00:05:51,477 can talk to no human CEO can talk to every employee at a company, every 107 00:05:51,560 --> 00:05:53,557 customer, be in every meeting, be an 108 00:05:53,640 --> 00:05:56,357 expert in every field. And so, 109 00:05:56,440 --> 00:05:59,317 more and more I think of these jobs will 110 00:05:59,400 --> 00:06:03,637 be supervising a bunch of AI, um providing 111 00:06:03,720 --> 00:06:06,237 oversight, deciding how to like trust 112 00:06:06,320 --> 00:06:08,037 the the outputs, how to provide 113 00:06:08,120 --> 00:06:12,317 guidance. And that threshold of when you 114 00:06:12,400 --> 00:06:16,637 really wouldn't want to be doing your job, 115 00:06:16,720 --> 00:06:20,877 running a large organization without heavily heavy reliance on AI, 116 00:06:20,960 --> 00:06:23,357 I think that's another sort of interesting threshold. That may take a 117 00:06:23,440 --> 00:06:26,277 little bit longer, but probably not a lot longer. 118 00:06:26,360 --> 00:06:29,077 >> And and as you do your job, how much are 119 00:06:29,160 --> 00:06:32,357 you finding yourself relying on 120 00:06:32,440 --> 00:06:36,117 some of the agents and some of the artificial intelligence that we're 121 00:06:36,200 --> 00:06:39,317 developing at OpenAI? >> Uh 122 00:06:39,400 --> 00:06:42,117 it's ramping incredibly 123 00:06:42,200 --> 00:06:45,757 quickly. Um if I have like a new idea 124 00:06:45,840 --> 00:06:51,520 for a business model, a strategy shift, a uh 125 00:06:51,680 --> 00:06:55,317 a product offering we should do, the very first thing I do before I even 126 00:06:55,400 --> 00:06:57,517 bounce it off somebody else is to ask 127 00:06:57,600 --> 00:07:02,517 our tools. And as they get more context, and I think this really is like 128 00:07:02,600 --> 00:07:06,157 the next big thing to happen is they can get close to full context of our 129 00:07:06,240 --> 00:07:10,717 company, access to all of our internal docs, communication, code, 130 00:07:10,800 --> 00:07:16,117 customer data, everything, the quality of the answers, thought, whatever you want to 131 00:07:16,200 --> 00:07:17,237 call it, gets better and better. 132 00:07:17,320 --> 00:07:21,437 >> Right. Okay. So, let's let's shift um a little bit. 2 133 00:07:21,520 --> 00:07:26,477 weeks ago, um you announced a $110 billion funding 134 00:07:26,560 --> 00:07:31,917 round. I asked ChatGPT, how does that compare compare to any 135 00:07:32,000 --> 00:07:34,717 other fundraising that's been done in the public markets? 136 00:07:34,800 --> 00:07:37,917 >> don't know. >> Uh four times as large, okay? The 137 00:07:38,000 --> 00:07:42,357 largest public offering ever done was roughly 25 billion that Aramco did 138 00:07:42,440 --> 00:07:45,477 several years ago. And the public markets are supposed to 139 00:07:45,560 --> 00:07:50,317 be the broadest and deepest sources of capital, okay? 140 00:07:50,400 --> 00:07:53,677 Three strategic partners, Amazon, 141 00:07:53,760 --> 00:07:56,997 Nvidia, and oh and SoftBank. 142 00:07:57,080 --> 00:07:59,477 Tell us a little bit about this. 143 00:07:59,560 --> 00:08:02,397 How is this an inflection point of the company? 144 00:08:02,480 --> 00:08:05,237 And one of the questions somebody asked me is, what are we spending all that 145 00:08:05,320 --> 00:08:08,797 money on? The 146 00:08:08,880 --> 00:08:12,357 There is many hard parts of this business, but one of the hardest ones is 147 00:08:12,440 --> 00:08:15,197 the infrastructure is so expensive. You 148 00:08:15,280 --> 00:08:20,360 need so much of it, and you have to commit so far in advance. 149 00:08:20,960 --> 00:08:25,757 I have never seen any other industry quite like this. I 150 00:08:25,840 --> 00:08:29,637 mean, there have been clearly many capital intensive industries throughout 151 00:08:29,720 --> 00:08:34,077 history, but as I look at the heat what's to come in 152 00:08:34,160 --> 00:08:35,637 front of us in the future in future 153 00:08:35,720 --> 00:08:41,117 years, if the ramp stays as steep as it looks like it is right now, the demand 154 00:08:41,200 --> 00:08:46,597 is growing as fast as it as it's growing, you have to do some pretty unusual things. 155 00:08:46,680 --> 00:08:49,157 Uh OpenAI does a lot of things that look 156 00:08:49,240 --> 00:08:51,237 weird. 157 00:08:51,320 --> 00:08:53,757 We spend a ton of money on 158 00:08:53,840 --> 00:08:56,837 infrastructure in advance of revenue. We 159 00:08:56,920 --> 00:08:59,517 do new business models like ads that 160 00:08:59,600 --> 00:09:04,477 seem like, you know, maybe not the most profitable thing we could do. Um 161 00:09:04,560 --> 00:09:08,237 a long list of other things, but we have this fundamental belief in abundance of 162 00:09:08,320 --> 00:09:12,357 intelligence. And that one of the most important things in the future is that 163 00:09:12,440 --> 00:09:16,997 we make intelligence, you know, to borrow an old phrase from 164 00:09:17,080 --> 00:09:19,757 the energy industry that didn't quite work, too cheap to meter. We want to 165 00:09:19,840 --> 00:09:21,997 flood the world with intelligence. We want people to just use it for 166 00:09:22,080 --> 00:09:24,997 everything. We want this to just be something that the future generation 167 00:09:25,080 --> 00:09:29,637 doesn't think about. They expect everywhere, and everybody has access to 168 00:09:29,720 --> 00:09:33,197 like geniuses, as many as they need in any 169 00:09:33,280 --> 00:09:37,240 area that they need. And 170 00:09:37,640 --> 00:09:41,917 this principle, which is one of our kind of like top guiding principles, does 171 00:09:42,000 --> 00:09:45,437 lead to a lot of behavior that would look 172 00:09:45,520 --> 00:09:49,197 less natural for other companies. And one of those is 173 00:09:49,280 --> 00:09:52,477 we really want to get out of this world that we have been in that we still think 174 00:09:52,560 --> 00:09:55,637 we're on a trajectory to stay on without changing what we do of always being 175 00:09:55,720 --> 00:09:58,237 capacity constrained. >> Right. 176 00:09:58,320 --> 00:10:01,477 And capacity constrained you mean compute? Yeah. 177 00:10:01,560 --> 00:10:06,917 And I've often heard you say a lot that compute is revenues. 178 00:10:07,000 --> 00:10:11,317 You want to talk a little bit about how you think how you think about that? 179 00:10:11,400 --> 00:10:14,797 >> Fundamentally, our business and I think the business of every other model 180 00:10:14,880 --> 00:10:19,237 provider is going to look like selling tokens. 181 00:10:19,320 --> 00:10:23,437 You know, they may come from bigger or smaller models which makes them more or 182 00:10:23,520 --> 00:10:26,917 less expensive. They may use more or less reasoning which also makes them 183 00:10:27,000 --> 00:10:29,997 more or less expensive. They may be running all the time in the background 184 00:10:30,080 --> 00:10:33,437 trying to help you out. Uh they may run only when you need them if you want to 185 00:10:33,520 --> 00:10:38,637 pay less. They may work super hard. You know, spend tens of 186 00:10:38,720 --> 00:10:41,877 millions, hundreds of millions, someday billions of dollars on a single problem. 187 00:10:41,960 --> 00:10:44,437 >> Right. >> That's really valuable. But 188 00:10:44,520 --> 00:10:46,997 we 189 00:10:47,080 --> 00:10:49,437 see a future where 190 00:10:49,520 --> 00:10:51,597 intelligence is a utility like 191 00:10:51,680 --> 00:10:56,517 electricity or water and people buy it from us 192 00:10:56,600 --> 00:11:00,957 um on a meter and use it for whatever they want to use it for. 193 00:11:01,040 --> 00:11:04,957 The demand that we see for that seems like it's going to continue to just go 194 00:11:05,040 --> 00:11:08,797 like this. And 195 00:11:08,880 --> 00:11:14,197 if we don't have enough, we either can't sell it or the price gets really high 196 00:11:14,280 --> 00:11:19,397 and it you know, it kind of goes to rich people or society makes a bunch of sort 197 00:11:19,480 --> 00:11:22,557 of central planning decisions that I think almost always go badly about, you 198 00:11:22,640 --> 00:11:25,997 know, we're going to use our limited compute supply for this and not that. 199 00:11:26,080 --> 00:11:30,397 So, the best thing to me throughout all the history of capitalism, innovation, 200 00:11:30,480 --> 00:11:32,637 whatever you want is to just flood the market. 201 00:11:32,720 --> 00:11:35,757 >> Yeah. And obviously a critical part of 202 00:11:35,840 --> 00:11:37,637 addressing that 203 00:11:37,720 --> 00:11:41,477 compute demand is Stargate. 204 00:11:41,560 --> 00:11:45,757 And this is an infrastructure conference and you know, you announced that several 205 00:11:45,840 --> 00:11:49,837 months ago, or almost a year ago, I guess it was, yeah. 206 00:11:49,920 --> 00:11:53,997 How's it going in the US? And then how is it going, because I 207 00:11:54,080 --> 00:11:59,680 also started in in Abu Dhabi. >> Yeah. Yeah. Um 208 00:12:00,240 --> 00:12:03,037 Honestly, there's many cool parts of the job. One 209 00:12:03,120 --> 00:12:06,317 of the coolest is getting to go visit these mega data centers under 210 00:12:06,400 --> 00:12:07,837 construction and operation. >> Yeah. 211 00:12:07,920 --> 00:12:11,317 >> Just the scale is, you know, of like these gigawatt campuses. It's it's it's 212 00:12:11,400 --> 00:12:15,237 really hard to explain and you see these photos and it's like, okay, it does look 213 00:12:15,320 --> 00:12:18,157 big. And then you go there and you're walking through it and building to 214 00:12:18,240 --> 00:12:21,597 building and it's just, you know, 10,000 people there, all these different 215 00:12:21,680 --> 00:12:24,197 skilled trades doing all these different things and 216 00:12:24,280 --> 00:12:28,797 you know, it looks like a spaceship inside. It's really quite incredible. Um 217 00:12:28,880 --> 00:12:32,557 we are training right now on the first site in Abilene, 218 00:12:32,640 --> 00:12:33,917 what I think will be the best model in 219 00:12:34,000 --> 00:12:38,797 the world, hopefully by a lot. And it is 220 00:12:38,880 --> 00:12:42,597 it's so amazing to have gone from, you know, many visits while it was under 221 00:12:42,680 --> 00:12:46,997 construction, and just really internalizing the scale and incredible 222 00:12:47,080 --> 00:12:49,277 complexity to the day that one 223 00:12:49,360 --> 00:12:51,277 researcher at OpenAI types in like one 224 00:12:51,360 --> 00:12:54,157 command and it pushes enter, 225 00:12:54,240 --> 00:12:58,557 unbelievable number of GPUs spin up and start doing this one huge computation 226 00:12:58,640 --> 00:13:02,517 all together. >> Yeah. It's very And and 227 00:13:02,600 --> 00:13:06,757 what have been the most pleasant surprises and what have 228 00:13:06,840 --> 00:13:11,237 been the hardest things about getting it up and running and scaling it going 229 00:13:11,320 --> 00:13:16,157 forward? >> Uh I mean, there have been all the expected 230 00:13:16,240 --> 00:13:19,517 challenges and then the unknown unknowns. Like there was a crazy weather 231 00:13:19,600 --> 00:13:22,677 event in Abilene that we hadn't, you know, was outside 232 00:13:22,760 --> 00:13:25,797 what we had planned for and that brought things down for a little while. 233 00:13:25,880 --> 00:13:30,237 Um there's all the supply chain challenges. Anything at this scale is 234 00:13:30,320 --> 00:13:32,437 just like so much stuff goes wrong. Lots 235 00:13:32,520 --> 00:13:39,080 goes right, but the just try trying to build something around the clock 236 00:13:39,280 --> 00:13:42,397 with such complexity, there's all the stuff that goes wrong. 237 00:13:42,480 --> 00:13:46,757 One of the biggest surprises on the upside is how 238 00:13:46,840 --> 00:13:50,877 many different organizations had to come together to do this in an incredibly 239 00:13:50,960 --> 00:13:53,197 short period of time. Uh, and how much 240 00:13:53,280 --> 00:13:54,997 we all like ended up working together as 241 00:13:55,080 --> 00:13:58,557 one team uh, under a lot of pressure. 242 00:13:58,640 --> 00:14:03,077 >> And of course, the power demand part of it is the one that a lot of people have 243 00:14:03,160 --> 00:14:06,237 have focused on. Um, 244 00:14:06,320 --> 00:14:10,357 are you optimistic that we will solve that challenge 245 00:14:10,440 --> 00:14:15,077 first in the US and then we can talk about in in other places? 246 00:14:15,160 --> 00:14:19,077 >> Uh, in the long term, I am. 247 00:14:19,160 --> 00:14:23,477 >> Okay. >> Uh, I have no doubt we will figure out how 248 00:14:23,560 --> 00:14:27,797 to build huge amounts of power generation. AI will help, of course. 249 00:14:27,880 --> 00:14:31,437 Uh, but I think the portfolio the world has in front of it 250 00:14:31,520 --> 00:14:35,157 gas, solar, nuclear fission, nuclear fusion, more 251 00:14:35,240 --> 00:14:39,237 like we're I feel good about what what we will be 252 00:14:39,320 --> 00:14:41,397 capable of and what we will eventually do. 253 00:14:41,480 --> 00:14:42,157 >> Mhm. 254 00:14:42,240 --> 00:14:43,957 >> Um, 255 00:14:44,040 --> 00:14:48,637 given the demand growth we are seeing, 256 00:14:48,720 --> 00:14:52,997 I am sort of hoping for a miracle in terms of figuring out how we can get way 257 00:14:53,080 --> 00:14:54,957 more efficient per watt >> Uh, okay. 258 00:14:55,040 --> 00:14:59,877 >> models to give us time to build out all of this infrastructure. Now, 259 00:14:59,960 --> 00:15:03,477 the track record there has been incredible. Um, 260 00:15:03,560 --> 00:15:07,317 people cite whatever amazing statistic they like about how much more efficient 261 00:15:07,400 --> 00:15:10,877 our models have gotten over our industries models have gotten over time, 262 00:15:10,960 --> 00:15:14,477 but one that I think is incredible, 263 00:15:14,560 --> 00:15:18,237 our first reasoning model was called 01, came out 264 00:15:18,320 --> 00:15:22,917 like 16 months ago. Uh, 265 00:15:23,000 --> 00:15:28,800 and our latest model where we're now integrated reasoning is 5.4. 266 00:15:29,040 --> 00:15:31,357 To get the same answer 267 00:15:31,440 --> 00:15:34,437 to a hard problem from that first model 268 00:15:34,520 --> 00:15:40,800 to 5.4 has been a reduction in cost of about 1,000 x. 269 00:15:41,320 --> 00:15:42,997 Maybe I was a little bit wrong on the timeline. Maybe it was a little bit 270 00:15:43,080 --> 00:15:45,197 longer. But in any case, since 01 till now. 271 00:15:45,280 --> 00:15:50,317 >> Yeah. >> 1,000 x. Uh, that is unbelievable in a relatively 272 00:15:50,400 --> 00:15:54,677 short period of time. And 273 00:15:54,760 --> 00:15:58,397 the thing that that the two things that I think that points to 274 00:15:58,480 --> 00:16:02,277 one, we are still so early and in this 275 00:16:02,360 --> 00:16:05,597 paradigm and we're still so we have so much more to gain about our 276 00:16:05,680 --> 00:16:08,957 understanding of how to develop these models and train them and run them 277 00:16:09,040 --> 00:16:13,597 efficiently >> Mhm. >> that we have 278 00:16:13,680 --> 00:16:17,757 were then and still are doing things in dumb ways and will get better and 279 00:16:17,840 --> 00:16:21,437 better. And two, is that 280 00:16:21,520 --> 00:16:25,477 human ingenuity and the ability to operate 281 00:16:25,560 --> 00:16:29,877 in constraints and to find ways to solve problems almost always surprise you on the 282 00:16:29,960 --> 00:16:32,757 upside. So, it's not just that the models have gotten better, it's that we 283 00:16:32,840 --> 00:16:36,277 have figured out, you know, kernel engineers came to help figure out how to 284 00:16:36,360 --> 00:16:40,237 write more efficient kernels and like power engineers, the people that design 285 00:16:40,320 --> 00:16:43,077 data centers found more efficient ways to do that. So, people are answering the 286 00:16:43,160 --> 00:16:45,237 call well beyond just the model side >> Right. 287 00:16:45,320 --> 00:16:46,237 >> to make this more efficient. 288 00:16:46,320 --> 00:16:48,837 >> Right. And then of course, um, 289 00:16:48,920 --> 00:16:53,717 obviously OpenAI is a big customer of the Nvidias and AMDs of the world, but 290 00:16:53,800 --> 00:16:57,117 um, you know, I think we signed a big contract with Amazon. 291 00:16:57,200 --> 00:16:58,357 But we're building our own chip. 292 00:16:58,440 --> 00:17:01,517 >> Yeah. >> What's what's the thinking behind that? 293 00:17:01,600 --> 00:17:03,717 >> Uh, so the chip we're doing is inference only. 294 00:17:03,800 --> 00:17:08,037 >> Okay. >> And the thinking behind it is that on this, you know, come 295 00:17:08,120 --> 00:17:09,997 rise up to solve problems in front of 296 00:17:10,080 --> 00:17:12,317 us, um, 297 00:17:12,400 --> 00:17:15,957 we think that a specialized chip to be 298 00:17:16,040 --> 00:17:19,717 not necessarily the fastest inference chip, but the cheapest inference chip, 299 00:17:19,800 --> 00:17:23,597 the most efficient per watt given the constraints we see in front of us, is 300 00:17:23,680 --> 00:17:27,197 going to be important for all the agent demand we see in the future. So, it's a 301 00:17:27,280 --> 00:17:31,277 it's an opinionated bet, it's a limited chip, but the thing that it does, 302 00:17:31,360 --> 00:17:34,157 in a world where we're energy constrained, I think it'll be very important. 303 00:17:34,240 --> 00:17:38,197 >> Maybe you can explain because I am not sure everybody in the audience is 304 00:17:38,280 --> 00:17:43,197 AI proficient with the difference between an inference chip and a training chip. 305 00:17:43,280 --> 00:17:44,557 >> Sorry, I should have done that first. 306 00:17:44,640 --> 00:17:45,517 So, 307 00:17:45,600 --> 00:17:49,957 there there are two main phases to uh 308 00:17:50,040 --> 00:17:53,477 AI workloads. Now, they're going to blend together eventually and it'll all 309 00:17:53,560 --> 00:17:55,837 be one continuous thing. But right now, 310 00:17:55,920 --> 00:17:58,757 first we train a model, 311 00:17:58,840 --> 00:18:00,717 you know, a gigantic number of GPUs 312 00:18:00,800 --> 00:18:03,157 crunch for weeks or months on a bunch of 313 00:18:03,240 --> 00:18:05,837 data and you can think of that like 314 00:18:05,920 --> 00:18:08,597 maybe it takes you 22 years in life to 315 00:18:08,680 --> 00:18:13,237 get your education and you learn a bunch of stuff starting when you're a baby and 316 00:18:13,320 --> 00:18:16,317 you drop things and see they fall and then eventually like in college physics 317 00:18:16,400 --> 00:18:20,157 understanding like at a very detailed level what's going on there. 318 00:18:20,240 --> 00:18:24,157 Um And then after that, if you ask a model 319 00:18:24,240 --> 00:18:26,637 to solve a physics question, it that's 320 00:18:26,720 --> 00:18:31,197 called inference uh and that's quite efficient. But this is true for people 321 00:18:31,280 --> 00:18:34,357 as well and when people talk about all these AI models are so 322 00:18:34,440 --> 00:18:37,557 efficient, they're usually comparing the 323 00:18:37,640 --> 00:18:40,597 22 years that a human takes to train to 324 00:18:40,680 --> 00:18:45,517 the 1 second that an adult takes to solve a physics problem. If you compare the 325 00:18:45,600 --> 00:18:48,717 you know, the model solving the physics problem to the human solving physics 326 00:18:48,800 --> 00:18:52,917 problem, actually the models already probably are more energy efficient. Um 327 00:18:53,000 --> 00:18:57,637 but training is this hugely massive amount of processing power that then 328 00:18:57,720 --> 00:19:01,037 produces really just like a file of numbers that 329 00:19:01,120 --> 00:19:03,877 you can then pose a question to and get a response. 330 00:19:03,960 --> 00:19:07,757 >> And are you optimistic about the progress we're making on the chip? 331 00:19:07,840 --> 00:19:12,037 >> Yeah, um we should have the first chips deployed at scale 332 00:19:12,120 --> 00:19:15,077 uh by the end of this year. We should have the first chips back just in a few 333 00:19:15,160 --> 00:19:18,237 months now. Um and looks like it'll be really good. 334 00:19:18,320 --> 00:19:21,877 >> Fantastic. So, you announced a new partnership this 335 00:19:21,960 --> 00:19:25,957 morning, actually, with the North American Building Trades Unions to 336 00:19:26,040 --> 00:19:30,357 expand training pathways for skilled construction, which has been a subject 337 00:19:30,440 --> 00:19:34,677 of my many of the discussions today. Can you tell us a little bit about what you 338 00:19:34,760 --> 00:19:38,517 and uh Sean McGarvey >> Yes. >> agreed and What's he And he's going to 339 00:19:38,600 --> 00:19:41,997 be on stage later on. So, let's hear your version of it and then see if his 340 00:19:42,080 --> 00:19:46,640 version is the same. >> So, 341 00:19:48,200 --> 00:19:52,557 we talked earlier and we've talked The world has talked many times 342 00:19:52,640 --> 00:19:53,997 in the past about the need for AI 343 00:19:54,080 --> 00:19:56,437 infrastructure, the need for the world 344 00:19:56,520 --> 00:20:00,357 to have the physical infrastructure, 345 00:20:00,440 --> 00:20:03,117 power plants, transmission lines, data 346 00:20:03,200 --> 00:20:05,997 center halls, chillers, 347 00:20:06,080 --> 00:20:10,477 obviously the racks and the GPUs and everything that goes inside of those. Um 348 00:20:10,560 --> 00:20:14,757 and I sort of wish everyone would go visit 349 00:20:14,840 --> 00:20:17,157 at some point one of these 350 00:20:17,240 --> 00:20:21,997 mega scale data centers because it it's hard, 351 00:20:22,080 --> 00:20:25,877 you know, when you ask ChatGPT a question, you get your answer back and it's really 352 00:20:25,960 --> 00:20:29,917 hard to visualize the scale of what it took to make that or what it takes to 353 00:20:30,000 --> 00:20:34,397 make that happen. Uh people talk about all of these different 354 00:20:34,480 --> 00:20:36,397 places that were limited. >> Mhm. 355 00:20:36,480 --> 00:20:38,477 >> We are limited in the, you know, number 356 00:20:38,560 --> 00:20:43,837 of turbines or now it's the voltage transformers or it's the memory fabs or 357 00:20:43,920 --> 00:20:46,557 it's building the data centers, whatever. All of these things have one 358 00:20:46,640 --> 00:20:49,757 thing in common, which is they are massively complex physical 359 00:20:49,840 --> 00:20:53,277 infrastructure that require skilled tradespeople 360 00:20:53,360 --> 00:20:57,077 >> Mhm. >> uh and a lot of them to do. And 361 00:20:57,160 --> 00:20:59,357 no matter where the choke point in the 362 00:20:59,440 --> 00:21:02,677 supply chain is at any one time, 363 00:21:02,760 --> 00:21:04,397 when I talk to people about what it 364 00:21:04,480 --> 00:21:07,597 would take to accelerate it, um 365 00:21:07,680 --> 00:21:11,997 it is more skilled trades workers to build out all of this infrastructure 366 00:21:12,080 --> 00:21:15,437 that we all depend on. I think this will be incredible jobs. 367 00:21:15,520 --> 00:21:19,997 Uh more than that, I will I think they will lay the foundation for the next 368 00:21:20,080 --> 00:21:23,837 generation of American infrastructure and economic prosperity. And we are 369 00:21:23,920 --> 00:21:28,197 thrilled to get to work uh together to to drive that faster. 370 00:21:28,280 --> 00:21:32,237 >> Uh just so you know, we at BlackRock share that view and we think it's really 371 00:21:32,320 --> 00:21:35,477 exciting that that you announced this partnership together. I think that's 372 00:21:35,560 --> 00:21:38,917 really really terrific. Uh I think another question that's on 373 00:21:39,000 --> 00:21:44,997 people's minds is competition with China. 374 00:21:45,080 --> 00:21:47,117 Where do you think that stands? 375 00:21:47,200 --> 00:21:51,397 Uh what what do we need to do to make sure I assume we're ahead, but I don't 376 00:21:51,480 --> 00:21:54,277 know. So, you might tell us whether we are or not. 377 00:21:54,360 --> 00:21:57,917 What do we need to do to make sure we stay ahead? 378 00:21:58,000 --> 00:22:02,157 >> So, a general framework thought first and 379 00:22:02,240 --> 00:22:06,200 then I'll answer the question more specifically. 380 00:22:07,240 --> 00:22:10,597 I think the discovery of deep learning is 381 00:22:10,680 --> 00:22:12,837 is closer to 382 00:22:12,920 --> 00:22:15,077 discovering an element or a fundamental 383 00:22:15,160 --> 00:22:18,077 property of physics than it is uh of 384 00:22:18,160 --> 00:22:22,117 like a secret technology. Um 385 00:22:22,200 --> 00:22:24,037 and that means that eventually, and 386 00:22:24,120 --> 00:22:26,157 eventually probably not being very long, 387 00:22:26,240 --> 00:22:31,157 the fundamental ideas that make a model so capable 388 00:22:31,240 --> 00:22:34,597 will be simplified. They'll be very well known. 389 00:22:34,680 --> 00:22:37,357 And just like we kind of understand how 390 00:22:37,440 --> 00:22:41,637 big parts of physics work, we will understand as a scientific principle how 391 00:22:41,720 --> 00:22:45,717 big parts of artificial intelligence work. Um 392 00:22:45,800 --> 00:22:50,197 this is this really we started to appreciate this with the scaling laws uh 393 00:22:50,280 --> 00:22:53,677 that OpenAI published maybe 7 years ago now. 394 00:22:53,760 --> 00:22:57,717 Um there was such a measurable 395 00:22:57,800 --> 00:23:01,477 beautiful correlation between the resources that go into a model and the 396 00:23:01,560 --> 00:23:03,277 intelligence of that model 397 00:23:03,360 --> 00:23:06,237 that it kind of felt 398 00:23:06,320 --> 00:23:10,437 like hair-raising but clear at the time that there was just something 399 00:23:10,520 --> 00:23:12,837 fundamental going on here as a scientific principle. 400 00:23:12,920 --> 00:23:14,757 >> Mhm. >> Now, there have been a lot of details 401 00:23:14,840 --> 00:23:18,677 we've discovered since. There will be more to come, but like other scientific 402 00:23:18,760 --> 00:23:23,437 frontiers, it is simplifying and becoming more clear over time. And 403 00:23:23,520 --> 00:23:27,757 eventually, this recipe will be well understood as a scientific principle. It 404 00:23:27,840 --> 00:23:32,957 will not be a trade secret in the sense that other things have been. 405 00:23:33,040 --> 00:23:38,397 Now, the analogy I like best historically uh from technological history is the 406 00:23:38,480 --> 00:23:41,877 transistor. >> Mhm. >> Transistor was also a sort of 407 00:23:41,960 --> 00:23:45,517 fundamental scientific breakthrough, very hard to 408 00:23:45,600 --> 00:23:49,797 uh discover, kind of chancy to discover, uh took us a little while to refine our 409 00:23:49,880 --> 00:23:53,397 discovery, but once we understood it, the scientific principle was clear to 410 00:23:53,480 --> 00:23:58,877 everyone. There were still massive amounts of operational knowledge that went around 411 00:23:58,960 --> 00:24:01,717 that. >> Uh TSMC >> can still do things that no one else in 412 00:24:01,800 --> 00:24:03,757 the world can do, and you know, I I 413 00:24:03,840 --> 00:24:06,517 expect the industrial process around 414 00:24:06,600 --> 00:24:09,117 this to have a lot of advantages, 415 00:24:09,200 --> 00:24:13,637 uh competitive advantages. I also expect that the integration in workflows and 416 00:24:13,720 --> 00:24:16,557 training data and other and usability of models, 417 00:24:16,640 --> 00:24:18,717 there will be a lot of differentiation there. 418 00:24:18,800 --> 00:24:21,797 Maybe most of all, I expect there will be differentiation on who has the 419 00:24:21,880 --> 00:24:26,597 infrastructure and how much of it. But the fundamental scientific principles 420 00:24:26,680 --> 00:24:28,757 I are going to be 421 00:24:28,840 --> 00:24:33,957 well known, and they will fit on a t-shirt, I think. Um 422 00:24:34,040 --> 00:24:38,237 In terms of where we are, uh the most capable models in the world, 423 00:24:38,320 --> 00:24:40,277 the frontier the US is leading on. 424 00:24:40,360 --> 00:24:43,477 >> Mhm. >> Um the cheapest 425 00:24:43,560 --> 00:24:48,237 inference usage for a two-generation earlier model, China is leading on. 426 00:24:48,320 --> 00:24:50,797 >> Mhm. >> Um infrastructure in the US is currently 427 00:24:50,880 --> 00:24:53,637 leading on, but China's moving much faster on. 428 00:24:53,720 --> 00:24:57,757 And that kind of industrialization, productization, whatever you want to 429 00:24:57,840 --> 00:25:01,637 call it. I The US is leading in closed source. 430 00:25:01,720 --> 00:25:04,957 China's leading in open source. Uh I think the US is probably leading 431 00:25:05,040 --> 00:25:09,237 overall. >> And you were recently in India. 432 00:25:09,320 --> 00:25:12,717 And um sounded like you were very excited about 433 00:25:12,800 --> 00:25:16,397 how India is thinking about the challenges and the opportunities. 434 00:25:16,480 --> 00:25:19,077 >> I was blown away uh talking to Indian 435 00:25:19,160 --> 00:25:25,000 startups and how they are using this technology. Um 436 00:25:25,160 --> 00:25:28,077 I like got there and someone like handed me a briefing sheet for India and it was 437 00:25:28,160 --> 00:25:30,237 like you know 438 00:25:30,320 --> 00:25:35,157 Codex usage in India has like 10x'd in some small number of months. And I was 439 00:25:35,240 --> 00:25:37,837 like that's got to be a bug. Like that >> It can't be right. 440 00:25:37,920 --> 00:25:43,517 >> It can't be right. Um but it was true. And the 441 00:25:43,600 --> 00:25:45,717 And then I started talking to these startups. 442 00:25:45,800 --> 00:25:48,560 And 443 00:25:49,320 --> 00:25:52,437 it's an even stronger version of what's happening in the US 444 00:25:52,520 --> 00:25:56,037 of people saying, "Hey, the world is different. Um 445 00:25:56,120 --> 00:25:59,317 you know, you talk about a one-person startup. I'm trying to build a 446 00:25:59,400 --> 00:26:02,357 zero-person startup. I'm trying to just write a prompt that's going to make my 447 00:26:02,440 --> 00:26:06,957 whole startup and you know, write my software, do my customer support, do my 448 00:26:07,040 --> 00:26:09,957 legal stuff, whatever. And then I want to go on vacation." 449 00:26:10,040 --> 00:26:13,597 Um And the companies 450 00:26:13,680 --> 00:26:17,237 big companies in India they were just saying like, "How much capacity can we 451 00:26:17,320 --> 00:26:20,117 buy from you? How long can we reserve it for? Can we negotiate this right now? 452 00:26:20,200 --> 00:26:22,917 We're not going to like let you leave the room until you agree on this with 453 00:26:23,000 --> 00:26:24,917 us." That just the level of aggression 454 00:26:25,000 --> 00:26:28,397 and speed and sort of a belief that AI 455 00:26:28,480 --> 00:26:30,157 was going to 456 00:26:30,240 --> 00:26:33,037 reshape the business landscape in India 457 00:26:33,120 --> 00:26:37,560 was was really quite impressive. Um 458 00:26:37,760 --> 00:26:42,117 Yeah. >> Is is that different from when we when 459 00:26:42,200 --> 00:26:44,357 we talk to customers in the US? 460 00:26:44,440 --> 00:26:48,597 >> Uh it's it's the same vector, but they seem a little further 461 00:26:48,680 --> 00:26:51,957 along. They're moving faster. >> Yeah. The other The other comment you I 462 00:26:52,040 --> 00:26:55,877 think I saw you made was that there's a difference between autocratic and 463 00:26:55,960 --> 00:26:58,797 democratic AI. 464 00:26:58,880 --> 00:27:03,320 What did you mean by that and what what do you think is at stake? 465 00:27:05,920 --> 00:27:09,477 >> Once in a while, I think a technological shift comes 466 00:27:09,560 --> 00:27:12,077 along that 467 00:27:12,160 --> 00:27:15,277 reshapes society to such a degree that 468 00:27:15,360 --> 00:27:18,277 it does not decisions about it do not 469 00:27:18,360 --> 00:27:22,437 belong to the handful of companies that happen to be it. I am a huge believer in 470 00:27:22,520 --> 00:27:25,517 capitalism. I am a huge believer in the rights of companies. 471 00:27:25,600 --> 00:27:28,357 Um I am a 472 00:27:28,440 --> 00:27:31,077 huge believer in governments shouldn't 473 00:27:31,160 --> 00:27:35,840 interfere too much. But 474 00:27:37,240 --> 00:27:38,837 e- 475 00:27:38,920 --> 00:27:40,757 I think this is one of those exceptional 476 00:27:40,840 --> 00:27:44,237 times where society has a legitimate 477 00:27:44,320 --> 00:27:48,717 interest in what the impact of this technology 478 00:27:48,800 --> 00:27:51,277 is going to be. I think the internet was one of these times, too, and I don't 479 00:27:51,360 --> 00:27:56,317 think we got all of that right. But I'd like us to learn 480 00:27:56,400 --> 00:28:00,717 what we can do better. And 481 00:28:00,800 --> 00:28:02,597 if what 482 00:28:02,680 --> 00:28:05,237 the AI companies say, and I also believe 483 00:28:05,320 --> 00:28:09,637 it comes true, and that this is going to reshape the economy. This is going to 484 00:28:09,720 --> 00:28:12,797 reshape geopolitical power. This is going to 485 00:28:12,880 --> 00:28:15,037 change how we all live our lives. 486 00:28:15,120 --> 00:28:20,037 Um then I don't think it should be up to 487 00:28:20,120 --> 00:28:22,517 companies or 488 00:28:22,600 --> 00:28:24,757 a government to impose 489 00:28:24,840 --> 00:28:27,317 a particular will on how this is going 490 00:28:27,400 --> 00:28:30,877 to get used. Um I think that this 491 00:28:30,960 --> 00:28:32,997 belongs to the will of the people and 492 00:28:33,080 --> 00:28:35,157 working through the democratic process. 493 00:28:35,240 --> 00:28:38,040 And 494 00:28:39,120 --> 00:28:42,357 companies like ours, I think are quickly, more quickly than companies in 495 00:28:42,440 --> 00:28:46,197 previous generations have had to do, moving into a sort of critical 496 00:28:46,280 --> 00:28:47,797 infrastructure role. >> Right. 497 00:28:47,880 --> 00:28:52,757 >> Where we have to say um we create this technology. 498 00:28:52,840 --> 00:28:54,717 We are experts in it and we should have 499 00:28:54,800 --> 00:28:58,077 a real voice and we have a we have 500 00:28:58,160 --> 00:29:01,837 opinions and understanding of where its limitations 501 00:29:01,920 --> 00:29:05,757 are and where it's not ready to be used and where great harm could come. 502 00:29:05,840 --> 00:29:09,197 Uh but the rules, the limitations have to 503 00:29:09,280 --> 00:29:13,317 be agreed upon by society through this process and because the technology is 504 00:29:13,400 --> 00:29:16,037 moving so fast, it'd be great for that 505 00:29:16,120 --> 00:29:20,357 democratic process to run a little bit faster. But companies, 506 00:29:20,440 --> 00:29:24,837 governments, uh need to be able to depend on companies like ours to 507 00:29:24,920 --> 00:29:28,157 integrate and be able to use the the technology. 508 00:29:28,240 --> 00:29:33,597 >> Okay. So coming back to the global AI race, 509 00:29:33,680 --> 00:29:37,800 where where do you think the US is the most vulnerable? 510 00:29:44,680 --> 00:29:47,197 >> Three three things come to mind. Uh 511 00:29:47,280 --> 00:29:51,757 one, there's been a ton of noise made about 512 00:29:51,840 --> 00:29:55,637 the global supply chain dependence and US infrastructure. 513 00:29:55,720 --> 00:29:58,077 I don't have anything new or deep to say there, 514 00:29:58,160 --> 00:30:03,597 but I can't overstate how scary this is to me. 515 00:30:03,680 --> 00:30:06,437 >> Yeah. >> Um the if we fall behind on 516 00:30:06,520 --> 00:30:08,757 infrastructure and can't catch back up, 517 00:30:08,840 --> 00:30:15,117 um if if if globalization falls apart in any of the many ways it could, 518 00:30:15,200 --> 00:30:19,557 >> Mhm. >> uh and we are not able to fairly independently [snorts] 519 00:30:19,640 --> 00:30:24,597 keep building AI infrastructure, that 520 00:30:24,680 --> 00:30:29,437 that seems like a big vulnerability and I don't I don't hate, but I don't love our 521 00:30:29,520 --> 00:30:31,597 global position right now. >> Mhm. 522 00:30:31,680 --> 00:30:37,520 >> Um the second is if we don't 523 00:30:39,200 --> 00:30:42,637 it's a competitive world and if we don't move as quickly as other countries on 524 00:30:42,720 --> 00:30:45,157 economic adoption of this >> Mhm. 525 00:30:45,240 --> 00:30:50,077 then I think we will lose the advantage that we have from being the economic 526 00:30:50,160 --> 00:30:54,757 powerhouse that we are. And this is about how quickly companies 527 00:30:54,840 --> 00:30:58,237 adopt it. This is about how quickly our scientists adopt it. This is how quickly 528 00:30:58,320 --> 00:31:01,557 our government adopts it. This is like 529 00:31:01,640 --> 00:31:03,477 on the positive side, I think this is a 530 00:31:03,560 --> 00:31:09,077 once in many generation opportunity to really improve the economy, 531 00:31:09,160 --> 00:31:11,917 really rewrite some of the rules of society that aren't working in light of 532 00:31:12,000 --> 00:31:13,597 this new incredible wealth fountain we 533 00:31:13,680 --> 00:31:18,037 have. So I can see the world where 534 00:31:18,120 --> 00:31:21,637 this is not a disadvantage at all, but this becomes our biggest competitive 535 00:31:21,720 --> 00:31:26,157 advantage. I don't 536 00:31:26,240 --> 00:31:29,757 I don't think it's like super obvious that we're on the trajectory we 537 00:31:29,840 --> 00:31:32,957 want to be there now. Again, I don't hate it. I just I think it could move 538 00:31:33,040 --> 00:31:35,397 faster and you can see 539 00:31:35,480 --> 00:31:40,157 a bunch of potential headwinds, you know, like AI is not very popular in the 540 00:31:40,240 --> 00:31:44,677 US right now. Data centers are getting blamed for electricity price hikes. 541 00:31:44,760 --> 00:31:48,957 Almost every company that does layoffs is getting 542 00:31:49,040 --> 00:31:52,797 is blaming AI whether or not it really is about AI. 543 00:31:52,880 --> 00:31:56,117 There is this real debate about the relative power between governments and 544 00:31:56,200 --> 00:32:01,280 companies going on. So So there's like a lot of 545 00:32:02,120 --> 00:32:03,557 There's a lot of stuff happening there. 546 00:32:03,640 --> 00:32:06,437 And then the third category is 547 00:32:06,520 --> 00:32:10,757 diffusion into the rest of the world. Is the world mostly going to build on the 548 00:32:10,840 --> 00:32:15,237 American AI tech stack? Chips, models, applications, whatever. Or are we going 549 00:32:15,320 --> 00:32:19,160 to enact a set of policies that make that harder? 550 00:32:19,360 --> 00:32:22,837 >> So it sounds to me like you think that AI could be 551 00:32:22,920 --> 00:32:26,237 the foundation of an immense 552 00:32:26,320 --> 00:32:28,237 productivity boom. 553 00:32:28,320 --> 00:32:31,637 If properly used and if if adopted. 554 00:32:31,720 --> 00:32:35,677 >> Yes, for sure. Although I think the way we measure that is going to have to 555 00:32:35,760 --> 00:32:39,720 change. Explain. 556 00:32:42,920 --> 00:32:45,757 I can see a world where we have an 557 00:32:45,840 --> 00:32:47,797 incredible productivity boom. 558 00:32:47,880 --> 00:32:51,237 Quality of life goes up and up. 559 00:32:51,320 --> 00:32:55,517 Most of the things that we think we we say we would care about get better and 560 00:32:55,600 --> 00:32:58,997 better. And yet GDP in the way we currently 561 00:32:59,080 --> 00:33:01,757 measure it goes down and down. 562 00:33:01,840 --> 00:33:04,477 Like deflation for a very long period of time. 563 00:33:04,560 --> 00:33:07,277 And 564 00:33:07,360 --> 00:33:09,997 I don't know what it means to like live 565 00:33:10,080 --> 00:33:12,877 in a forever deflationary world. 566 00:33:12,960 --> 00:33:17,397 I don't know what it means to think about GDP in a 567 00:33:17,480 --> 00:33:21,837 and GDP's correlation of quality to quality of life in a world where you 568 00:33:21,920 --> 00:33:25,757 know more of the economic capability or the sorry intellectual capability is 569 00:33:25,840 --> 00:33:29,637 inside of data centers than outside of them. But you know maybe we're going to 570 00:33:29,720 --> 00:33:34,317 find out and and I think there's going to be a lot of 571 00:33:34,400 --> 00:33:37,637 debate in the coming years about what the right 572 00:33:37,720 --> 00:33:41,757 thing to measure is. >> Do you think we're thinking about these 573 00:33:41,840 --> 00:33:46,117 challenges and issues the right way? Are we having that Are we starting to talk 574 00:33:46,200 --> 00:33:48,717 about that? 575 00:33:48,800 --> 00:33:52,117 >> I I think we yes, we are starting to. 576 00:33:52,200 --> 00:33:54,797 If there was an easy consensus answer >> [laughter] 577 00:33:54,880 --> 00:33:57,677 >> have done it by now. So I don't think anyone knows 578 00:33:57,760 --> 00:34:02,757 what to do. The all of these things that we've 579 00:34:02,840 --> 00:34:05,917 depended on for so long as a society are sort of 580 00:34:06,000 --> 00:34:08,117 coming into question at the same time. 581 00:34:08,200 --> 00:34:10,917 There was this quote that I saw on the internet a few weeks ago. It really 582 00:34:11,000 --> 00:34:14,677 stuck in my head and it was something like 583 00:34:14,760 --> 00:34:17,437 for centuries maybe millennia we have 584 00:34:17,520 --> 00:34:19,957 learned a lot about how to structure 585 00:34:20,040 --> 00:34:22,957 society to manage scarcity. 586 00:34:23,040 --> 00:34:29,840 Almost none of that helps us as we have to quickly learn towards managing abundance. 587 00:34:30,400 --> 00:34:35,077 So that's like a real change to how capitalism has worked. Um 588 00:34:35,160 --> 00:34:39,357 capitalism has also depended on this balance between 589 00:34:39,440 --> 00:34:41,157 somewhat of a power balance between 590 00:34:41,240 --> 00:34:44,357 labor and capital. But, if it's hard in 591 00:34:44,440 --> 00:34:48,677 many of our current jobs to outwork a GPU, then that changes. And I see we're 592 00:34:48,760 --> 00:34:52,597 out of time, but I had like a list of 10 things like that changing all at once. I 593 00:34:52,680 --> 00:34:55,917 I I'm not a long-term jobs doomer. I I 594 00:34:56,000 --> 00:34:59,997 think we will figure out new things to do. I'm also certainly not a long-term 595 00:35:00,080 --> 00:35:02,037 capitalism doomer. I believe in it very, 596 00:35:02,120 --> 00:35:04,477 very deeply. Um 597 00:35:04,560 --> 00:35:07,637 but I think the next few years 598 00:35:07,720 --> 00:35:11,637 are going to be a painful adjustment as we get to this future we all get to 599 00:35:11,720 --> 00:35:17,237 redefine of what the new system and just this incredible prosperity looks like. And 600 00:35:17,320 --> 00:35:20,077 there are going to be some very intense and uncomfortable debates on the way 601 00:35:20,160 --> 00:35:23,157 there. >> Okay, so thank you, Sam, for that very 602 00:35:23,240 --> 00:35:26,797 thoughtful discussion. Here's what I will say. 603 00:35:26,880 --> 00:35:31,557 Five years from now, we'll come back here 604 00:35:31,640 --> 00:35:36,877 and we'll see where we are and how we've navigated our way through this. 605 00:35:36,960 --> 00:35:38,757 Deal? >> Deal. I look forward to that. 606 00:35:38,840 --> 00:35:42,447 >> Don't drink. Thank you. >> Thank you. Thank you. 607 00:35:42,530 --> 00:35:44,550 >> [applause] 608 00:35:47,080 --> 00:35:48,837 >> We'll now take [music] a short break. 609 00:35:48,920 --> 00:35:50,517 Please join us in the East Green Room 610 00:35:50,600 --> 00:35:56,360 for refreshments and return to your seats in 15 minutes. Thank you. 611 00:35:59,560 --> 00:36:01,580 >> [music] 612 00:38:24,760 --> 00:38:26,880 >> Mhm. 613 00:39:00,400 --> 00:39:02,440 >> Mhm.