WEBVTT - Netflix Co-Founder Reed Hastings Talks AI, Future of TV 0:00:02.520 --> 0:00:07.400 Bloomberg Audio Studios, podcasts, radio news. 0:00:07.800 --> 0:00:12.040 A master's and computer science from Stanford nineteen eighty eight. 0:00:12.320 --> 0:00:13.360 An artificial intelligence. 0:00:13.440 --> 0:00:15.520 I was a new partner at a law firm called 0:00:15.560 --> 0:00:18.239 lower Color and Pickering in Washington during National trade? What 0:00:18.320 --> 0:00:20.600 did you see in artificial intelligence? 0:00:20.600 --> 0:00:26.200 As on It's honored to be with you all, I 0:00:26.239 --> 0:00:30.000 wish I had been a behavioral economist. That was my 0:00:30.240 --> 0:00:33.960 only real big regret in life. And so what did 0:00:33.960 --> 0:00:37.599 I see in AI? Which is sort of related, what's 0:00:37.640 --> 0:00:42.360 the nature of intelligence? I think I saw a dream 0:00:42.800 --> 0:00:48.720 of understanding ourselves and our society a little better, and 0:00:48.760 --> 0:00:52.680 that was sort of the primary motivator. I will say 0:00:52.760 --> 0:00:56.520 that the AI that I studied was the equivalent of 0:00:56.680 --> 0:01:01.160 the sun goes around the earth field. So we had 0:01:01.160 --> 0:01:04.679 a fundamental theory in the nineteen eighties that was completely 0:01:04.720 --> 0:01:06.920 false and didn't work, but we didn't know it at 0:01:06.920 --> 0:01:11.720 the time. So me and many other people then fled 0:01:11.760 --> 0:01:15.960 to all kinds of other fields. And it's exciting now 0:01:15.959 --> 0:01:20.319 to see a new set of techniques really be transformative. 0:01:21.200 --> 0:01:27.360 And to Mike's reference on the news this week, you know, 0:01:27.520 --> 0:01:33.160 you know that I'm enamored with subscription models and now 0:01:33.920 --> 0:01:39.240 I've become more aware about tender offers and I like 0:01:39.400 --> 0:01:45.200 TV channels. So we're announcing today my tender offer for Bloomberg. 0:01:47.440 --> 0:01:51.800 We hope you will consider it appropriately. Your board members 0:01:51.840 --> 0:01:56.840 are willing to debate the transaction. But unfortunately, as you 0:01:56.840 --> 0:01:59.480 would suspect as a board member, that's all I can 0:01:59.520 --> 0:02:01.600 say on the news of the week. 0:02:01.720 --> 0:02:04.160 So you're not going to answer my bosses question, who's going. 0:02:04.080 --> 0:02:06.000 To I think it is all answered. 0:02:07.800 --> 0:02:10.359 So looking back on AI as you saw it then 0:02:10.400 --> 0:02:12.919 and whatever you saw then, and comparing it to now, 0:02:13.840 --> 0:02:16.640 I hear a range of things being said. As extreme 0:02:16.760 --> 0:02:19.480 as this is, as big an invention as the industrial 0:02:19.480 --> 0:02:23.440 evolution sometimes say is the discovery of fire. Are we 0:02:23.600 --> 0:02:27.360 overestimating or underestimating the effect generative A. 0:02:27.400 --> 0:02:32.480 I will have us, Well, let's think back thirty forty 0:02:32.480 --> 0:02:35.240 thousand years ago, and we're a bunch of Neanderthals and 0:02:35.240 --> 0:02:38.320 we're in the entrance to our cave, and you know, 0:02:38.360 --> 0:02:42.080 we've been the dominant human humanoid species for a couple 0:02:42.120 --> 0:02:45.560 hundred thousand years. And the two of us look down 0:02:45.840 --> 0:02:48.919 and we see Homo sapiens. We're like, look at those guys, 0:02:48.919 --> 0:02:54.959 they're so skinny. Hairless, but they're pitiful. But those guys 0:02:54.960 --> 0:03:00.400 were really intelligent, and ultimately those homelostapiens dominated the earth 0:03:00.880 --> 0:03:06.880 and killed off us neandervals. So I would say intelligence 0:03:07.040 --> 0:03:11.880 per species has been highly selected for, and that Homo 0:03:11.919 --> 0:03:16.000 sapiens intelligence has allowed us to become the dominant species 0:03:16.040 --> 0:03:19.520 on Earth because we use our intelligence to make tools 0:03:19.600 --> 0:03:23.320 and do other things. So I think it's a lot 0:03:23.560 --> 0:03:27.839 different than say, mechanizing muscle power. So if you think 0:03:27.880 --> 0:03:33.639 of bulldozers and how did that transform you know, society, 0:03:34.120 --> 0:03:36.400 it did a lot, but it did it over one 0:03:36.440 --> 0:03:41.200 hundred years. And although muscles are important, they're not the 0:03:41.240 --> 0:03:44.680 core human attribute. Thus the dislocation of the end with 0:03:44.760 --> 0:03:49.520 the Neanderthals, which were much stronger than humans. So if 0:03:49.600 --> 0:03:54.920 AI develops to actually be super intelligence, then it will 0:03:54.920 --> 0:03:58.360 be a lot more profound, I think, than anything else, 0:03:58.560 --> 0:04:03.160 and that we will have actually real species threats because 0:04:04.400 --> 0:04:07.480 the AI will keep getting smarter and smarter and smarter 0:04:07.600 --> 0:04:11.720 without limit. A natural selection works quite slowly in terms 0:04:11.760 --> 0:04:15.240 of making humans more intelligent, so then we have to 0:04:15.280 --> 0:04:18.800 sort of say, okay, how fast and how you know? 0:04:19.080 --> 0:04:21.640 Will the computer is really thank Obviously when we all 0:04:21.720 --> 0:04:24.600 use AI as consumers we can find. You know, it's 0:04:24.640 --> 0:04:28.839 pretty miraculous in some cases, but in other cases it's 0:04:28.960 --> 0:04:33.800 not very effective. But you know, we're understanding more and 0:04:33.839 --> 0:04:36.719 more of the techniques. And again, one theory is it'll 0:04:36.760 --> 0:04:39.360 be like Moore's law and just AI will get better 0:04:39.360 --> 0:04:43.599 and better and better. But the other theory is it's 0:04:43.600 --> 0:04:46.680 sort of like the War on cancer, where cancer developed 0:04:46.720 --> 0:04:50.000 over a very long time. It has lots of different etiologies, 0:04:50.480 --> 0:04:52.560 and we keep coming up with a solution for one 0:04:52.600 --> 0:04:57.040 cancer but not another, and our overall progress and society 0:04:57.200 --> 0:05:01.080 against cancer has been you know, pretty steady but flat. 0:05:01.120 --> 0:05:05.520 It definitely not exponential, and so you know, it may 0:05:05.560 --> 0:05:08.880 be that as AI gets better, it hits various walls 0:05:09.400 --> 0:05:12.320 and that we've got some time to deal with it. 0:05:13.040 --> 0:05:15.520 Or it may be that it stays on this exponential 0:05:16.240 --> 0:05:18.600 So we're just going to have to watch. But I 0:05:18.640 --> 0:05:21.480 think we need to be prepared for it to be 0:05:21.640 --> 0:05:25.480 on the exponential, in which case we're going to have 0:05:25.760 --> 0:05:28.480 a lot of societal stress over the next twenty years. 0:05:28.640 --> 0:05:31.560 So your analogy is very helpful, but a little disturbing 0:05:31.720 --> 0:05:34.400 because I don't see many Neanderthals around anymore. 0:05:34.839 --> 0:05:35.960 Right, we have a lot of. 0:05:35.880 --> 0:05:39.880 Homo saviens in your analogy, What can we do to 0:05:40.000 --> 0:05:42.279 make sure that we survive, that that this thing we're 0:05:42.320 --> 0:05:46.520 creating doesn't become so smart and perhaps I can have 0:05:46.560 --> 0:05:48.920 our best interests so are well. 0:05:48.920 --> 0:05:53.440 It's a big challenge. It's hard to slow down because 0:05:54.160 --> 0:05:57.000 unlike chemical weapons, like we're kind of in a war 0:05:57.080 --> 0:06:00.360 with a West in Russia and yet no one using 0:06:00.440 --> 0:06:03.760 chemical weapons and no one's using nuclear weapons. So even 0:06:03.800 --> 0:06:06.480 in this incredible state of war, we're able to put 0:06:06.520 --> 0:06:10.000 some limits on what happens. The problem with AI is 0:06:10.040 --> 0:06:13.640 it's very continuous. You know, your thermostat is an AI thermostat. 0:06:13.720 --> 0:06:17.040 So there's no good way or even the major powers 0:06:17.080 --> 0:06:20.320 to agree not to use AI because it's so integrated 0:06:20.360 --> 0:06:24.719 through everything else that we do. So there's no real 0:06:24.800 --> 0:06:28.480 practical scenario to take a break as human society, even 0:06:28.520 --> 0:06:30.320 if we could get the politics to work on that. 0:06:31.080 --> 0:06:34.120 So instead we're in the situation where we in America 0:06:34.360 --> 0:06:38.360 better well damn win the race, and so everyone acknowledges that, 0:06:38.600 --> 0:06:43.599 and so all the companies are going full out, both 0:06:43.680 --> 0:06:48.760 for their own you know, competitive reasons, and so you know, 0:06:48.920 --> 0:06:52.520 it is a race, hopefully to the top of what 0:06:52.760 --> 0:06:56.320 can AI do for us? And there will be amazing 0:06:56.480 --> 0:07:00.680 positive scenarios. You know, all kinds of medical yours, all 0:07:00.760 --> 0:07:05.640 kinds of productivity. I certainly six years ago, AI was 0:07:05.720 --> 0:07:09.279 very good at image of analysis. That was one of 0:07:09.279 --> 0:07:11.480 the kind of workloads that it first got good at, 0:07:11.760 --> 0:07:15.760 and in particular in radiology, it got significantly better than 0:07:15.800 --> 0:07:19.040 the typical radiologists pretty quickly and has been better than 0:07:19.080 --> 0:07:23.080 the typical radiologist for at least five years. And so 0:07:23.240 --> 0:07:25.880 I would say, you know, weddings and with friends who 0:07:25.920 --> 0:07:28.680 are radiologists, what's good that your kids are not going 0:07:28.720 --> 0:07:31.520 in the business, because radiology is going to be the 0:07:31.560 --> 0:07:36.000 first profession that's wrecked by AI. Good for humanity, but 0:07:36.200 --> 0:07:41.080 not good for radiologists. Well, what happened is as radiology 0:07:41.120 --> 0:07:46.080 got better because of AI analysis that helped radiologists, the 0:07:46.160 --> 0:07:49.800 price of scans went down, and the number of scans 0:07:49.880 --> 0:07:52.240 turned out to be hugely elastic. Now you walk in, 0:07:52.320 --> 0:07:55.160 you've got a cough, boom, you get a scan, okay. 0:07:55.320 --> 0:07:57.840 And so what's happened now is there's about thirty four 0:07:57.880 --> 0:08:01.360 thousand radiologists in the US, which is a shortage. We 0:08:01.440 --> 0:08:06.240 now have a labor shortage of radiologists okay, by about 0:08:06.280 --> 0:08:09.760 five thousand. So it's an incredible story. Of of course, 0:08:09.840 --> 0:08:14.880 elasticity of demand to improve lives, i e. More scans, 0:08:15.280 --> 0:08:17.600 and so we may well see with AI that it 0:08:17.640 --> 0:08:22.120 makes Wall Street analysts more productive and makes software engineers 0:08:22.200 --> 0:08:26.000 more productive, and that in fact the elasticity and that 0:08:26.120 --> 0:08:28.800 we grows and that we grow the economy. And when 0:08:28.840 --> 0:08:32.200 people ask, you know, we're spending collectively as an industry 0:08:32.240 --> 0:08:34.640 at half a trillion dollars on AI data center, how 0:08:34.720 --> 0:08:37.320 is that ever going to get paid back? Well, you know, 0:08:37.600 --> 0:08:39.920 add one or two points to GDP growth and it 0:08:39.960 --> 0:08:44.400 gets paid back fast. So you know, I don't want 0:08:44.440 --> 0:08:47.319 to like guarantee that it's always going to be like radiology. 0:08:48.040 --> 0:08:52.959 I do think software engineering is particularly fascinating because all 0:08:52.960 --> 0:08:55.839 the major AI companies are working on it, and so 0:08:56.120 --> 0:08:59.200 it is the white collar canary in the coal mine. 0:09:00.280 --> 0:09:04.640 If software engineering engineering jobs go down a lot over 0:09:04.679 --> 0:09:08.240 the next five years, then that's probably going to happen 0:09:08.240 --> 0:09:11.640 to law on architecture and many other things. If in fact, 0:09:11.720 --> 0:09:14.600 because of the increased productivity, people are building a lot 0:09:14.640 --> 0:09:19.200 more software, sort of the radiology example, then I think 0:09:19.240 --> 0:09:22.880 in many professions we will see a big expansion and 0:09:22.920 --> 0:09:26.600 we can be more confident of the high productivity. None 0:09:26.600 --> 0:09:29.320 of that really answers the long term question that you asked, 0:09:29.320 --> 0:09:31.880 which is how are we not the Neanderthals? So I 0:09:31.880 --> 0:09:37.160 think reasonable chance high productivity rather than mass unemployment. Okay, 0:09:37.240 --> 0:09:40.120 but then ultimately, what if they're smarter and smarter and 0:09:40.160 --> 0:09:42.640 smarter than us. We're going to have to find ways, 0:09:42.679 --> 0:09:45.920 and I don't know what they are, to both continue 0:09:45.960 --> 0:09:48.480 to insist on alignment, and that's where you train the 0:09:48.559 --> 0:09:52.480 AIS to care about human beings so they are aligned 0:09:52.480 --> 0:09:56.959 with our values. Okay, but if somebody doesn't train, somebody 0:09:57.000 --> 0:09:59.600 programs their AI to try to take over the world, 0:10:00.120 --> 0:10:02.440 we're going to have to enlist the other AIS on 0:10:02.520 --> 0:10:06.160 our defense to protect us. Okay. So you know there's 0:10:06.400 --> 0:10:10.839 a number of scenarios out there, and probably for ten 0:10:10.920 --> 0:10:13.040 twenty years, we're not going to know how serious the 0:10:13.080 --> 0:10:15.720 threat is, but we will have tools. It's not just 0:10:15.760 --> 0:10:21.360 that the AI biological species we have been selected for 0:10:21.520 --> 0:10:25.200 dominance to try to grow our species. So AI is 0:10:25.240 --> 0:10:28.680 not naturally trying to expand, not naturally it could be 0:10:28.720 --> 0:10:31.320 programmed for that, but it can be also programmed to 0:10:31.400 --> 0:10:34.880 keep humans on top. So it's not as scary as 0:10:34.920 --> 0:10:38.120 a super powerful human, which we all kind of into it. 0:10:38.240 --> 0:10:41.680 A super powerful human would be hard to hold back 0:10:41.720 --> 0:10:44.440 from taking over the world. It's not as dire as that. 0:10:45.640 --> 0:10:49.239 Insofar as we do make progress on trying to reshape 0:10:49.960 --> 0:10:52.600 general of AI in a more positive direction. Does that 0:10:52.640 --> 0:10:55.240 come from outside the industry or inside the industry? Does 0:10:55.280 --> 0:10:58.440 that come from government regulation and agreement or does it 0:10:58.480 --> 0:11:00.800 have to come from inside? Because as I talk to 0:11:00.880 --> 0:11:04.240 many people in the business, they're much more focused on 0:11:04.280 --> 0:11:06.240 the race you talk about than they are on the 0:11:06.240 --> 0:11:06.760 safety part. 0:11:06.840 --> 0:11:08.360 They don't want to get slowed down in that. 0:11:08.400 --> 0:11:10.120 I mean, you're on the border anthropic, which I know 0:11:10.240 --> 0:11:11.920 is trying to make a move in the other direction. 0:11:12.200 --> 0:11:14.160 But can we do it from the outside or does 0:11:14.240 --> 0:11:17.080 the industry itself somehow have to internalize the risk. 0:11:18.080 --> 0:11:20.800 Well, I think lots of the industry is working on it. 0:11:20.800 --> 0:11:24.880 So there's different sides of safety. So there's when you're 0:11:25.679 --> 0:11:31.080 treating AI like a counselor and you know it, you 0:11:31.080 --> 0:11:33.320 know helps you tie a news that's not a good thing. 0:11:34.400 --> 0:11:37.480 And so those cases across the industry are getting you know, 0:11:37.559 --> 0:11:42.000 more and more watched for it eliminated, So there's inevitable 0:11:42.040 --> 0:11:46.640 safety bumps as any technology grows, so then there's more 0:11:46.720 --> 0:11:49.760 macro safety, like none of the major AIS can be 0:11:49.840 --> 0:11:54.120 used to design chemical weapons or biological weapons. But those 0:11:54.160 --> 0:11:58.800 defenses in the AIS you know, aren't perfect, and we 0:11:58.880 --> 0:12:01.760 have to constantly and best in them to prevent people 0:12:01.800 --> 0:12:05.840 from using this super powerful technology plus some crisper to 0:12:05.920 --> 0:12:09.600 do some really bad things. So there's active work across 0:12:09.640 --> 0:12:14.240 the whole industry on those scenarios. Because from a commercial standpoint, 0:12:14.520 --> 0:12:17.320 nobody wants their AI brand to be the brand that 0:12:17.400 --> 0:12:21.679 develops some bad virus, so there is an incentive there 0:12:21.800 --> 0:12:26.120 to protect against the negative cases. But the sort of 0:12:26.160 --> 0:12:28.720 big long term case is that the AIS get so 0:12:28.840 --> 0:12:33.120 smart that then we have to do things to make 0:12:33.120 --> 0:12:35.679 sure that they are aligned with human values, that they 0:12:35.720 --> 0:12:39.480 are programmed so that success is humans flourishing. 0:12:40.160 --> 0:12:42.319 Give us your view, but we are likely to see 0:12:42.320 --> 0:12:46.320 the biggest effect. You mentioned biomedicine. You mentioned computer software progners. 0:12:46.640 --> 0:12:48.000 Let's talk turning the news. 0:12:47.800 --> 0:12:51.600 Today with the deal between Disney and Open AI and 0:12:51.679 --> 0:12:53.520 the entertainment space that you know well, I mean you 0:12:53.880 --> 0:12:57.160 manage transition from my perspective, least from more traditional TV 0:12:57.440 --> 0:12:59.040 cable broadcast. 0:12:58.840 --> 0:12:59.600 Through to streaming. 0:13:00.080 --> 0:13:03.240 What does the transition look like from streaming through the AI. 0:13:03.960 --> 0:13:07.360 So just as an example, we can look at video 0:13:07.480 --> 0:13:11.560 creation storytelling and sort of say, let's look at what 0:13:11.600 --> 0:13:14.520 AI is going to do and the mechanics of generating 0:13:14.600 --> 0:13:18.960 video and frankly, whether that's a news channel trying to 0:13:19.000 --> 0:13:24.000 illustrate a concept or entertainers trying to do amazing special effects, 0:13:24.840 --> 0:13:28.600 we're going to have higher quality special effects and that 0:13:28.679 --> 0:13:32.479 will shift to be AI generated instead of manually generated 0:13:32.520 --> 0:13:36.560 with sort of overseas visual effects companies. So there is 0:13:36.840 --> 0:13:40.320 a shift there, But the core thing of storytelling is 0:13:40.440 --> 0:13:44.440 very hard. How do you do long form character development, 0:13:44.600 --> 0:13:49.360 creating tension, resolving tension. That's not a case that the 0:13:49.400 --> 0:13:53.960 AI does well today. Eventually, ten twenty years, the AI 0:13:54.120 --> 0:13:57.640 may win the Booker Prize, okay, and then if it does, 0:13:57.720 --> 0:14:00.720 it's going to be able to do a movie script all. So, 0:14:00.760 --> 0:14:04.600 but for now, the AI is really helping on sort 0:14:04.640 --> 0:14:07.760 of industrial aspects of what we do, like an amazing 0:14:07.840 --> 0:14:12.040 visual effects shot, and so it's you know, important, and 0:14:12.320 --> 0:14:15.160 we want to stay on the front as Disney does, 0:14:16.520 --> 0:14:19.600 but it's not doing the core thing, which is which 0:14:19.680 --> 0:14:22.160 story is going to captivate human attention. 0:14:22.560 --> 0:14:25.160 I'm not sure it says much about my job, about 0:14:25.160 --> 0:14:27.600 whether I keep my job in the AI world, because 0:14:27.600 --> 0:14:29.400 there's a lot of things that are done on television 0:14:29.480 --> 0:14:32.560 with news and other things that frankly, are not creating 0:14:32.560 --> 0:14:33.520 the next Lion King. 0:14:34.640 --> 0:14:38.400 Well, I think interpreting the world in your case is 0:14:38.440 --> 0:14:42.240 a broad general wisdom, and so I think you'll have 0:14:42.280 --> 0:14:47.440 a relatively longstanding role in your success. 0:14:47.480 --> 0:14:48.360 Given my age, I'm good. 0:14:48.600 --> 0:14:52.520 Given your age, you're good. But you know, it is 0:14:52.840 --> 0:14:55.800 a very big change. Again, you know, people want to 0:14:55.840 --> 0:14:59.320 compare it to the Industrial revolution that happened over two 0:14:59.400 --> 0:15:02.880 hundred years. This is going to happen over ten or twenty. 0:15:03.880 --> 0:15:08.280 And it might be that what's caused political polarization in 0:15:08.280 --> 0:15:11.880 the last decade or two is rate of change. So 0:15:12.040 --> 0:15:15.240 think about you know, since NAFTA, you know, sort of 0:15:15.280 --> 0:15:21.800 the rise of globalization, the redefining of marriage equality, immigration, 0:15:22.360 --> 0:15:25.960 all kinds of change in society. And one view is 0:15:27.040 --> 0:15:31.120 enough of our fellow Americans, it's just too much, and 0:15:31.200 --> 0:15:33.920 it radicalizes them and they're willing to vote for things 0:15:34.000 --> 0:15:38.160 they wouldn't normally vote for. And if that's fundamentally what's 0:15:38.200 --> 0:15:41.280 going on in US society as well as you know, 0:15:41.440 --> 0:15:45.720 Brexit and a couple other places. Then we're in for 0:15:45.760 --> 0:15:48.520 a pretty big storm because the rate of change is 0:15:48.560 --> 0:15:51.520 not going to slow down. The rate of change of society, 0:15:51.840 --> 0:15:55.080 you know, partially or maybe largely driven by AI, is 0:15:55.120 --> 0:15:58.440 going to be large. And so that may be that 0:15:59.360 --> 0:16:03.080 you know, how everyone today is nostalgic for Reagan, you know, 0:16:04.200 --> 0:16:06.480 there may be a day when we're nostalgic for Trump. 0:16:07.400 --> 0:16:13.560 That our polarization has really continued to increase because the 0:16:13.680 --> 0:16:18.160 rate of change has continued to increase. So that's why 0:16:18.400 --> 0:16:21.440 I think it's so important that leaders like all of 0:16:21.440 --> 0:16:25.040 you are thinking about how do we build bonds, how 0:16:25.040 --> 0:16:28.240 do we have people of Americans care about each other 0:16:29.200 --> 0:16:33.640 so that we can keep the society coherent and caring 0:16:34.400 --> 0:16:38.520 despite the rapid amounts of change in generated by technology. 0:16:38.600 --> 0:16:40.480 I want to come back to something you referred to 0:16:40.480 --> 0:16:42.360 just explore a bit more, and that is how are 0:16:42.360 --> 0:16:44.600 we going to pay for it? There is so much 0:16:44.640 --> 0:16:47.240 money going into data center's investment right now, and to 0:16:47.280 --> 0:16:49.360 something said, I think it's been supporting the markets because 0:16:49.400 --> 0:16:50.480 there's been so much investment. 0:16:51.320 --> 0:16:51.920 How do you get a. 0:16:51.920 --> 0:16:54.560 Return on that investment without laying off an awful lot 0:16:54.560 --> 0:16:55.280 of people. 0:16:55.960 --> 0:16:59.960 With GDP growth, so you can either automate away the job. 0:17:00.320 --> 0:17:03.000 That's one theory, and the other is that people will 0:17:03.040 --> 0:17:06.399 produce more and I'm sure it'll be a mix, but 0:17:06.520 --> 0:17:10.639 efficiency often generates more growth. I mean, we did the 0:17:10.720 --> 0:17:14.119 radiology example, so it's sort of that at larger scale. 0:17:14.440 --> 0:17:17.520 So the way we pay for it is more GDP growth. 0:17:17.800 --> 0:17:20.919 You can do GDP growth with fewer people. And so 0:17:21.040 --> 0:17:22.760 what happens to the people who are out of work 0:17:22.880 --> 0:17:25.440 or have to take much more menial jobs than they 0:17:25.440 --> 0:17:27.359 had before because it's the knowledge workers. 0:17:27.560 --> 0:17:30.480 Well, I think it's a great question. Compared to globalization. 0:17:30.560 --> 0:17:33.040 Probably this is a room full of globalists. We believe 0:17:33.080 --> 0:17:36.920 in the benefits of trade, and yet despite that that 0:17:37.000 --> 0:17:40.000 really did deliver in terms of the overall economy. There 0:17:40.080 --> 0:17:43.600 was enough dislocation and enough of the countries that our 0:17:43.680 --> 0:17:48.080 politics has been shifted. So I think there probably will 0:17:48.119 --> 0:17:51.960 be things like that because of AI, But at least 0:17:52.000 --> 0:17:56.600 it's not again, it's not a mass layoff scenario. Most likely, 0:17:56.680 --> 0:18:00.159 it's much more likely that there's a growth response. We 0:18:00.280 --> 0:18:05.000 also see significant GDP growth, you know, increased beyond what 0:18:05.040 --> 0:18:08.879 we typically see because of the productivity of AI. 0:18:09.440 --> 0:18:12.640 What I hear from you is AI has enormous potential, 0:18:13.080 --> 0:18:15.399 a lot of it for good, some of for decidedly 0:18:15.560 --> 0:18:19.119 not good. It's very complicated, it's coming very very fast. 0:18:19.800 --> 0:18:24.040 Where will the leadership in directing AI come from? Are 0:18:24.080 --> 0:18:26.439 there people who understand this, who get it, who have 0:18:26.480 --> 0:18:27.879 the right values, who can help us go in the 0:18:27.960 --> 0:18:30.560 right direction, particularly given how fast it's coming. 0:18:33.080 --> 0:18:35.480 I would say that's an emerging area. I mean, I 0:18:35.520 --> 0:18:38.720 think this was likely to be a big political issue 0:18:38.760 --> 0:18:43.960 in the next couple cycles because many Americans will be 0:18:44.000 --> 0:18:47.760 concerned about world changing too fast and too much, and 0:18:47.840 --> 0:18:51.199 I think it's up to our politicians to sort of 0:18:51.359 --> 0:18:56.880 understand that and channel it in some productive way. There's 0:18:56.960 --> 0:19:00.640 leaders in the industry like Jeffrey Hinton that spends full 0:19:00.680 --> 0:19:03.600 time now on these how do we keep humans at 0:19:03.600 --> 0:19:08.560 the center of the system. So there definitely are emerging 0:19:08.640 --> 0:19:11.639 leaders in that way. But again it's a new area. 0:19:11.960 --> 0:19:17.119 Like you know, nuclear energy you know, or other DNA 0:19:17.280 --> 0:19:19.080 for long time is very controversial. 0:19:19.880 --> 0:19:23.040 Timing is everything, and it comes. The develops you talk 0:19:23.080 --> 0:19:25.360 about come at a time when there's a decided rise 0:19:25.400 --> 0:19:27.520 in populism in the United States as well as in 0:19:27.600 --> 0:19:30.840 much of the Western world. Right now, the early things 0:19:30.920 --> 0:19:33.520 I'm hearing this but on political from the people is 0:19:34.200 --> 0:19:37.760 it increases our energy costs and it reduces our jobs. 0:19:38.200 --> 0:19:40.920 We're against it, and some politicians are showing signs now 0:19:40.920 --> 0:19:43.600 of trading on that and saying we should be against 0:19:43.640 --> 0:19:46.000 Actually data centers, we should be against them. How do 0:19:46.040 --> 0:19:48.600 you put together the rise of populism with the likely 0:19:48.640 --> 0:19:50.040 effects of general AI. 0:19:51.119 --> 0:19:54.320 Well, you know, the heart of democracy is we're going 0:19:54.400 --> 0:19:56.800 to live in this country together and we're going to 0:19:56.840 --> 0:19:59.480 have different views, and so you want a system that's 0:19:59.480 --> 0:20:03.479 not civil war to work out those views. And I 0:20:03.480 --> 0:20:05.800 think I think a lot of people will have those concerns. 0:20:05.800 --> 0:20:08.840 They'll be those effects, and we have to take them 0:20:08.960 --> 0:20:12.920 thoughtfully and seriously. And arguably during the era of globalization, 0:20:13.040 --> 0:20:16.840 there was sort of lip service to retraining, but we 0:20:16.880 --> 0:20:19.919 didn't really understand or take seriously the devastation of a 0:20:19.960 --> 0:20:23.480 lot of communities. So I think, you know, the elites 0:20:23.480 --> 0:20:25.720 in this room as an example of do a better 0:20:25.840 --> 0:20:28.400 job at that, which is, you know, where is AI 0:20:28.560 --> 0:20:32.280 really providing benefits to people's healthcare, to their day to 0:20:32.320 --> 0:20:34.600 day lives, and if we can do that, then there's 0:20:34.680 --> 0:20:38.680 more toleration of the dislocation, which is as you said, 0:20:38.920 --> 0:20:42.000 cost of electricity, data centers, those things. 0:20:42.200 --> 0:20:43.320 Are we over investing. 0:20:47.200 --> 0:20:51.200 It's unlikely, but you know it's definitely possible. But again, 0:20:51.800 --> 0:20:54.520 you know the Telkom boom in two thousand, I mean, 0:20:54.560 --> 0:20:57.879 you know it's a slight over investment, you know, so 0:20:59.600 --> 0:21:01.679 it could. But if you look at the in the 0:21:01.880 --> 0:21:04.399 in the telecom one, everyone got so excited about the 0:21:04.400 --> 0:21:08.840 Internet that they forward invested and were disappointed for a while, 0:21:08.920 --> 0:21:12.320 but ultimately the Internet really did deliver. But if you 0:21:12.320 --> 0:21:15.320 look at the mobile phone, it never had a bubble. 0:21:15.440 --> 0:21:17.280 It came out and just grew and grew and grew 0:21:17.280 --> 0:21:19.960 in impact. So you can get both things that live 0:21:20.080 --> 0:21:23.040 up to the hype and things where the hype gets 0:21:23.080 --> 0:21:26.640 bigger than the actual technology in the short term. And 0:21:26.720 --> 0:21:28.680 you know, the market's a good way to work that out. 0:21:28.680 --> 0:21:31.399 On AI, does it matter who gets their first? 0:21:32.040 --> 0:21:34.560 I mean, is this like beta and VHS where the 0:21:34.560 --> 0:21:36.560 point that gets their first gets to really set this up. 0:21:36.640 --> 0:21:38.560 You're in countries, it matters. 0:21:38.240 --> 0:21:41.000 A lot, So US versus China take that correct. 0:21:41.400 --> 0:21:45.119 If you're talking companies you know, I care a lot 0:21:45.119 --> 0:21:47.280 because I'm on the board of Entropic. But from a 0:21:47.320 --> 0:21:51.440 society standpoint, it probably just matters that the technology is 0:21:51.520 --> 0:21:53.200 developed and deployed in great ways. 0:21:53.800 --> 0:21:55.840 But it does matter in the US versus China. 0:21:56.720 --> 0:22:00.280 Yeah, absolutely. I mean, if another country doesn't have to 0:22:00.280 --> 0:22:04.320 be China right gets ahead of us. There are significant 0:22:04.359 --> 0:22:07.680 scenarios referred to in a class of the singularity where 0:22:07.680 --> 0:22:10.480 the AI gets so good that it rights the next 0:22:10.520 --> 0:22:13.800 DAI and that AI is much better, and then that 0:22:13.840 --> 0:22:16.520 writes the next day I And so you know, being 0:22:16.680 --> 0:22:21.040 first by even just a year gives you an enormous advantage. 0:22:21.080 --> 0:22:24.640 So that's the exponential. But again that's a theory that's 0:22:24.680 --> 0:22:29.240 not yet proven, but even its slight possibility means we 0:22:29.440 --> 0:22:31.720 better don't get there first. And I think we're on 0:22:31.760 --> 0:22:32.359 a good bathroom. 0:22:32.440 --> 0:22:33.240 Well that's my question. 0:22:33.280 --> 0:22:35.880 If that were your one goal in the United States, 0:22:36.200 --> 0:22:38.679 what are the policies that get you there? What are 0:22:38.720 --> 0:22:40.760 the things we need to do or avoid doing to 0:22:40.800 --> 0:22:42.440 make sure we're as competitive as possible. 0:22:43.440 --> 0:22:44.960 I think we're doing a pretty good job on that. 0:22:45.200 --> 0:22:48.760 I mean, there's you would push the edge on light regulation, 0:22:49.000 --> 0:22:52.679 make it easy to build data centers. You know, I 0:22:52.720 --> 0:22:58.960 think the government's done open field running, so I can't 0:22:58.960 --> 0:23:00.480 think I don't I don't think we need like a 0:23:00.520 --> 0:23:05.040 government investment program or something like that. So there's a 0:23:05.080 --> 0:23:09.000 lot of progress, and we are ahead of our competitors 0:23:09.040 --> 0:23:12.240 as far as we can tell. And again, we don't 0:23:12.240 --> 0:23:15.439 want to make it zero some because you'd like, just 0:23:15.480 --> 0:23:18.040 like we did in globalization, we would like to bring 0:23:18.080 --> 0:23:21.080 the rest of the world along with us so that 0:23:21.119 --> 0:23:24.640 we've got a stable we have less likely chance of war. 0:23:25.720 --> 0:23:30.080 So I mean, but that's happening. 0:23:30.680 --> 0:23:32.320 Okay, So let's talk about what you really love. 0:23:32.359 --> 0:23:39.000