WEBVTT - Empires of AI With Karen Hao 0:00:02.400 --> 0:00:06.840 Also media, Hello and welcome to Better Offline. I'm your 0:00:06.840 --> 0:00:20.840 host ed zich Tron. As ever, remember you can buy 0:00:20.840 --> 0:00:23.840 Better Offline merchandise linkers in the episode notes. Today, I 0:00:23.840 --> 0:00:26.040 am joined by Karen Howe, the author of the upcoming 0:00:26.079 --> 0:00:28.680 book Empire of Ai, which tells the story of Open 0:00:28.720 --> 0:00:31.440 Ai and the arms rais surrounding large language models. Karen, 0:00:31.480 --> 0:00:32.280 thank you for joining me. 0:00:32.840 --> 0:00:34.120 Thank you so much for having me ed. 0:00:35.000 --> 0:00:39.279 So you describe the progress of these models in these 0:00:39.280 --> 0:00:41.600 companies as a kind of colonialism, Can you get into 0:00:41.640 --> 0:00:42.120 that for me? 0:00:43.520 --> 0:00:46.239 Yeah, So, if you think about the way that empires 0:00:46.240 --> 0:00:49.960 of old operated during the very long history of European colonialism, 0:00:50.280 --> 0:00:53.480 they were essentially taking resources that were not their own, 0:00:54.480 --> 0:00:57.880 exploiting massive amounts of labor as in not paying them 0:00:58.160 --> 0:01:01.680 or paying them extremely small amounts of money, and they 0:01:01.680 --> 0:01:05.320 were doing this all under a civilizing mission, this idea 0:01:05.480 --> 0:01:09.560 that they were bringing modernity and progress to all of humanity, 0:01:09.600 --> 0:01:11.600 when in fact what was actually happening was they were 0:01:11.600 --> 0:01:14.920 just fortifying themselves and the empire, and the people at 0:01:14.959 --> 0:01:17.640 the top of the empire, and everyone else that kind 0:01:17.640 --> 0:01:20.160 of lived in the world had to live in the 0:01:20.160 --> 0:01:22.559 thrash of what the people at the top decided based 0:01:22.600 --> 0:01:25.000 on their whims for what was part of their self 0:01:25.040 --> 0:01:28.680 serving agenda. And that's essentially what we're seeing with Empires 0:01:28.680 --> 0:01:32.000 of AI today, where they are taking data that is 0:01:32.040 --> 0:01:34.480 not their own, their laying claim to it. They're taking land, 0:01:34.640 --> 0:01:38.720 they're taking energy, they're taking water. They are exploiting massive 0:01:38.720 --> 0:01:41.679 amounts of labor, both labor that goes into the inputs 0:01:41.959 --> 0:01:45.080 for developing these am models, but also exploiting labor in 0:01:45.080 --> 0:01:48.440 the sense that they are ultimately creating labor automating technologies 0:01:49.200 --> 0:01:52.200 that is eroding away people's labor rights as we speak, 0:01:53.240 --> 0:01:55.760 and they're doing it under this civilizing mission of they 0:01:55.800 --> 0:01:58.600 are doing it for the benefit of all of humanity. 0:01:59.800 --> 0:02:02.120 And what I say in the book is Empires of 0:02:02.160 --> 0:02:05.040 AI they're not as overtly violent as empires of old. 0:02:05.240 --> 0:02:09.799 And so maybe that can become confusing and people think, oh, well, 0:02:09.800 --> 0:02:12.359 it can't be that bad. But the thing is, we've 0:02:12.360 --> 0:02:15.399 had one hundred and fifty years of social and moral progress, 0:02:15.480 --> 0:02:18.280 and so empires of modern day are going to look 0:02:18.320 --> 0:02:22.280 different in the way that empires of old operated. And 0:02:22.520 --> 0:02:26.480 when you look just at like the actual parallels, there 0:02:26.480 --> 0:02:31.480 are just so many extraordinary parallels between the kind of 0:02:31.600 --> 0:02:35.120 basis of empire building back then and now that I 0:02:35.160 --> 0:02:39.080 think it is fundamentally the only frame that I have 0:02:39.160 --> 0:02:42.120 found to really help understand and grapple with the sheer 0:02:42.160 --> 0:02:46.280 scope and scale and the actual like what is actually 0:02:46.360 --> 0:02:48.280 happening here within the AI industry. 0:02:49.440 --> 0:02:51.799 One theme from the book I also noticed was that 0:02:52.320 --> 0:02:55.160 despite all of the backs and forths between all the people, 0:02:55.760 --> 0:02:58.680 very rarely product came out. Though, Like it was interesting, 0:02:58.720 --> 0:03:01.360 there seemed to be all that these sations about research 0:03:01.440 --> 0:03:04.360 and all of these things they were saying, but it 0:03:04.440 --> 0:03:06.800 usually just ended with like some sort of release and 0:03:06.840 --> 0:03:10.560 then kind of just moved on. Yeah, it almost it 0:03:10.639 --> 0:03:12.960 almost makes me wonder what they're always what they're working 0:03:13.000 --> 0:03:13.840 on half the time. 0:03:15.000 --> 0:03:18.000 Yeah, you know, it's I think it's a product of 0:03:18.080 --> 0:03:21.120 two different things that that you noticed that in the book. 0:03:21.200 --> 0:03:24.160 One is that I finished writing the book before a 0:03:24.200 --> 0:03:28.080 lot of the most recent product releases came out, Right, 0:03:28.440 --> 0:03:31.680 That's just the nature of writing things on the timescale of. 0:03:31.639 --> 0:03:34.160 Books, is it's not fun. 0:03:34.680 --> 0:03:39.920 Yeah. I froze the manuscript in like the early days 0:03:39.920 --> 0:03:43.520 of January, right before Deep Seek, right before Stargate, right 0:03:43.520 --> 0:03:48.760 before you know, a string of other releases. So that's 0:03:48.800 --> 0:03:53.080 one is that through most of open AI's history, it 0:03:53.160 --> 0:03:56.240 is really it was really more focused on research conversations, 0:03:56.280 --> 0:03:58.680 and it's only been in the last year or so 0:03:58.800 --> 0:04:02.520 that it's really dramatically shifted much more to talking about product. 0:04:03.080 --> 0:04:06.480 But the second reason is that I personally like that 0:04:06.600 --> 0:04:11.480 is my expertise. I came up in AI reporting covering 0:04:11.560 --> 0:04:15.160 the research, and so I wanted to focus on that 0:04:15.200 --> 0:04:18.280 in the book and really unpack it, especially because there's 0:04:18.360 --> 0:04:22.320 not as much reporting on the research these days, and 0:04:22.400 --> 0:04:24.640 I wanted to kind of track that history and the 0:04:24.640 --> 0:04:27.880 internal conversations that happened when people say that they're developing 0:04:28.200 --> 0:04:30.360 so called AGI and. 0:04:30.360 --> 0:04:32.560 You talk about in twenty nineteen in the book the 0:04:32.560 --> 0:04:35.480 your rose colored Glasses got knocked off by a story? 0:04:35.760 --> 0:04:39.040 What was it that really made you start being suspicious 0:04:39.080 --> 0:04:40.040 of these companies? 0:04:41.520 --> 0:04:45.360 Yeah, So in twenty nineteen was when I started covering 0:04:45.440 --> 0:04:48.000 Opening Eye, and I embedded within the company for three 0:04:48.080 --> 0:04:52.040 days in August of twenty nineteen to profile what had 0:04:52.080 --> 0:04:57.040 then become a newly minted capped profit nested in a nonprofit. 0:04:58.680 --> 0:05:00.880 And I think the thing that but that really started 0:05:00.880 --> 0:05:05.360 tipping me off was it was actually really small things. Initially, 0:05:05.640 --> 0:05:09.839 the first thing was they publicly professed to be this 0:05:10.000 --> 0:05:12.760 bastion of transparency and they were going to share all 0:05:12.839 --> 0:05:14.880 their research to the world, and they had accumulated a 0:05:14.880 --> 0:05:18.839 significant amount of good will on the basis of this idea, 0:05:19.320 --> 0:05:23.080 and they were raising not literally fundraising, but they had 0:05:23.120 --> 0:05:26.240 amassed a lot of capital on the basis of this idea. 0:05:26.560 --> 0:05:29.960 And when I started embedding within the company, I realized 0:05:29.960 --> 0:05:33.320 that they were incredibly secretive, like they were not. They 0:05:33.320 --> 0:05:37.479 wouldn't allow me to see anything or time to anyone 0:05:37.680 --> 0:05:42.720 beyond very strictly sanctioned conversations. And even in those conversations, 0:05:42.760 --> 0:05:45.719 I would notice that researchers were giving side eye to 0:05:45.760 --> 0:05:49.839 the communications officer every other sentence because they were worried 0:05:49.880 --> 0:05:53.520 about stepping into a lane that was considered proprietary. And 0:05:53.560 --> 0:05:55.800 I was like, wait a minute, why are there things 0:05:55.839 --> 0:05:58.280 that are proprietary, and why are the people being secretive 0:05:58.279 --> 0:06:02.200 if all this is supposed to ultimately be shared with 0:06:02.240 --> 0:06:05.800 the public. But the other thing was when I was 0:06:05.880 --> 0:06:09.680 talking with executives, Like the very first interview that I 0:06:09.760 --> 0:06:12.520 had was with Greg Brockman and Ilia Setskav, the CTO 0:06:12.520 --> 0:06:16.880 and she's scientist, And I just asked them very basic questions, 0:06:16.920 --> 0:06:19.160 like why do you think we should spend this much 0:06:19.200 --> 0:06:22.720 money on Agi and not on something else? And can 0:06:22.760 --> 0:06:25.599 you articulate for me what does AGI look like? What 0:06:25.640 --> 0:06:29.560 would you even want Agi to do? And can you 0:06:29.640 --> 0:06:32.559 articulate for me? You know, part of their origin story 0:06:32.600 --> 0:06:36.080 as a company was they want to build agi good 0:06:36.120 --> 0:06:39.640 agi first before the bad people built bad Agi. So 0:06:39.680 --> 0:06:42.159 I was like, well, what would bad agi look like 0:06:42.200 --> 0:06:45.280 as well? Or like what are the harms that are 0:06:45.279 --> 0:06:48.359 coming out of some of this rapid AI progress? And 0:06:48.920 --> 0:06:53.200 they weren't able to answer any of those questions. And 0:06:53.640 --> 0:06:56.359 that was when I thought, hold on a second, Like 0:06:56.480 --> 0:07:00.000 I thought that this was a nonprofit meant to care 0:07:00.080 --> 0:07:04.320 counter some of the ills of Silicon Valley, one of 0:07:04.360 --> 0:07:08.719 the ills being that the most companies end up being 0:07:08.760 --> 0:07:12.640 thrown boatloads of cash without like clear articulated ideas about 0:07:12.640 --> 0:07:15.520 what they're going to do with that cash. And here 0:07:15.560 --> 0:07:17.920 I am in this meeting room trying to just ask 0:07:18.000 --> 0:07:21.640 the most basic question, like the most boiler plate stuff 0:07:21.680 --> 0:07:25.920 that there should be some kind of answer too, uh, 0:07:25.960 --> 0:07:28.840 And they can't even answer that. So it seems like 0:07:29.040 --> 0:07:33.320 it is actually very much just an animal of Silicon Valley. 0:07:33.680 --> 0:07:36.640 This is not actually something different from what we're seeing 0:07:36.680 --> 0:07:37.880 with the rest of the tech industry. 0:07:38.840 --> 0:07:41.480 It felt as well, that was a comment and forgive 0:07:41.480 --> 0:07:43.640 me for forgetting it exactly where where it was like 0:07:43.720 --> 0:07:45.600 the all secrets could be written on a grain of 0:07:45.680 --> 0:07:50.800 rice or something like that. Yeah, and I have to admit, 0:07:51.000 --> 0:07:53.120 as I read it, I got this weird feeling like 0:07:53.160 --> 0:07:56.720 does anyone actually have any ip because when you actually 0:07:56.720 --> 0:07:59.480 look at the conversations they're having, and you likely privy 0:07:59.520 --> 0:08:02.040 to more hair, it felt like they wouldn't talk about 0:08:02.040 --> 0:08:05.080 what they were doing at all, and not I say, 0:08:05.120 --> 0:08:07.960 this is run apof written a lot about the valley. 0:08:08.520 --> 0:08:11.760 It feels like they'd say more, but no one wanted 0:08:11.800 --> 0:08:14.679 to say anything, not even secret. It's like nobody really knew. 0:08:15.080 --> 0:08:17.800 And you even described some of the managerial stuff in there, 0:08:17.800 --> 0:08:19.760 like no one really knew what was going on anyway. 0:08:19.960 --> 0:08:23.880 It just feels like a remarkably disorganized company considering the scale. 0:08:25.160 --> 0:08:28.000 Yeah, so I think early on at Opening Eye was 0:08:28.040 --> 0:08:31.360 completely disorganized in the sense that they had no idea. 0:08:31.920 --> 0:08:34.680 You know, they decide, Okay, we're going to build this 0:08:34.720 --> 0:08:36.559 ag I think, but then they were like, what does 0:08:36.600 --> 0:08:40.000 that even mean? We have no idea, And there was 0:08:40.240 --> 0:08:42.760 a lot of there weren't real managers at the company either, 0:08:42.800 --> 0:08:46.080 because they had just gathered up a bunch of researchers 0:08:46.120 --> 0:08:49.560 from academia, and they didn't really have much of a 0:08:49.600 --> 0:08:52.360 sense of how to organize themselves other than a traditional 0:08:52.400 --> 0:08:56.440 academic lab where there's a professor and grad students, and 0:08:57.679 --> 0:09:02.199 I mean academia, you know, as it's has its function, 0:09:02.600 --> 0:09:06.040 But that ultimately wasn't the right structure for trying to 0:09:06.559 --> 0:09:10.720 move a group of people towards a similar goal. And 0:09:10.760 --> 0:09:14.199 over time, Opening Eye did start cleaning itself up a 0:09:14.240 --> 0:09:16.800 little bit, It did start restructuring itself. It started focusing 0:09:16.840 --> 0:09:20.199 more on GPT models because they hit on that in 0:09:20.320 --> 0:09:26.080 around twenty eighteen twenty nineteen. But similarly, there's still just 0:09:26.240 --> 0:09:31.160 because there's no clarity about its mission and ultimately what 0:09:31.240 --> 0:09:34.680 it is trying to build, you end up with just 0:09:34.920 --> 0:09:38.800 a lot of riffs within the organization over this very 0:09:38.840 --> 0:09:42.320 fundamental question. People fundamentally disagree about what Opening I is. 0:09:43.000 --> 0:09:46.400 They disagree about what AGI is, They disagree about what 0:09:46.440 --> 0:09:49.520 it means to ensure that something should benefit all of humanity. 0:09:51.240 --> 0:09:56.040 And I think because there was all this confusion or 0:09:56.120 --> 0:09:59.520 there were all these different interpretations ultimately of these like 0:09:59.600 --> 0:10:04.200 basic tenants of the organization. I think people also just 0:10:04.920 --> 0:10:08.120 there was they wouldn't quite clearly articulate to one another 0:10:08.400 --> 0:10:10.679 what they were doing. It wasn't necessarily that they were 0:10:10.720 --> 0:10:13.600 trying to be secretive to one another. It was more 0:10:13.760 --> 0:10:16.880 just that they weren't really on the same page. And 0:10:18.840 --> 0:10:23.080 this eventually became sort of less and less true in 0:10:23.080 --> 0:10:26.480 the sense that as sam Oltman's installed himself as CEO 0:10:26.520 --> 0:10:31.760 and started really exerting a particular type of path for 0:10:31.920 --> 0:10:35.160 AI progress, then they started having, you know, research documents 0:10:35.200 --> 0:10:38.480 that explicitly articulated we are a scaling lab, like we 0:10:38.520 --> 0:10:41.160 are going after scale way. 0:10:41.280 --> 0:10:44.480 How long did it take them to put those documents 0:10:44.480 --> 0:10:46.360 together though? What year about? 0:10:47.520 --> 0:10:50.360 I think their first research roadmap was in twenty seventeen, 0:10:50.400 --> 0:10:52.480 so it was one and a half years to them 0:10:53.000 --> 0:10:57.520 so bad into the non profit Yeah, yeah. 0:10:56.640 --> 0:10:59.600 So I will admit there is another colonial thing that 0:10:59.640 --> 0:11:04.320 stood out well. Two specifically, One, it definitely feels that 0:11:04.320 --> 0:11:07.240 there are a lot of unintelligent cousin types who were 0:11:07.240 --> 0:11:09.840 put in there because their mate was there throughout the company. 0:11:10.240 --> 0:11:13.800 But two it's this kind of religious view around agi, 0:11:14.280 --> 0:11:18.200 this kind of nebulous justification for just about every I 0:11:18.240 --> 0:11:20.240 was disappointed, and I understand what you mentioned it like 0:11:20.640 --> 0:11:23.480 did Yadowski was in there? I think the less wrong people. 0:11:23.640 --> 0:11:26.640 This is a personal belief. Just no need to mention 0:11:26.720 --> 0:11:29.679 them again for anyone. I think that Yadowski, anyone who 0:11:29.679 --> 0:11:31.840 writes a six hundred thousand word Harry Potter book should 0:11:31.880 --> 0:11:35.760 be put in prison, including JK. Rowling. But it feels 0:11:35.840 --> 0:11:38.960 like there really is this belief system that's pushed throughout 0:11:38.960 --> 0:11:43.080 this industry which mirrors colonialism. Mirror is the very judao 0:11:43.200 --> 0:11:46.880 Christian push of the British and many other colonial entities. 0:11:47.760 --> 0:11:50.320 Yeah. Absolutely so. One of the things that I was 0:11:50.480 --> 0:11:54.760 most surprised by when reporting the book is I had 0:11:54.840 --> 0:11:58.320 seen all the divisions around boomers and boomers people saying hey, 0:11:58.400 --> 0:12:00.760 I can bring us to Utopiera, people saying I can 0:12:00.840 --> 0:12:04.080 kill us all. I really did think initially that it 0:12:04.160 --> 0:12:07.200 was just rhetoric and that it was just a tool 0:12:07.360 --> 0:12:11.120 for accruing more power. And the thing that surprised me 0:12:11.280 --> 0:12:16.319 most was how many people I met that genuinely deeply 0:12:16.320 --> 0:12:19.240 believed in both, especially the Dumer ideology like I was 0:12:19.240 --> 0:12:23.200 interviewing people whose voices were quivering because they were talking 0:12:23.240 --> 0:12:27.320 about their anxiety around the potential end of the world, 0:12:27.400 --> 0:12:31.880 and that was a very sincere reaction. And I think 0:12:31.920 --> 0:12:34.000 that is part You're exactly right that it is a 0:12:34.080 --> 0:12:37.720 huge parallel with empire building in the past, is that 0:12:38.280 --> 0:12:43.680 empires need to have an ideology that convinces themselves why 0:12:43.720 --> 0:12:46.800 they are ultimately doing something that is for the benefit 0:12:46.840 --> 0:12:49.680 of the world. So in the past, when they had 0:12:49.679 --> 0:12:52.280 the civilizing mission, we're bringing this to the world. It 0:12:52.360 --> 0:12:57.320 also wasn't rhetoric. It was also a deeply seated religious 0:12:57.840 --> 0:13:02.600 and spiritual and scientific belief that they were doing something 0:13:02.640 --> 0:13:04.440 that was better off for everyone. 0:13:04.920 --> 0:13:07.600 I mean, the origins of the BBC in England were 0:13:07.880 --> 0:13:12.600 religious indoctrination on some level. It kind of it's I 0:13:12.679 --> 0:13:16.800 admit I'm surprised to hear the quivering voices stuff I 0:13:16.840 --> 0:13:21.360 think because I think that again, this personal opinion, Ydowski, 0:13:21.480 --> 0:13:23.000 I think is full. I think a lot of those 0:13:23.040 --> 0:13:24.840 less wrong guys are full of shit. I think they're 0:13:24.880 --> 0:13:27.440 doing it for not for the bit. But it's the 0:13:27.480 --> 0:13:31.800 same kind of horse trading shit that people do around anything. 0:13:31.840 --> 0:13:33.480 It's like, we don't have anything to believe in, so 0:13:33.559 --> 0:13:36.680 let's all agree on this. But it's interesting to hear 0:13:36.720 --> 0:13:40.320 that people are I don't know how to put this, 0:13:40.480 --> 0:13:44.360 actually believing this crap even though it doesn't feel like 0:13:44.520 --> 0:13:46.560 there's any real evidence. You know. 0:13:46.880 --> 0:13:50.280 Yeah, well, I think the analogy that I've started using 0:13:50.400 --> 0:13:54.520 is I really feel like opening Eyes done, where you know, 0:13:54.679 --> 0:13:57.920 in Dune, there is a mythology that is created by 0:13:57.960 --> 0:14:00.960 a certain group of people with four understanding that they're 0:14:00.960 --> 0:14:06.120 creating a mythology, right right, But then as they start 0:14:06.160 --> 0:14:10.640 to embody and act out this mythology, not only do 0:14:11.160 --> 0:14:13.360 many many people who didn't know that it was originally 0:14:13.400 --> 0:14:16.400 created come to believe it, also the people who created 0:14:16.400 --> 0:14:20.000 it come to believe it themselves. And I think this 0:14:20.240 --> 0:14:26.960 is essentially exactly what is happening within AI with the ideologies, 0:14:27.040 --> 0:14:30.520 is that maybe there was at some point someone who 0:14:30.960 --> 0:14:36.520 was more aware that there was some kind of rhetorical 0:14:36.600 --> 0:14:39.880 trick that they were playing around, really propagating this kind 0:14:39.920 --> 0:14:43.040 of belief, but it is not We're not at that 0:14:43.120 --> 0:14:46.160 point anymore. Like, there are lots and lots of people 0:14:46.160 --> 0:14:49.840 who genuinely believe these things, and I think it's self 0:14:49.880 --> 0:14:53.640 perpetuating because when you believe it, you look for signs 0:14:53.680 --> 0:14:58.320 of it, and you research things that would suggest more 0:14:58.400 --> 0:15:01.800 evidence for your belief and so they're kind of continuing 0:15:01.840 --> 0:15:04.960 to reinforce their beliefs. And the more these A models 0:15:04.960 --> 0:15:09.320 have progressed, the stronger these beliefs have become. Because whether 0:15:09.400 --> 0:15:12.840 you believe AI will bring utopia or dystopia, there is 0:15:12.880 --> 0:15:16.320 an abundance of evidence that you can point to now 0:15:16.800 --> 0:15:21.000 to reinforce your own yeah, exactly, to reinforce your own 0:15:21.800 --> 0:15:24.160 starting point. And so it's sort of like a microcosm 0:15:25.120 --> 0:15:29.520 of society today where you know, most the average person 0:15:29.720 --> 0:15:32.880 no longer encounters information that can change their minds. It 0:15:32.960 --> 0:15:36.600 just continues to entrench whatever they already believed before. 0:15:37.120 --> 0:15:40.840 Do you believe Sam Moultman believes this shite? Do you 0:15:40.840 --> 0:15:44.240 think he believes in the AGI? Is he part of it? 0:15:44.240 --> 0:15:49.360 It's really It's really interesting because I think the no 0:15:49.440 --> 0:15:52.560 matter who I interviewed, and no matter how long they 0:15:52.600 --> 0:15:55.880 worked with Sam Aultman or how closely they worked with 0:15:55.960 --> 0:15:58.360 Sam Altman, and not a single person was able to 0:15:58.560 --> 0:16:02.760 fully articulate what his beliefs are. And I think that 0:16:02.840 --> 0:16:03.920 is very much by design. 0:16:04.240 --> 0:16:08.000 Is that it's beautiful that's. 0:16:07.720 --> 0:16:11.360 Yeah, and and and and and it wasn't. And they 0:16:11.360 --> 0:16:14.800 would explicitly say this too. They would they would call out, 0:16:15.000 --> 0:16:19.720 I'm not actually sure what he believes. And this was 0:16:19.800 --> 0:16:34.120 the most consistent thing that people said about him. 0:16:34.400 --> 0:16:37.840 I really noticed as well. If you read your book 0:16:38.120 --> 0:16:40.800 and you really look, you actually can't get much of 0:16:40.800 --> 0:16:43.040 an idea of who Sam Mortman is at all, And 0:16:43.080 --> 0:16:46.200 in fact, you can't work out why he's brilliant at all. 0:16:46.200 --> 0:16:47.880 And I've read a lot of stuff about Sam Moultman. 0:16:48.120 --> 0:16:50.120 The long and short of it I can understand is 0:16:50.200 --> 0:16:54.400 that he's got good psychology and he's really charming. Everyone 0:16:54.440 --> 0:16:57.360 talks about the psychology and the charming, and it's just 0:16:57.520 --> 0:17:01.040 really it. It is so busy. He's so a bizarre 0:17:01.120 --> 0:17:04.160 man like everything about him is so like. The way 0:17:04.200 --> 0:17:07.680 that people talk about him is so strange. 0:17:08.000 --> 0:17:10.560 Yeah, So this is what I sort of concluded about 0:17:10.600 --> 0:17:15.160 why he's able to pull off what he does. He 0:17:15.320 --> 0:17:17.600 is a once in a generation talent when it comes 0:17:17.600 --> 0:17:23.720 to storytelling, and he has a loose relationship with the truth, yes, 0:17:23.800 --> 0:17:27.600 which is a really powerful combination. And so when he's 0:17:27.640 --> 0:17:31.960 talking to someone, he shines most when he's talking to 0:17:32.040 --> 0:17:35.639 small groups of people in one on one meetings, and 0:17:35.720 --> 0:17:40.280 what he says is more tightly correlated with what that 0:17:40.440 --> 0:17:45.080 person needs to hear rather than what he believes, which 0:17:45.119 --> 0:17:47.040 is part of the reason why people say, ultimately like 0:17:47.080 --> 0:17:50.000 they don't really know what he believes because he doesn't 0:17:50.119 --> 0:17:54.439 really indicate it. And so I think that is what 0:17:54.520 --> 0:17:58.600 makes him incredibly persuasive. And he is really good at 0:17:58.640 --> 0:18:01.960 understanding people and what they and what they want, and 0:18:02.080 --> 0:18:04.160 you know, he's well resourced, so he's able to then 0:18:04.200 --> 0:18:07.000 deliver to them what they need and want. And what 0:18:07.080 --> 0:18:13.679 I realized is with that kind of talent, you would 0:18:13.720 --> 0:18:19.280 inherently be incredibly polarizing as a figure, because if the 0:18:19.320 --> 0:18:23.479 person agrees with you, you're the best asset in the 0:18:23.520 --> 0:18:27.200 world for what they want to achieve. You're incredibly persuasive, 0:18:27.440 --> 0:18:31.200 You're able to get the resources you can do exactly 0:18:31.840 --> 0:18:34.680 what that person can do for you, exactly what you 0:18:34.720 --> 0:18:37.879 want them to do. But if you disagree with this person, 0:18:38.119 --> 0:18:42.639 that person becomes the greatest threat ever because they are 0:18:42.680 --> 0:18:46.159 so persuasive. You have fear that they're going to be 0:18:46.200 --> 0:18:48.399 able to carry out exactly what you don't want them 0:18:48.440 --> 0:18:51.400 to carry out. And so that kind of boils down 0:18:51.440 --> 0:18:56.639 to why he's just such an enigmatic and extremely polarizing person. 0:18:57.440 --> 0:18:59.520 Is it really depends on whether or not someone agrees 0:18:59.600 --> 0:19:00.440 or disagree with him. 0:19:00.680 --> 0:19:02.680 He also doesn't seem that smart. I don't know, he 0:19:02.720 --> 0:19:04.800 seems quite good at talking to people, but when I 0:19:04.840 --> 0:19:07.760 hear him talk, he doesn't seem that eloquent. And it 0:19:07.800 --> 0:19:10.800 makes me wonder if perhaps Sam Mortman is a symptom 0:19:10.840 --> 0:19:12.840 of a greater problem that so much of our power, 0:19:12.880 --> 0:19:16.200 structure and money is based on someone making decisions based 0:19:16.240 --> 0:19:19.399 on the last intelligent person or intelligent seeming person they 0:19:19.480 --> 0:19:19.840 talk to. 0:19:21.200 --> 0:19:24.639 Yeah, I mean, I think our society is also just 0:19:24.800 --> 0:19:31.360 we still have such a have we worship people that 0:19:31.480 --> 0:19:36.199 are wealthy. Yeah, like, And so even if he's not 0:19:36.480 --> 0:19:40.199 saying something that is convincing you in real time, he 0:19:40.320 --> 0:19:43.040 has all of the kind of indicators that this person 0:19:43.119 --> 0:19:46.679 has been remarkably successful and you should listen to what 0:19:46.720 --> 0:19:49.200 he says because then that will make you successful too, right, 0:19:50.280 --> 0:19:53.040 And so I think that is part of the part 0:19:53.200 --> 0:19:56.960 of the kind of mythos around him, is that if 0:19:56.960 --> 0:20:00.240 you can join up with him, it will greatly enrich you. 0:20:00.800 --> 0:20:02.439 And you know, like, there's a lot of evidence to 0:20:02.440 --> 0:20:05.520 suggest that too, that like, there have been plenty of 0:20:05.520 --> 0:20:08.560 people that I've allied themselves with Sam Aulman and that 0:20:08.600 --> 0:20:12.879 have become much richer for it. And so whether or 0:20:12.920 --> 0:20:16.879 not people are joining up with him because they necessarily 0:20:16.920 --> 0:20:19.320 one hundred percent agree with like his ideology or his 0:20:19.400 --> 0:20:22.720 actions or anything like that, or if it's if it's 0:20:22.760 --> 0:20:26.280 more because ultimately they get to benefit from that alliance, 0:20:27.200 --> 0:20:29.080 I think is Yeah, what. 0:20:29.000 --> 0:20:31.720 Most feels like, it's people connecting with other people to 0:20:31.760 --> 0:20:34.560 see how far they can get, far more than AI. 0:20:34.800 --> 0:20:37.639 Because one of the other things I really noticed when 0:20:37.680 --> 0:20:41.160 you were telling the story of the firing Sam Wulman 0:20:41.160 --> 0:20:44.080 getting fired in No. Twenty twenty three, As much as 0:20:44.080 --> 0:20:47.440 people wanted to pretend, they kept bringing up the tender 0:20:48.000 --> 0:20:50.080 and to explain for the listeners the tender was the 0:20:50.119 --> 0:20:52.960 open AI hadn't had a plans to let have plans 0:20:52.960 --> 0:20:55.479 to let people sell their stock. It really felt like 0:20:55.560 --> 0:20:59.680 that was more the primary concern than any loyalty to Altman. 0:21:00.800 --> 0:21:02.960 It was. I wouldn't say it was the primary concern, 0:21:03.000 --> 0:21:06.080 but I mean, yeah, it really depended on. 0:21:06.080 --> 0:21:06.639 Who I was. 0:21:07.320 --> 0:21:11.679 It's taught, Yeah, exactly like every employee sort of had 0:21:11.680 --> 0:21:15.639 a different calculus that ultimately led them to revolt against 0:21:15.640 --> 0:21:18.720 the board and want Altman back, And there were different 0:21:18.720 --> 0:21:21.760 calculuses among Microsoft and investors. But one of the key 0:21:21.800 --> 0:21:25.679 things that I think is uh necessary to understand just 0:21:25.800 --> 0:21:29.040 why there is so much seeming loyalty around Altman in general, 0:21:29.200 --> 0:21:33.960 is he is very very good at at establishing relationships 0:21:34.840 --> 0:21:38.159 with money involved, where he is the lynchpin to the 0:21:38.200 --> 0:21:42.160 other person accessing that much yes, and so the tender 0:21:42.200 --> 0:21:44.720 offer is a perfect example of this, and that employees 0:21:45.000 --> 0:21:48.680 ultimately they realized that Altman is just really he's really 0:21:48.680 --> 0:21:52.320 good at fundraising, and whether or not an employee believes 0:21:52.359 --> 0:21:56.000 in the AGI thing, they all agreed that open AA 0:21:56.040 --> 0:21:59.080 ultimately needs an enormous amount of capital, and also many 0:21:59.160 --> 0:22:01.160 of them are doing it in part because they can 0:22:01.200 --> 0:22:05.239 then like guarantee their own financial future. And so with 0:22:05.960 --> 0:22:09.679 Altman gone, it became increasingly clear that open a I 0:22:09.720 --> 0:22:12.760 wouldn't survive, and so that's not something that a lot 0:22:12.760 --> 0:22:16.440 of employees wanted. It became clear that even if Opening 0:22:16.480 --> 0:22:18.960 I did survive, they would be a lot more short 0:22:19.080 --> 0:22:21.240 changed in terms of the amount of capital that they 0:22:21.280 --> 0:22:23.840 would be able to get because he would no longer 0:22:23.880 --> 0:22:27.960 be their champion for that, and also the tender offer 0:22:28.000 --> 0:22:30.480 could potentially go away and they would not be personally 0:22:30.920 --> 0:22:35.000 enriched as well, and you know, many of them. The 0:22:35.040 --> 0:22:39.000 thing to also understand is like it's very expensive to 0:22:39.000 --> 0:22:42.040 live in the Bay Area, and so for the worker, 0:22:42.240 --> 0:22:45.240 for the employees, in the moment, losing the tender offer 0:22:45.359 --> 0:22:48.480 wasn't like oh no, I'm going to lose like my retirement. 0:22:48.640 --> 0:22:51.560 It was also this sense of like, I'm literally going 0:22:51.640 --> 0:22:55.760 to lose my financial security right now, like I already 0:22:56.280 --> 0:22:59.320 tried to, I already bought, you know, a house based 0:22:59.320 --> 0:23:02.600 on the fact. I feel like you mentioned that yeah, yeah. 0:23:02.640 --> 0:23:04.400 That was someone who put money. 0:23:04.680 --> 0:23:07.960 Yeah, who put plenty of money down, and that the 0:23:09.040 --> 0:23:13.440 tender offer dissolving was a real financial stress. It was 0:23:13.760 --> 0:23:16.040 a threat to their financial existence in the life. 0:23:16.080 --> 0:23:19.359 I imagine so, But I'm just the way it was 0:23:19.440 --> 0:23:22.800 framed in public was that this was some big loyalty 0:23:22.880 --> 0:23:25.480 thing where everyone was like, I love open Ai and 0:23:25.560 --> 0:23:29.879 Sam Altman, and that just didn't feel like it was 0:23:29.920 --> 0:23:32.760